Dual-Licensing Linux Kernel with GPL V2 and GPL V3

Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Tarkan Erimer]

Hi,

As we know the forthcoming GPL V3 will be not compatible with the GPL V2 
and Linux Kernel is GPL V2 only.
So, another point is, which is previously mentioned by Linus and others, 
that if it is decided to upgrade the Linux Kernel's License to GPL V3, 
it is needed the permission of all the maintainers permission who 
contributed to the Linux Kernel and there are a lot of lost or dead 
maintainers. Which makes it impossible to get all the maintainers' 
permission.
But; if the Linux kernel should Dual-Licensed (GPL V2 and GPL V3), it 
will allow us the both worlds' fruits like code exchanging from other 
Open Source Projects (OpenSolaris etc.) that is compatible with GPL V3 
and not with GPL V2 and of course the opposite is applicable,too.

So;at this situation, what is possibility to make the Linux Kernel 
Dual-Licensed as I mentioned above and what is your opinions and 
suggestions about this idea ?

Regards,

Tarkan Erimer

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Neil Brown]

On Saturday June 9, tarkan@netone.net.tr wrote:
> Hi,
> 
> As we know the forthcoming GPL V3 will be not compatible with the GPL V2 
> and Linux Kernel is GPL V2 only.
> So, another point is, which is previously mentioned by Linus and others, 
> that if it is decided to upgrade the Linux Kernel's License to GPL V3, 
> it is needed the permission of all the maintainers permission who 
> contributed to the Linux Kernel and there are a lot of lost or dead 
> maintainers. Which makes it impossible to get all the maintainers' 
> permission.

You don't need the permission of maintainers.  You need the permission
of copyright owners.  The two groups overlap, but are not the same.
Dead people cannot own anything, even copyright.  Their estate
probably can.  I don't think it is theoretically impossible to get
everyone's permission, though it may be quite close to practically
impossible. 

> But; if the Linux kernel should Dual-Licensed (GPL V2 and GPL V3), it 
> will allow us the both worlds' fruits like code exchanging from other 
> Open Source Projects (OpenSolaris etc.) that is compatible with GPL V3 
> and not with GPL V2 and of course the opposite is applicable,too.
> 
> So;at this situation, what is possibility to make the Linux Kernel 
> Dual-Licensed as I mentioned above and what is your opinions and 
> suggestions about this idea ?

Dual licensing is no easier.  It means it is licensed to be used under
either license.  You already have permission to use it under GPLv2.
So to get a dual license, you precisely need to get access under GPLv3
i.e. to convince copyright owners to make that license grant.  A thing
that we have already agreed is at least "hard".

NeilBrown

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Jan-Benedict Glaw]

[-- Attachment #1: Type: text/plain, Size: 1508 bytes --]

On Sat, 2007-06-09 15:57:55 +1000, Neil Brown <neilb@suse.de> wrote:
> On Saturday June 9, tarkan@netone.net.tr wrote:
> > As we know the forthcoming GPL V3 will be not compatible with the GPL V2 
> > and Linux Kernel is GPL V2 only.
> > So, another point is, which is previously mentioned by Linus and others, 
> > that if it is decided to upgrade the Linux Kernel's License to GPL V3, 
> > it is needed the permission of all the maintainers permission who 
> > contributed to the Linux Kernel and there are a lot of lost or dead 
> > maintainers. Which makes it impossible to get all the maintainers' 
> > permission.
> 
> You don't need the permission of maintainers.  You need the permission
> of copyright owners.  The two groups overlap, but are not the same.
> Dead people cannot own anything, even copyright.  Their estate
> probably can.  I don't think it is theoretically impossible to get
> everyone's permission, though it may be quite close to practically
> impossible. 

And the next question is: How much copyright does a copyright owner
own?  For example, think of drivers written by one person, but a small
number of lines changed here and there by others to adopt the code to
new APIs.  Ask them all, I think?

MfG, JBG

-- 
      Jan-Benedict Glaw      jbglaw@lug-owl.de              +49-172-7608481
Signature of:             God put me on earth to accomplish a certain number of
the second  :            things. Right now I am so far behind I will never die.

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Tarkan Erimer]

Jan-Benedict Glaw wrote:
> On Sat, 2007-06-09 15:57:55 +1000, Neil Brown <neilb@suse.de> wrote:
>   
>> On Saturday June 9, tarkan@netone.net.tr wrote:
>>     
>>> As we know the forthcoming GPL V3 will be not compatible with the GPL V2 
>>> and Linux Kernel is GPL V2 only.
>>> So, another point is, which is previously mentioned by Linus and others, 
>>> that if it is decided to upgrade the Linux Kernel's License to GPL V3, 
>>> it is needed the permission of all the maintainers permission who 
>>> contributed to the Linux Kernel and there are a lot of lost or dead 
>>> maintainers. Which makes it impossible to get all the maintainers' 
>>> permission.
>>>       
>> You don't need the permission of maintainers.  You need the permission
>> of copyright owners.  The two groups overlap, but are not the same.
>> Dead people cannot own anything, even copyright.  Their estate
>> probably can.  I don't think it is theoretically impossible to get
>> everyone's permission, though it may be quite close to practically
>> impossible. 
>>     
>
> And the next question is: How much copyright does a copyright owner
> own?  For example, think of drivers written by one person, but a small
> number of lines changed here and there by others to adopt the code to
> new APIs.  Ask them all, I think?
>
> MfG, JBG
>
>   

And maybe another questions should be : How long a copyright owner can 
hold the copyright, if died or lost for sometime ? if died, the 
copyright still should be valid or not ? If lost, what the law orders at 
this point for copyright holding ?

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Tarkan Erimer]

david@lang.hm wrote:
> On Sun, 10 Jun 2007, Tarkan Erimer wrote:
>
>> Date: Sun, 10 Jun 2007 11:43:28 +0300
>> From: Tarkan Erimer <tarkan@netone.net.tr>
>> To: linux-kernel@vger.kernel.org
>> Subject: Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3
>>
>> Jan-Benedict Glaw wrote:
>>>  On Sat, 2007-06-09 15:57:55 +1000, Neil Brown <neilb@suse.de> wrote:
>>>
>>> >  On Saturday June 9, tarkan@netone.net.tr wrote:
>>> > > >  As we know the forthcoming GPL V3 will be not compatible with 
>>> the GPL > >  V2 and Linux Kernel is GPL V2 only.
>>> > >  So, another point is, which is previously mentioned by Linus 
>>> and > >  others, that if it is decided to upgrade the Linux Kernel's 
>>> License to > >  GPL V3, it is needed the permission of all the 
>>> maintainers permission > >  who contributed to the Linux Kernel and 
>>> there are a lot of lost or > >  dead maintainers. Which makes it 
>>> impossible to get all the > >  maintainers' permission.
>>> > > >  You don't need the permission of maintainers.  You need the 
>>> permission
>>> >  of copyright owners.  The two groups overlap, but are not the same.
>>> >  Dead people cannot own anything, even copyright.  Their estate
>>> >  probably can.  I don't think it is theoretically impossible to get
>>> >  everyone's permission, though it may be quite close to practically
>>> >  impossible. >
>>>  And the next question is: How much copyright does a copyright owner
>>>  own?  For example, think of drivers written by one person, but a small
>>>  number of lines changed here and there by others to adopt the code to
>>>  new APIs.  Ask them all, I think?
>>>
>>>  MfG, JBG
>>>
>>>
>>
>> And maybe another questions should be : How long a copyright owner 
>> can hold the copyright, if died or lost for sometime ? if died, the 
>> copyright still should be valid or not ? If lost, what the law orders 
>> at this point for copyright holding ?
>
> I believe that in the US it's life + 90 years.
>
> David Lang
Hmm... Really,it is damn too much time to wait! It's really better idea 
to replace the code of this person as said before instead of waiting 
such  90+ years!

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[david]

On Sun, 10 Jun 2007, Tarkan Erimer wrote:

>> >  And maybe another questions should be : How long a copyright owner can 
>> >  hold the copyright, if died or lost for sometime ? if died, the 
>> >  copyright still should be valid or not ? If lost, what the law orders at 
>> >  this point for copyright holding ?
>>
>>  I believe that in the US it's life + 90 years.
>>
>>  David Lang
> Hmm... Really,it is damn too much time to wait! It's really better idea to 
> replace the code of this person as said before instead of waiting such  90+ 
> years!

exactly, however as others are pointing out, there are a lot of active 
developers who do not agree with some of the key points of the GPLv3 
(including Linus), so until you convince them that the GPLv3 is better it 
really doesn't matter how hard it is to deal with the people who you can't 
contact.

David Lang

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[debian developer]

On 6/10/07, david@lang.hm <david@lang.hm> wrote:
> On Sun, 10 Jun 2007, Tarkan Erimer wrote:
>
> >> >  And maybe another questions should be : How long a copyright owner can
> >> >  hold the copyright, if died or lost for sometime ? if died, the
> >> >  copyright still should be valid or not ? If lost, what the law orders at
> >> >  this point for copyright holding ?
> >>
> >>  I believe that in the US it's life + 90 years.
> >>
> >>  David Lang
> > Hmm... Really,it is damn too much time to wait! It's really better idea to
> > replace the code of this person as said before instead of waiting such  90+
> > years!
>
> exactly, however as others are pointing out, there are a lot of active
> developers who do not agree with some of the key points of the GPLv3
> (including Linus), so until you convince them that the GPLv3 is better it

Last heard, Linus was quite impressed with the toned down version of
the final draft of GPLv3. I think Linus, and other major developers
should make their stand on this issue clear so that the kernel
community can discuss the future steps.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Tarkan Erimer]

debian developer wrote:
> On 6/10/07, david@lang.hm <david@lang.hm> wrote:
>> On Sun, 10 Jun 2007, Tarkan Erimer wrote:
>>
>> >> >  And maybe another questions should be : How long a copyright 
>> owner can
>> >> >  hold the copyright, if died or lost for sometime ? if died, the
>> >> >  copyright still should be valid or not ? If lost, what the law 
>> orders at
>> >> >  this point for copyright holding ?
>> >>
>> >>  I believe that in the US it's life + 90 years.
>> >>
>> >>  David Lang
>> > Hmm... Really,it is damn too much time to wait! It's really better 
>> idea to
>> > replace the code of this person as said before instead of waiting 
>> such  90+
>> > years!
>>
>> exactly, however as others are pointing out, there are a lot of active
>> developers who do not agree with some of the key points of the GPLv3
>> (including Linus), so until you convince them that the GPLv3 is 
>> better it
>
> Last heard, Linus was quite impressed with the toned down version of
> the final draft of GPLv3. I think Linus, and other major developers
> should make their stand on this issue clear so that the kernel
> community can discuss the future steps.
Yep, the GPLv3 probably will release around July time. So;luckily, we 
had very little time to see the final decision about it :-) I hope we 
should upgrade to GPLv3 and Sun should "Dual License" the OpenSolaris 
via GPLv3 (or at least,GPLv3 should be CDDL compatible.). So,we should 
have more fruits (like ZFS,DTrace etc.) ;-)

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Greg KH]

On Sun, Jun 10, 2007 at 05:21:53PM +0300, Tarkan Erimer wrote:
>  I hope we should upgrade to GPLv3 and Sun should "Dual License" the
>  OpenSolaris via GPLv3 (or at least,GPLv3 should be CDDL compatible.).

The OpenSolaris community has already stated that they do not want to
accept GPLv3, why not discuss this with them if you want to try to
change their minds?

>  So,we should have more fruits (like ZFS,DTrace etc.) ;-)

I think the transfer would be more the other way, we have a zillion more
things that they do not than the other way around.

thanks,

greg k-h

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Tarkan Erimer]

Greg KH wrote:
> On Sun, Jun 10, 2007 at 05:21:53PM +0300, Tarkan Erimer wrote:
>   
>>  I hope we should upgrade to GPLv3 and Sun should "Dual License" the
>>  OpenSolaris via GPLv3 (or at least,GPLv3 should be CDDL compatible.).
>>     
>
> The OpenSolaris community has already stated that they do not want to
> accept GPLv3, why not discuss this with them if you want to try to
> change their minds?
>   
It was just an example came to my mind at first when thinking about 
"Dual Licensing" or upgrading Linux Kernel to the GPLv3. Yeah maybe,the 
"OpenSolaris Community" do not want GPLv3. But; IMHO, it is in the hands 
of "Sun" not the "OpenSolaris Community".

Regards,

Tarkan

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Al Viro]

On Mon, Jun 11, 2007 at 09:46:18AM +0300, Tarkan Erimer wrote:
> Greg KH wrote:
> >On Sun, Jun 10, 2007 at 05:21:53PM +0300, Tarkan Erimer wrote:
> >  
> >> I hope we should upgrade to GPLv3 and Sun should "Dual License" the
> >> OpenSolaris via GPLv3 (or at least,GPLv3 should be CDDL compatible.).
> >>    
> >
> >The OpenSolaris community has already stated that they do not want to
> >accept GPLv3, why not discuss this with them if you want to try to
> >change their minds?
> >  
> It was just an example came to my mind at first when thinking about 
> "Dual Licensing" or upgrading Linux Kernel to the GPLv3. Yeah maybe,the 
> "OpenSolaris Community" do not want GPLv3. But; IMHO, it is in the hands 
> of "Sun" not the "OpenSolaris Community".

Perhaps.  However, since the only thing in hands of your kind of advocates
is best not mentioned on a family-friendly maillist, may I suggest taking
that exciting thread to more appropriate place?

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Tarkan Erimer]

Al Viro wrote:
> Perhaps.  However, since the only thing in hands of your kind of advocates
> is best not mentioned on a family-friendly maillist, may I suggest taking
> that exciting thread to more appropriate place?
>   
I don't think that this thread is going unfriendly or harmfully. 
However, what is your suggestion ?

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Ingo Molnar]

* Tarkan Erimer <tarkan@netone.net.tr> wrote:

> > > >> I hope we should upgrade to GPLv3 and Sun should "Dual License" 
> > > >> the OpenSolaris via GPLv3 (or at least,GPLv3 should be CDDL 
> > > >> compatible.).
> > > >
> > > > The OpenSolaris community has already stated that they do not 
> > > > want to accept GPLv3, why not discuss this with them if you want 
> > > > to try to change their minds?
> > >
> > > It was just an example came to my mind at first when thinking 
> > > about "Dual Licensing" or upgrading Linux Kernel to the GPLv3. 
> > > Yeah maybe,the "OpenSolaris Community" do not want GPLv3. But; 
> > > IMHO, it is in the hands of "Sun" not the "OpenSolaris Community".
> >
> > Perhaps.  However, since the only thing in hands of your kind of 
> > advocates is best not mentioned on a family-friendly maillist, may I 
> > suggest taking that exciting thread to more appropriate place?
>
> I don't think that this thread is going unfriendly or harmfully. 
> However, what is your suggestion ?

if you want to change the minds of the OpenSolaris community, i'd 
proffer that it's perhaps more efficient to talk to them, not to the 
linux-kernel mailing list. Thanks,

	Ingo

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Tarkan Erimer]

Ingo Molnar wrote:
> if you want to change the minds of the OpenSolaris community, i'd 
> proffer that it's perhaps more efficient to talk to them, not to the 
> linux-kernel mailing list. Thanks,
>
> 	Ingo
>   

I do not want to and try to change anyone's mind: nor the Open Solaris 
Community nor the Linux Community. Just, I asked simple question and 
included a simple example in it. Son, including an example related to 
OpenSolaris does not mean that I want to push OpenSolaris things. That's 
all.

Regards,

Tarkan

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Ingo Molnar]

* Tarkan Erimer <tarkan@netone.net.tr> wrote:

> [...] Just, I asked simple question and included a simple example in 
> it. [...]

actually, what you said was this:

" I hope we should upgrade to GPLv3 and Sun should "Dual License"
  the OpenSolaris via GPLv3 (or at least,GPLv3 should be CDDL
  compatible.). "

and to that the answer was:

" The OpenSolaris community has already stated that they do not want to
  accept GPLv3 [...] "

in other words: your hypothetical is false today. You called us to do a 
specific action, but why did you then include a factually false 
'example' to underline that point of yours? Or if you simply did not 
know about the OpenSolaris community's position beforehand, why dont you 
just admit that and withdraw from that line of argument gracefully?

	Ingo

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Tarkan Erimer]

Ingo Molnar wrote:
> * Tarkan Erimer <tarkan@netone.net.tr> wrote:
>
>   
>> [...] Just, I asked simple question and included a simple example in 
>> it. [...]
>>     
>
> actually, what you said was this:
>
> " I hope we should upgrade to GPLv3 and Sun should "Dual License"
>   the OpenSolaris via GPLv3 (or at least,GPLv3 should be CDDL
>   compatible.). "
>   
Why don't you include the last sentence I wrote: "So,we should have more 
fruits (like ZFS,DTrace etc.) ;-) "
So, that's why I said it. Because, as all the time, we did it: Importing 
and exporting codes to/from different open source projects.

> and to that the answer was:
>
> " The OpenSolaris community has already stated that they do not want to
>   accept GPLv3 [...] "
>
> in other words: your hypothetical is false today. You called us to do a 
> specific action, but why did you then include a factually false 
> 'example' to underline that point of yours? Or if you simply did not 
> know about the OpenSolaris community's position beforehand, why dont you 
> just admit that and withdraw from that line of argument gracefully?
>
> 	Ingo
>   

As I mentioned in my previous posts: This is **not** in the hands of the 
"OpenSolaris Community" to make and apply such decision. Sun itself 
**will decide** it. Also, there are strong indications that Sun is very 
interested to make "OpenSolaris" at least "Dual-Licensed" with GPLv3.

Regards,

Tarkan

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Ingo Molnar]

* Tarkan Erimer <tarkan@netone.net.tr> wrote:

> Ingo Molnar wrote:
> >* Tarkan Erimer <tarkan@netone.net.tr> wrote:
> >
> >  
> >>[...] Just, I asked simple question and included a simple example in 
> >>it. [...]
> >>    
> >
> >actually, what you said was this:
> >
> >" I hope we should upgrade to GPLv3 and Sun should "Dual License"
> >  the OpenSolaris via GPLv3 (or at least,GPLv3 should be CDDL
> >  compatible.). "
> >  
> Why don't you include the last sentence I wrote: "So,we should have more 
> fruits (like ZFS,DTrace etc.) ;-) "

You might as well have said "the moon is made out of cheese" and i'd not 
have quoted it either. Why? Because it's irrelevant to the fundamental 
point that was raised and which you keep ignoring: that the only 
"example" you cited is a hypothetical that is currently false. In any 
case, speculation about what Sun might or might not do, up until the 
point it actually does it, is not something i feel compelled to do 
anything over, so please stop wasting my time by Cc:-ing me. Thanks,

	Ingo

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Tarkan Erimer]

Ingo Molnar wrote:
> You might as well have said "the moon is made out of cheese" and i'd not 
> have quoted it either. Why? Because it's irrelevant to the fundamental 
> point that was raised and which you keep ignoring: that the only 
> "example" you cited is a hypothetical that is currently false. In any 
> case, speculation about what Sun might or might not do, up until the 
> point it actually does it, is not something i feel compelled to do 
> anything over, so please stop wasting my time by Cc:-ing me. Thanks,
>
> 	Ingo
>   

I think, you do not want to understand what I really  mean. OK,I 
stopping here. Because, you already wasted a lot of my time via always 
not understanding what I really mean.

Regards,

Tarkan

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Sun, 10 Jun 2007, Tarkan Erimer wrote:
> > 
> > Last heard, Linus was quite impressed with the toned down version of
> > the final draft of GPLv3.

I was impressed in the sense that it was a hell of a lot better than the 
disaster that were the earlier drafts.

I still think GPLv2 is simply the better license.

I consider dual-licensing unlikely (and technically quite hard), but at 
least _possible_ in theory. I have yet to see any actual *reasons* for 
licensing under the GPLv3, though. All I've heard are shrill voices about 
"tivoization" (which I expressly think is ok) and panicked worries about 
Novell-MS (which seems way overblown, and quite frankly, the argument 
seems to not so much be about the Novell deal, as about an excuse to push 
the GPLv3).

		Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Sun, 10 Jun 2007, Linus Torvalds wrote:
>
> I have yet to see any actual *reasons* for licensing under the GPLv3, 
> though.

Btw, if Sun really _is_ going to release OpenSolaris under GPLv3, that 
_may_ be a good reason. I don't think the GPLv3 is as good a license as 
v2, but on the other hand, I'm pragmatic, and if we can avoid having two 
kernels with two different licenses and the friction that causes, I at 
least see the _reason_ for GPLv3. As it is, I don't really see a reason at 
all.

I personally doubt it will happen, but hey, I didn't really expect them to 
open-source Java either(*), so it's not like I'm infallible in my 
predictions.

				Linus

(*) And I've been pushing for that since before they even released it - I 
walked out on Bill Joy at a private event where they discussed their 
horrible previous Java license.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Tarkan Erimer]

Linus Torvalds wrote:
> On Sun, 10 Jun 2007, Linus Torvalds wrote:
>   
>> I have yet to see any actual *reasons* for licensing under the GPLv3, 
>> though.
>>     
>
> Btw, if Sun really _is_ going to release OpenSolaris under GPLv3, that 
> _may_ be a good reason. I don't think the GPLv3 is as good a license as 
> v2, but on the other hand, I'm pragmatic, and if we can avoid having two 
> kernels with two different licenses and the friction that causes, I at 
> least see the _reason_ for GPLv3. As it is, I don't really see a reason at 
> all.
>
> I personally doubt it will happen, but hey, I didn't really expect them to 
> open-source Java either(*), so it's not like I'm infallible in my 
> predictions.
>
> 				Linus
>
> (*) And I've been pushing for that since before they even released it - I 
> walked out on Bill Joy at a private event where they discussed their 
> horrible previous Java license.
>   

Thanks for making things more clear :-) Some really strong indications 
that Sun is very willing to,at least, "Dual-License" the OpenSolaris 
with GPLv3. I think; in a very short time; we will see when the GPLv3 
finalized and released.

Regards,

Tarkan

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Ingo Molnar]

* Tarkan Erimer <tarkan@netone.net.tr> wrote:

> > (*) And I've been pushing for that since before they even released 
> > it - I walked out on Bill Joy at a private event where they 
> > discussed their horrible previous Java license.
> 
> Thanks for making things more clear :-) Some really strong indications 
> that Sun is very willing to,at least, "Dual-License" the OpenSolaris 
> with GPLv3. I think; in a very short time; we will see when the GPLv3 
> finalized and released.

that would certainly be a good and productive move from them. Note the 
issue that others have pointed out to you: OpenSolaris is probably more 
interested in picking up code from Linux than the other way around! :-) 
You mentioned "dtrace" and "ZFS". Firstly, Linux already has a "dtrace" 
equivalent. Secondly, ZFS might be interesting in theory, although our 
prior experience of having compatibly-licensed filesystems ported over 
to Linux has been pretty negative: XFS ended up being an integration 
nightmare - and that doesnt have to do anything with the qualities of 
XFS (it's one of the cleanest Linux filesystems, if not the cleanest), 
the problem is that components within a kernel are very tightly 
integrated and rarely does it make sense to port over more than just 
drivers or maybe libraries. And that's i guess what OpenSolaris lacks 
and which i suspect it is mostly interested in: lots of nice Linux 
drivers ;-) XFS, the largest Linux filesystem is 100K lines of code - 
and ZFS (i've never seen it) is very likely smaller than that. Linux 
drivers on the other hand, as of today, are _3.7 million_ lines of code 
and enable Linux to run on 99% of the hardware that is produced today. 
Guess which one has the larger strategic significance? ;-)

	Ingo

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bron Gondwana]

On Mon, Jun 11, 2007 at 11:03:48AM +0200, Ingo Molnar wrote:
> 
> * Tarkan Erimer <tarkan@netone.net.tr> wrote:
> 
> > > (*) And I've been pushing for that since before they even released 
> > > it - I walked out on Bill Joy at a private event where they 
> > > discussed their horrible previous Java license.
> > 
> > Thanks for making things more clear :-) Some really strong indications 
> > that Sun is very willing to,at least, "Dual-License" the OpenSolaris 
> > with GPLv3. I think; in a very short time; we will see when the GPLv3 
> > finalized and released.
> 
> that would certainly be a good and productive move from them. Note the 
> issue that others have pointed out to you: OpenSolaris is probably more 
> interested in picking up code from Linux than the other way around! :-) 
> You mentioned "dtrace" and "ZFS". Firstly, Linux already has a "dtrace" 
> equivalent. Secondly, ZFS might be interesting in theory
                        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

Having just got my first Sun box (one of those 48 disk Thumper machines)
I can say it's very interesting in practice too.  The admin tools for
zfs are a dream to use (unlike just about everything else on Solaris
which is a bugwards compatible nightmare to use) and the attitude of
checksumming everything on to disk and checking that those checksums
match on the way back out presses my "do your own safety checking and
don't trust the hardware" buttons very much the right way - especially
with that much hardware in there.

I would love to see Sun GPL3 OpenSolaris, not so much for the code
itself (maybe not portable into Linux) but for the clarity it would
give to the patent position.  The patent peace would be portable back
to GPL2.

Bron.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Tarkan Erimer]

Ingo Molnar wrote:
> * Tarkan Erimer <tarkan@netone.net.tr> wrote:
>
>   
>>> (*) And I've been pushing for that since before they even released 
>>> it - I walked out on Bill Joy at a private event where they 
>>> discussed their horrible previous Java license.
>>>       
>> Thanks for making things more clear :-) Some really strong indications 
>> that Sun is very willing to,at least, "Dual-License" the OpenSolaris 
>> with GPLv3. I think; in a very short time; we will see when the GPLv3 
>> finalized and released.
>>     
>
> that would certainly be a good and productive move from them. Note the 
> issue that others have pointed out to you: OpenSolaris is probably more 
> interested in picking up code from Linux than the other way around! :-) 
>   
Totally agreed :-)
> You mentioned "dtrace" and "ZFS". Firstly, Linux already has a "dtrace" 
> equivalent. Secondly, ZFS might be interesting in theory, although our 
> prior experience of having compatibly-licensed filesystems ported over 
> to Linux has been pretty negative: XFS ended up being an integration 
> nightmare - and that doesnt have to do anything with the qualities of 
> XFS (it's one of the cleanest Linux filesystems, if not the cleanest), 
> the problem is that components within a kernel are very tightly 
> integrated and rarely does it make sense to port over more than just 
> drivers or maybe libraries. And that's i guess what OpenSolaris lacks 
> and which i suspect it is mostly interested in: lots of nice Linux 
> drivers ;-) XFS, the largest Linux filesystem is 100K lines of code - 
> and ZFS (i've never seen it) is very likely smaller than that. Linux 
> drivers on the other hand, as of today, are _3.7 million_ lines of code 
> and enable Linux to run on 99% of the hardware that is produced today. 
> Guess which one has the larger strategic significance? ;-)
>
> 	Ingo
>   
Yep, it is clear that sun needs more things like drivers etc. to make 
OpenSolaris more usable and user friendly. Here is an article about this 
subject and some thoughts of Ian (Murdock) about it ;-)

http://www.zdnet.com.au/news/software/soa/Sun-hopes-for-Linux-like-Solaris/0,130061733,339276057,00.htm

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 11, 2007, Ingo Molnar <mingo@elte.hu> wrote:

> And that's i guess what OpenSolaris lacks and which i suspect it is
> mostly interested in: lots of nice Linux drivers ;-) XFS, the
> largest Linux filesystem is 100K lines of code - and ZFS (i've never
> seen it) is very likely smaller than that. Linux drivers on the
> other hand, as of today, are _3.7 million_ lines of code and enable
> Linux to run on 99% of the hardware that is produced today.  Guess
> which one has the larger strategic significance? ;-)

Per this reasoning, Sun wouldn't be waiting for GPLv3, and it would
have already released the OpenSolaris kernel under GPLv2, would it
not? ;-)

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Tue, 12 Jun 2007, Alexandre Oliva wrote:
> 
> Per this reasoning, Sun wouldn't be waiting for GPLv3, and it would
> have already released the OpenSolaris kernel under GPLv2, would it
> not? ;-)

Umm. You are making the fundamental mistake of thinking that Sun is in 
this to actually further some open-source agenda.

Here's a cynical prediction (but backed up by past behaviour of Sun):

 - first off: they may be talking a lot more than they are or ever will
   be doing. How many announcements about Sun and Linux have you seen over
   the years? And how much of that has actually happened?

 - They may like open source, but Linux _has_ hurt them in the 
   marketplace. A lot.

   They almost used to own the chip design market, and it took quite a 
   long time before the big EDA vendors ported to Linux (and x86-64 in 
   particular). But when they did, their chip design market just basically 
   disappeared: sparc performance is so horribly bad (especially on a 
   workstation kind of setup), that to do chip design on them is just 
   idiotic. Which is not to say that there aren't holdouts, but let's face 
   it, for a lot of things, Solaris is simply the wrong choice these days.

   Ergo: they sure as hell don't want to help Linux. Which is fine. 
   Competition is good.

 - So they want to use Linux resources (_especially_ drivers), but they do 
   *not* want to give anything back (especially ZFS, which seems to be one 
   of their very very few bright spots).

 - Ergo: they'll not be releasing ZFS and the other things that people are 
   drooling about in a way that lets Linux use them on an equal footing. I 
   can pretty much guarantee that. They don't like competition on that 
   level. They'd *much* rather take our drivers and _not_ give anythign 
   back, or give back the stuff that doesn't matter (like core Solaris: 
   who are you kidding - Linux code is _better_).

End result:

 - they'll talk about it. They not only drool after our drivers, they 
   drool after all the _people_ who write drivers. They'd love to get 
   kernel developers from Linux, they see that we have a huge amount of 
   really talented people. So they want to talk things up, and the more 
   "open source" they can position themselves, the better.

 - They may release the uninteresting parts under some fine license. See 
   the OpenSolaris stuff - instead of being blinded by the code they _did_ 
   release under an open source license, ask yourself what they did *not* 
   end up releasing. Ask yourself why the open source parts are not ready 
   to bootstrap a competitive system, or why they are released under 
   licenses that Sun can make sure they control.

So the _last_ thing they want to do is to release the interesting stuff 
under GPLv2 (quite frankly, I think the only really interesting thing they 
have is ZFS, and even there, I suspect we'd be better off talking to 
NetApp, and seeing if they are interested in releasing WAFL for Linux).

Yes, they finally released Java under GPLv2, and they should be commended 
for that. But you should also ask yourself why, and why it took so long. 
Maybe it had something to do with the fact that other Java implementations 
started being more and more relevant?

Am I cynical? Yes. Do I expect people to act in their own interests? Hell 
yes! That's how things are _supposed_ to happen. I'm not at all berating 
Sun, what I'm trying to do here is to wake people up who seem to be living 
in some dream-world where Sun wants to help people. 

So to Sun, a GPLv3-only release would actually let them look good, and 
still keep Linux from taking their interesting parts, and would allow them 
to take at least parts of Linux without giving anything back (ahh, the 
joys of license fragmentation). 

Of course, they know that. And yes, maybe ZFS is worthwhile enough that 
I'm willing to go to the effort of trying to relicense the kernel. But 
quite frankly, I can almost guarantee that Sun won't release ZFS under the 
GPLv3 even if they release other parts. Because if they did, they'd lose 
the patent protection.

And yes, I'm cynical, and yes, I hope I'm wrong. And if I'm wrong, I'll 
very happily retract anything cynical I said about Sun. They _have_ done 
great things, and maybe I'm just too pessimistic about all the history 
I've seen of Sun with open source.

The _good_ news is that Jonathan Schwartz actually does seem to have made 
a difference, and I hope to God he is really as serious about 
open-sourcing things as he says he is. And don't get me wrong: I think a 
truly open-source GPLv3 Solaris would be a really really _good_ thing, 
even if it does end up being a one-way street as far as code is concerned!

			Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 12, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:

> On Tue, 12 Jun 2007, Alexandre Oliva wrote:
>> 
>> Per this reasoning, Sun wouldn't be waiting for GPLv3, and it would
>> have already released the OpenSolaris kernel under GPLv2, would it
>> not? ;-)

> Umm. You are making the fundamental mistake of thinking that Sun is in 
> this to actually further some open-source agenda.

Err, no.  I was merely questioning Ingo's reasoning that Sun wanted
Linux's drivers as badly as he made it seem.  All the fuss about
waiting for and going to GPLv3 wouldn't get them that.  Moving to
GPLv2 would, and still, they aren't doing it.  That was my point.

FWIW, I share most of your assessment and wait-and-see attitude about
Sun's situation.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Matthias Kaehlcke]

El Tue, Jun 12, 2007 at 08:45:46AM -0700 Linus Torvalds ha dit:

> 
> 
> On Tue, 12 Jun 2007, Alexandre Oliva wrote:
> > 
> > Per this reasoning, Sun wouldn't be waiting for GPLv3, and it would
> > have already released the OpenSolaris kernel under GPLv2, would it
> > not? ;-)
> 
> Umm. You are making the fundamental mistake of thinking that Sun is in 
> this to actually further some open-source agenda.
> 
> Here's a cynical prediction (but backed up by past behaviour of Sun):
> 
>  - first off: they may be talking a lot more than they are or ever will
>    be doing. How many announcements about Sun and Linux have you seen over
>    the years? And how much of that has actually happened?
> 
>  - They may like open source, but Linux _has_ hurt them in the 
>    marketplace. A lot.
> 
>    They almost used to own the chip design market, and it took quite a 
>    long time before the big EDA vendors ported to Linux (and x86-64 in 
>    particular). But when they did, their chip design market just basically 
>    disappeared: sparc performance is so horribly bad (especially on a 
>    workstation kind of setup), that to do chip design on them is just 
>    idiotic. Which is not to say that there aren't holdouts, but let's face 
>    it, for a lot of things, Solaris is simply the wrong choice these days.
> 
>    Ergo: they sure as hell don't want to help Linux. Which is fine. 
		 >    Competition is good.
> 
>  - So they want to use Linux resources (_especially_ drivers), but they do 
>    *not* want to give anything back (especially ZFS, which seems to be one 
>    of their very very few bright spots).
> 
>  - Ergo: they'll not be releasing ZFS and the other things that people are 
>    drooling about in a way that lets Linux use them on an equal footing. I 
>    can pretty much guarantee that. They don't like competition on that 
>    level. They'd *much* rather take our drivers and _not_ give anythign 
>    back, or give back the stuff that doesn't matter (like core Solaris: 
>    who are you kidding - Linux code is _better_).
> 
> End result:
> 
>  - they'll talk about it. They not only drool after our drivers, they 
>    drool after all the _people_ who write drivers. They'd love to get 
>    kernel developers from Linux, they see that we have a huge amount of 
>    really talented people. So they want to talk things up, and the more 
>    "open source" they can position themselves, the better.
> 
>  - They may release the uninteresting parts under some fine license. See 
>    the OpenSolaris stuff - instead of being blinded by the code they _did_ 
>    release under an open source license, ask yourself what they did *not* 
>    end up releasing. Ask yourself why the open source parts are not ready 
>    to bootstrap a competitive system, or why they are released under 
>    licenses that Sun can make sure they control.
> 
> So the _last_ thing they want to do is to release the interesting stuff 
> under GPLv2 (quite frankly, I think the only really interesting thing they 
> have is ZFS, and even there, I suspect we'd be better off talking to 
> NetApp, and seeing if they are interested in releasing WAFL for Linux).
> 
> Yes, they finally released Java under GPLv2, and they should be commended 
> for that. But you should also ask yourself why, and why it took so long. 
> Maybe it had something to do with the fact that other Java implementations 
> started being more and more relevant?
> 
> Am I cynical? Yes. Do I expect people to act in their own interests? Hell 
> yes! That's how things are _supposed_ to happen. I'm not at all berating 
> Sun, what I'm trying to do here is to wake people up who seem to be living 
> in some dream-world where Sun wants to help people. 
> 
> So to Sun, a GPLv3-only release would actually let them look good, and 
> still keep Linux from taking their interesting parts, and would allow them 
> to take at least parts of Linux without giving anything back (ahh, the 
> joys of license fragmentation). 
> 
> Of course, they know that. And yes, maybe ZFS is worthwhile enough that 
> I'm willing to go to the effort of trying to relicense the kernel. But 
> quite frankly, I can almost guarantee that Sun won't release ZFS under the 
> GPLv3 even if they release other parts. Because if they did, they'd lose 
> the patent protection.
> 
> And yes, I'm cynical, and yes, I hope I'm wrong. And if I'm wrong, I'll 
> very happily retract anything cynical I said about Sun. They _have_ done 
> great things, and maybe I'm just too pessimistic about all the history 
> I've seen of Sun with open source.
> 
> The _good_ news is that Jonathan Schwartz actually does seem to have made 
> a difference, and I hope to God he is really as serious about 
> open-sourcing things as he says he is. And don't get me wrong: I think a 
> truly open-source GPLv3 Solaris would be a really really _good_ thing, 
> even if it does end up being a one-way street as far as code is
> concerned!

FYI, Jonathan Schwartz' response:

Linus,

First, I'm glad you give credit to Sun for the contributions we've
made to the open source world, and Linux specifically - we take the
commitment seriously. It's why we freed OpenOffice, elements of Gnome,
Mozilla, delivered Java, and a long list of other contributions that
show up in almost every distro. Individuals will always define
communities, but Sun as a company has done its part to grow the market
- for others as much as ourselves.

But I disagree with a few of your points. Did the Linux community hurt
Sun? No, not a bit. It was the companies that leveraged their work. I
draw a very sharp distinction - even if our competition is
conveniently reckless. They like to paint the battle as Sun vs. the
community, and it's not. Companies compete, communities simply
fracture.

And OpenSolaris has come a very long way since you last looked. It and
its community are growing, as a result of more than ZFS (although we
seem to be generating a lot of interest there, not all intentional) -
OpenSolaris scales on any hardware, has built in virtualization, great
web service infrastucture, fault management, diagnosability, and tons
more. Feel free to try for yourself (and yes, we're fixing
installability, no fair knocking us for that.)

Now despite what you suggest, we love where the FSF's GPL3 is
headed. For a variety of mechanical reasons, GPL2 is harder for us
with OpenSolaris - but not impossible, or even out of the
question. This has nothing to do with being afraid of the community
(if it was, we wouldn't be so interested in seeing ZFS everywhere,
including Linux, with full patent indemnity). Why does open sourcing
take so long? Because we're starting from products that exist, in
which a diversity of contributors and licensors/licensees have rights
we have to negotiate. Indulge me when I say It's different than
starting from scratch. I would love to go faster, and we are all doing
everything under our control to accelerate progress. (Remember, we
can't even pick GPL3 yet - it doesn't officially exist.) It's also a
delicate dance to manage this transition while growing a corporation.

But most of all, from where I sit, we should put the swords down -
you're not the enemy for us, we're not the enemy for you. Most of the
world doesn't have access to the internet - that's the enemy to slay,
the divide that separates us. By joining our communities, we can bring
transparency and opportunity to the whole planet. Are we after your
drivers? No more than you're after ZFS or Crossbow or dtrace - it's
not predation, it's prudence. Let's stop wasting time recreating
wheels we both need to roll forward.

I wanted you to hear this from me directly. We want to work together,
we want to join hands and communities - we have no intention of
holding anything back, or pulling patent nonsense. And to prove the
sincerity of the offer, I invite you to my house for dinner. I'll
cook, you bring the wine. A mashup in the truest sense.

Best,
Jonathan

President, Chief Executive Officer,
Sun Microsystems, Inc. 

http://blogs.sun.com/jonathan/entry/one_plus_one_is_fifty

-- 
Matthias Kaehlcke
Linux Application Developer
Barcelona

    Yo soy como soy y tú eres como eres, construyamos un mundo donde yo
    pueda ser sin dejar de ser yo, donde tú puedas ser sin dejar de ser
    tú, y donde ni yo ni tú obliguemos al otro a ser como yo o como tú
                                                                 .''`.
    using free software / Debian GNU/Linux | http://debian.org  : :'  :
                                                                `. `'`
gpg --keyserver pgp.mit.edu --recv-keys 47D8E5D4                  `-

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Tarkan Erimer]

Matthias Kaehlcke wrote:
> FYI, Jonathan Schwartz' response:
>
> I wanted you to hear this from me directly. We want to work together,
> we want to join hands and communities - we have no intention of
> holding anything back, or pulling patent nonsense. And to prove the
> sincerity of the offer, I invite you to my house for dinner. I'll
> cook, you bring the wine. A mashup in the truest sense.
>
> Best,
> Jonathan
>
> President, Chief Executive Officer,
> Sun Microsystems, Inc. 
>
> http://blogs.sun.com/jonathan/entry/one_plus_one_is_fifty
>
>   

Really very very interesting! This words reminded me the same dialogues  
and affair, between "Linus" and "Steve Jobs", that have been happened 
several years ago :-)

Regards,

Tarkan

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Tarkan Erimer]

Hi Linus,

Linus Torvalds wrote:
> On Tue, 12 Jun 2007, Alexandre Oliva wrote:
>   
>> Per this reasoning, Sun wouldn't be waiting for GPLv3, and it would
>> have already released the OpenSolaris kernel under GPLv2, would it
>> not? ;-)
>>     
>
> Umm. You are making the fundamental mistake of thinking that Sun is in 
> this to actually further some open-source agenda.
>
> Here's a cynical prediction (but backed up by past behaviour of Sun):
>
>  - first off: they may be talking a lot more than they are or ever will
>    be doing. How many announcements about Sun and Linux have you seen over
>    the years? And how much of that has actually happened?
>
>  - They may like open source, but Linux _has_ hurt them in the 
>    marketplace. A lot.
>
>    They almost used to own the chip design market, and it took quite a 
>    long time before the big EDA vendors ported to Linux (and x86-64 in 
>    particular). But when they did, their chip design market just basically 
>    disappeared: sparc performance is so horribly bad (especially on a 
>    workstation kind of setup), that to do chip design on them is just 
>    idiotic. Which is not to say that there aren't holdouts, but let's face 
>    it, for a lot of things, Solaris is simply the wrong choice these days.
>
>    Ergo: they sure as hell don't want to help Linux. Which is fine. 
>    Competition is good.
>
>  - So they want to use Linux resources (_especially_ drivers), but they do 
>    *not* want to give anything back (especially ZFS, which seems to be one 
>    of their very very few bright spots).
>
>  - Ergo: they'll not be releasing ZFS and the other things that people are 
>    drooling about in a way that lets Linux use them on an equal footing. I 
>    can pretty much guarantee that. They don't like competition on that 
>    level. They'd *much* rather take our drivers and _not_ give anythign 
>    back, or give back the stuff that doesn't matter (like core Solaris: 
>    who are you kidding - Linux code is _better_).
>
>   
Completely agreed :-)

> End result:
>
>  - they'll talk about it. They not only drool after our drivers, they 
>    drool after all the _people_ who write drivers. They'd love to get 
>    kernel developers from Linux, they see that we have a huge amount of 
>    really talented people. So they want to talk things up, and the more 
>    "open source" they can position themselves, the better.
>
>   
Definitely. They already began to pull some people like Ian Murdock. And 
I'm really very disappointed of this move,Ian did. Especially, such a 
person who has very good reputation and high profile in the Linux 
Community. He immediately shut down his company (also leaved 
Linux-Foundation) and joined to sun. After joining, he made statements 
like "How to make Solaris more like Linux ?" etc. Like a 40 years 
employee at Sun. Another interesting thing is the timing of this hiring. 
So, this situation is a good example of it.

>  - They may release the uninteresting parts under some fine license. See 
>    the OpenSolaris stuff - instead of being blinded by the code they _did_ 
>    release under an open source license, ask yourself what they did *not* 
>    end up releasing. Ask yourself why the open source parts are not ready 
>    to bootstrap a competitive system, or why they are released under 
>    licenses that Sun can make sure they control.
>
> So the _last_ thing they want to do is to release the interesting stuff 
> under GPLv2 (quite frankly, I think the only really interesting thing they 
> have is ZFS, and even there, I suspect we'd be better off talking to 
> NetApp, and seeing if they are interested in releasing WAFL for Linux).
>
> Yes, they finally released Java under GPLv2, and they should be commended 
> for that. But you should also ask yourself why, and why it took so long. 
> Maybe it had something to do with the fact that other Java implementations 
> started being more and more relevant?
>
> Am I cynical? Yes. Do I expect people to act in their own interests? Hell 
> yes! That's how things are _supposed_ to happen. I'm not at all berating 
> Sun, what I'm trying to do here is to wake people up who seem to be living 
> in some dream-world where Sun wants to help people. 
>
> So to Sun, a GPLv3-only release would actually let them look good, and 
> still keep Linux from taking their interesting parts, and would allow them 
> to take at least parts of Linux without giving anything back (ahh, the 
> joys of license fragmentation). 
>
> Of course, they know that. And yes, maybe ZFS is worthwhile enough that 
> I'm willing to go to the effort of trying to relicense the kernel. But 
> quite frankly, I can almost guarantee that Sun won't release ZFS under the 
> GPLv3 even if they release other parts. Because if they did, they'd lose 
> the patent protection.
>
> And yes, I'm cynical, and yes, I hope I'm wrong. And if I'm wrong, I'll 
> very happily retract anything cynical I said about Sun. They _have_ done 
> great things, and maybe I'm just too pessimistic about all the history 
> I've seen of Sun with open source.
>
> The _good_ news is that Jonathan Schwartz actually does seem to have made 
> a difference, and I hope to God he is really as serious about 
> open-sourcing things as he says he is. And don't get me wrong: I think a 
> truly open-source GPLv3 Solaris would be a really really _good_ thing, 
> even if it does end up being a one-way street as far as code is concerned!
>
> 			Linus
>   
One more time,agreed ;-)

Regards,

Tarkan

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alan Cox]

> licensing under the GPLv3, though. All I've heard are shrill voices about 
> "tivoization" (which I expressly think is ok) and panicked worries about 

GPLv2 probably forbids Tivoisation anyway. Which is good IMHO even if not
yours 8)

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[debian developer]

On 6/10/07, Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:
> > licensing under the GPLv3, though. All I've heard are shrill voices about
> > "tivoization" (which I expressly think is ok) and panicked worries about
>
> GPLv2 probably forbids Tivoisation anyway. Which is good IMHO even if not
              ^^^^^^^^

Now that is a bit waving in the air. GPLv2 forbids Tivoisation
theoretically but practically it didnt stop them doing it practically.
I agree with Linus that software licenses should have their influence
only on the software part and leave the freedom of the hardware on
which the software runs to the hardware manufacturers.

But was it the goal of GPLv2??

And what does Andrew Morton think of all this? I really want to know
his opinions....

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Andrew Morton]

On Mon, 11 Jun 2007 01:02:42 +0530 "debian developer" <debiandev@gmail.com> wrote:

> And what does Andrew Morton think of all this? I really want to know
> his opinions....

I have yet to see Linus make a statement on these matters with which
I didn't agree.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alan Cox]

> > GPLv2 probably forbids Tivoisation anyway. Which is good IMHO even if not
>               ^^^^^^^^
> 
> Now that is a bit waving in the air. GPLv2 forbids Tivoisation
> theoretically but practically it didnt stop them doing it practically.

They've never been given permission and there is no caselaw yet, doesn't
mean they are allowed to.

GPL2 actually in some ways was saner than GPL3 on this - you could
sensibly argue the key was part of the source/build environment but it
didn't then get muddled in with questions like ROMs which the new GPL3
wording is a bit messy about still.

Alan

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Greg KH]

On Sun, Jun 10, 2007 at 09:54:58PM +0100, Alan Cox wrote:
> > > GPLv2 probably forbids Tivoisation anyway. Which is good IMHO even if not
> >               ^^^^^^^^
> > 
> > Now that is a bit waving in the air. GPLv2 forbids Tivoisation
> > theoretically but practically it didnt stop them doing it practically.
> 
> They've never been given permission and there is no caselaw yet, doesn't
> mean they are allowed to.

Are you sure?  Tivo went and got a FSF "verification" of their system a
number of years ago and got their blessing that what they were doing was
just fine with regards to the GPLv2.

This is one reason Tivo's lawyers were so perplexed when the FSF then
turned around and made their company's name into a term to describe DRM
stuff and started preaching how it was so bad.  It seemed to be in
direct crontridiction from what they had previously been told by the
very same people.

Now yes, they didn't consult with the individual owners of the kernel,
who might hold different views as to if v2 covers keys like you have
stated in the past, but the FSF's position in this area does hold some
ammount of weight, especially in court if it were to come to that.

thanks,

greg k-h

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alan Cox]

> Now yes, they didn't consult with the individual owners of the kernel,
> who might hold different views as to if v2 covers keys like you have
> stated in the past, but the FSF's position in this area does hold some
> ammount of weight, especially in court if it were to come to that.

The authors position does have rather a lot of weight too. Especially as
they have been made plain to Tivo and various other relevant parties.

Alan

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Rob Landley]

On Sunday 10 June 2007 15:32:42 debian developer wrote:
> On 6/10/07, Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:
> > > licensing under the GPLv3, though. All I've heard are shrill voices
> > > about "tivoization" (which I expressly think is ok) and panicked
> > > worries about
> >
> > GPLv2 probably forbids Tivoisation anyway. Which is good IMHO even if not
>
>               ^^^^^^^^
>
> Now that is a bit waving in the air. GPLv2 forbids Tivoisation
> theoretically but practically it didnt stop them doing it practically.

A law never stops anybody from doing anything.  Enforcing the law does.

First of all, let's not confuse civil with criminal law:

http://nielsenhayden.com/makinglight/archives/009034.html
>  The difference between tort—breach of private rights—and crime—commission
> of an offence designated as such by the State—is one of the key legal
> concepts which the pro lawyer understands and the fan lawyer does not

Most variants of copyright infringement are a civil, not criminal matter.  
This means the state has no interest in enforcing the, it's your job to sue 
for damages and a restraining order if you want to exercise these rights 
(some people choose not to), and you have to have standing (I.E. be a holder 
of one of the infringed copyrights, or a designated legal representative 
thereof) in order to sue.  If none of the copyright holders sue to stop it, 
then it doesn't get stopped no matter how blatantly infringing it is.  Did 
anybody even bother to send Tivo a cease and desist?

Erik Andersen burned himself out trying to enforce the copyrights on BusyBox 
before Pamela Jones hooked that project (and uClibc) up with the SFLC.  
Harald Welte's been burning the candle at both ends with gpl-violations.org, 
but he's focusing on stuff sold in Germany.

As for anti-tivoisation, you can make a case that your signed binary is a 
derived work of the GPL source code just like a non-signed binary is, 
therefore the signing key is part of the source code used to create that 
binary, therefore GPLv2 requires the signing key be handed over on request.  
(I don't know if you'd WIN, I just know you could reasonably argue it in 
court and probably get past the inevitable initial motions to dismiss.)

GPLv2 is a nice, elegant license.  It's not perfect but it's very simple for 
what it does.

GPLv3 is not simple, not elegant, and contains numerous of special cases.  
Lots of the programmers here have an instinctive aversion to it because it 
reads like bad code.  We don't necessarily have to program in legalese to 
sense bad code in that language, at least compared to a "good code" example 
we've been using for some time.

Rob

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Jeff Garzik]

Linus Torvalds wrote:
> I still think GPLv2 is simply the better license.

Ditto.

	Jeff

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[James Bruce]

Linus Torvalds wrote:
[ snip ]
> I consider dual-licensing unlikely (and technically quite hard), but at 
> least _possible_ in theory. I have yet to see any actual *reasons* for 
> licensing under the GPLv3, though.
[ snip ]

One thing that would make that easier in the future is if contributers
at least started to dual-license their submissions.  I.e. if instead
of "GPL version 2", one could say "GPL version 2 or GPL version 3".
It isn't the same thing as the problematic "GPL version 2 or later",
because the developer is not agreeing to an unseen license (GPLv4,
etc).  What it does do is make it easier to move to GPLv3 a few years
from now, if that is decided then, as a significant fraction of the
code will already be GPLv3 compatible.  That way, if a reason is ever
found to move to v3, at least some of the work will already be done.

  - Jim Bruce

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Jesper Juhl]

On 10/06/07, James Bruce <bruce@andrew.cmu.edu> wrote:
> Linus Torvalds wrote:
> [ snip ]
> > I consider dual-licensing unlikely (and technically quite hard), but at
> > least _possible_ in theory. I have yet to see any actual *reasons* for
> > licensing under the GPLv3, though.
> [ snip ]
>
> One thing that would make that easier in the future is if contributers
> at least started to dual-license their submissions.  I.e. if instead
> of "GPL version 2", one could say "GPL version 2 or GPL version 3".
> It isn't the same thing as the problematic "GPL version 2 or later",
> because the developer is not agreeing to an unseen license (GPLv4,
> etc).  What it does do is make it easier to move to GPLv3 a few years
> from now, if that is decided then, as a significant fraction of the
> code will already be GPLv3 compatible.  That way, if a reason is ever
> found to move to v3, at least some of the work will already be done.
>
Good luck convincing all contributors to do that.

Personally I'm happy with GPL v2, and I for one won't be
dual-licensing anything I contribute until I see a clear benefit of
doing so (and I don't yet).

In any case, this whole debate is still a bit premature since GPL v3
has not even arrived in its final form yet.

-- 
Jesper Juhl <jesper.juhl@gmail.com>
Don't top-post  http://www.catb.org/~esr/jargon/html/T/top-post.html
Plain text mails only, please      http://www.expita.com/nomime.html

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[James Bruce]

Jesper Juhl wrote:
>> One thing that would make that easier in the future is if contributers
>> at least started to dual-license their submissions.  I.e. if instead
>> of "GPL version 2", one could say "GPL version 2 or GPL version 3".
>> It isn't the same thing as the problematic "GPL version 2 or later",
>> because the developer is not agreeing to an unseen license (GPLv4,
>> etc).  What it does do is make it easier to move to GPLv3 a few years
>> from now, if that is decided then, as a significant fraction of the
>> code will already be GPLv3 compatible.  That way, if a reason is ever
>> found to move to v3, at least some of the work will already be done.
>>
> Good luck convincing all contributors to do that.

Well, it's something that pro-GPLv3 people can do right now, instead of 
just lobbying/complaining.  Given 1000 developers, if 400 start dual 
licensing now, and down the road some compelling reason for GPLv3 does 
arise (read: a lawsuit with teeth), that's 600 people you need to 
contact/convince to change, not 1000.  This is made more interesting by 
that fact that 40% of the kernel code is already "GPLv2 or later", as 
someone else pointed out.

> Personally I'm happy with GPL v2, and I for one won't be
> dual-licensing anything I contribute until I see a clear benefit of
> doing so (and I don't yet).

Well, all my personal (non-kernel) stuff is still GPLv2 only right now 
(Linus' opinion is what convinced me that "or later" is dumb), and like 
many I disliked the original GPLv3 draft.  I'm willing to wait until the 
final one is out though, and I think my libraries will end up being 
dual-licensed, with contributions required to be dual-licensed.  I want 
to avoid v3 lock-in, but I don't want to cripple v3 projects either.

> In any case, this whole debate is still a bit premature since GPL v3
> has not even arrived in its final form yet.

Agreed.

  - Jim Bruce

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Florian Weimer]

* Linus Torvalds:

> I consider dual-licensing unlikely (and technically quite hard), but at 
> least _possible_ in theory. I have yet to see any actual *reasons* for 
> licensing under the GPLv3, though. All I've heard are shrill voices about 
> "tivoization" (which I expressly think is ok)

In a strange twist, the last-call draft contains a clause that
expressly permits some forms of "tivoization", provided a suitable
contractual arrangement exists ("Basic Permissions", second
paragraph).

Now a lot of the free software market follows this "sell yourself into
slavery" model (and even the FSF recommends to make money this way),
but I'm not sure if it's a good idea to state it so plainly in the
license.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[debian developer]

On 6/10/07, Tarkan Erimer <tarkan@netone.net.tr> wrote:
> debian developer wrote:
> > On 6/10/07, david@lang.hm <david@lang.hm> wrote:
> >> On Sun, 10 Jun 2007, Tarkan Erimer wrote:
> >>
> >> >> >  And maybe another questions should be : How long a copyright
> >> owner can
> >> >> >  hold the copyright, if died or lost for sometime ? if died, the
> >> >> >  copyright still should be valid or not ? If lost, what the law
> >> orders at
> >> >> >  this point for copyright holding ?
> >> >>
> >> >>  I believe that in the US it's life + 90 years.
> >> >>
> >> >>  David Lang
> >> > Hmm... Really,it is damn too much time to wait! It's really better
> >> idea to
> >> > replace the code of this person as said before instead of waiting
> >> such  90+
> >> > years!
> >>
> >> exactly, however as others are pointing out, there are a lot of active
> >> developers who do not agree with some of the key points of the GPLv3
> >> (including Linus), so until you convince them that the GPLv3 is
> >> better it
> >
> > Last heard, Linus was quite impressed with the toned down version of
> > the final draft of GPLv3. I think Linus, and other major developers
> > should make their stand on this issue clear so that the kernel
> > community can discuss the future steps.
> Yep, the GPLv3 probably will release around July time. So;luckily, we
> had very little time to see the final decision about it :-) I hope we
> should upgrade to GPLv3 and Sun should "Dual License" the OpenSolaris
> via GPLv3 (or at least,GPLv3 should be CDDL compatible.). So,we should
> have more fruits (like ZFS,DTrace etc.) ;-)
>
>
>
I don't think that upgrading to GPLv3 just for the sake of tools
present in some other software should be the reason. We are capable
enough of developing our own tools, and many experienced people are
working on equivalent(etx4 etc.,) and much sophisticated tools for the
linux kernel.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Al Viro]

On Mon, Jun 11, 2007 at 12:52:41AM +0530, debian developer wrote:

> I don't think that upgrading to GPLv3 just for the sake of tools
> present in some other software should be the reason. We are capable
> enough of developing our own tools, and many experienced people are
> working on equivalent(etx4 etc.,) and much sophisticated tools for the
> linux kernel.

I don't think that switch to GPLv3 can be described as upgrade.  I certainly
have no intention to do that to my code; some of it I might release under BSD
license, and that can be used in any project.  The rest of the kernel stuff
I've done (and that's the majority of my contributions) is under GPLv2 *only*.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Tarkan Erimer]

debian developer wrote:
> On 6/10/07, Tarkan Erimer <tarkan@netone.net.tr> wrote:
>> debian developer wrote:
>> > On 6/10/07, david@lang.hm <david@lang.hm> wrote:
>> >> On Sun, 10 Jun 2007, Tarkan Erimer wrote:
>> >>
>> >> >> >  And maybe another questions should be : How long a copyright
>> >> owner can
>> >> >> >  hold the copyright, if died or lost for sometime ? if died, the
>> >> >> >  copyright still should be valid or not ? If lost, what the law
>> >> orders at
>> >> >> >  this point for copyright holding ?
>> >> >>
>> >> >>  I believe that in the US it's life + 90 years.
>> >> >>
>> >> >>  David Lang
>> >> > Hmm... Really,it is damn too much time to wait! It's really better
>> >> idea to
>> >> > replace the code of this person as said before instead of waiting
>> >> such  90+
>> >> > years!
>> >>
>> >> exactly, however as others are pointing out, there are a lot of 
>> active
>> >> developers who do not agree with some of the key points of the GPLv3
>> >> (including Linus), so until you convince them that the GPLv3 is
>> >> better it
>> >
>> > Last heard, Linus was quite impressed with the toned down version of
>> > the final draft of GPLv3. I think Linus, and other major developers
>> > should make their stand on this issue clear so that the kernel
>> > community can discuss the future steps.
>> Yep, the GPLv3 probably will release around July time. So;luckily, we
>> had very little time to see the final decision about it :-) I hope we
>> should upgrade to GPLv3 and Sun should "Dual License" the OpenSolaris
>> via GPLv3 (or at least,GPLv3 should be CDDL compatible.). So,we should
>> have more fruits (like ZFS,DTrace etc.) ;-)
>>
>>
>>
> I don't think that upgrading to GPLv3 just for the sake of tools
> present in some other software should be the reason. We are capable
> enough of developing our own tools, and many experienced people are
> working on equivalent(etx4 etc.,) and much sophisticated tools for the
> linux kernel.
It is not because of the sake of the tools and we have no capable enough 
developers. It's just about an example that came to my mind, as I 
mentioned before and also,it is the same thing as we, all the time, did. 
I mean getting and sharing codes from many different open source 
projects like BSD and countless others. So, OpenSolaris makes no 
difference at this.

Regards,

Tarkan

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Greg KH]

On Sun, Jun 10, 2007 at 04:25:55PM +0530, debian developer wrote:
>  On 6/10/07, david@lang.hm <david@lang.hm> wrote:
> > On Sun, 10 Jun 2007, Tarkan Erimer wrote:
> >
> > >> >  And maybe another questions should be : How long a copyright owner 
> > can
> > >> >  hold the copyright, if died or lost for sometime ? if died, the
> > >> >  copyright still should be valid or not ? If lost, what the law orders 
> > at
> > >> >  this point for copyright holding ?
> > >>
> > >>  I believe that in the US it's life + 90 years.
> > >>
> > >>  David Lang
> > > Hmm... Really,it is damn too much time to wait! It's really better idea 
> > to
> > > replace the code of this person as said before instead of waiting such  
> > 90+
> > > years!
> >
> > exactly, however as others are pointing out, there are a lot of active
> > developers who do not agree with some of the key points of the GPLv3
> > (including Linus), so until you convince them that the GPLv3 is better it
> 
>  Last heard, Linus was quite impressed with the toned down version of
>  the final draft of GPLv3. I think Linus, and other major developers
>  should make their stand on this issue clear so that the kernel
>  community can discuss the future steps.

"future steps"?  Hah.

My code is going to stay GPLv2 as the v3 license is horrible for kernel
code for all of the reasons I have said in the past, plus a few more
(what, I can make an "industrial" product but not a commercial one?
That's horrible...)

thanks,

greg k-h

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[debian developer]

On 6/10/07, Greg KH <greg@kroah.com> wrote:
> On Sun, Jun 10, 2007 at 04:25:55PM +0530, debian developer wrote:
> >  On 6/10/07, david@lang.hm <david@lang.hm> wrote:
> > > On Sun, 10 Jun 2007, Tarkan Erimer wrote:
> > >
> > > >> >  And maybe another questions should be : How long a copyright owner
> > > can
> > > >> >  hold the copyright, if died or lost for sometime ? if died, the
> > > >> >  copyright still should be valid or not ? If lost, what the law orders
> > > at
> > > >> >  this point for copyright holding ?
> > > >>
> > > >>  I believe that in the US it's life + 90 years.
> > > >>
> > > >>  David Lang
> > > > Hmm... Really,it is damn too much time to wait! It's really better idea
> > > to
> > > > replace the code of this person as said before instead of waiting such
> > > 90+
> > > > years!
> > >
> > > exactly, however as others are pointing out, there are a lot of active
> > > developers who do not agree with some of the key points of the GPLv3
> > > (including Linus), so until you convince them that the GPLv3 is better it
> >
> >  Last heard, Linus was quite impressed with the toned down version of
> >  the final draft of GPLv3. I think Linus, and other major developers
> >  should make their stand on this issue clear so that the kernel
> >  community can discuss the future steps.
>
> "future steps"?  Hah.
>
> My code is going to stay GPLv2 as the v3 license is horrible for kernel
> code for all of the reasons I have said in the past, plus a few more
> (what, I can make an "industrial" product but not a commercial one?

     ^^^^^^^^^^^^^
What exactly in GPLv3 forbids you from making a commercial product?

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Greg KH]

On Tue, Jun 12, 2007 at 11:37:11PM +0530, debian developer wrote:
>  On 6/10/07, Greg KH <greg@kroah.com> wrote:
> > On Sun, Jun 10, 2007 at 04:25:55PM +0530, debian developer wrote:
> > >  On 6/10/07, david@lang.hm <david@lang.hm> wrote:
> > > > On Sun, 10 Jun 2007, Tarkan Erimer wrote:
> > > >
> > > > >> >  And maybe another questions should be : How long a copyright 
> > owner
> > > > can
> > > > >> >  hold the copyright, if died or lost for sometime ? if died, the
> > > > >> >  copyright still should be valid or not ? If lost, what the law 
> > orders
> > > > at
> > > > >> >  this point for copyright holding ?
> > > > >>
> > > > >>  I believe that in the US it's life + 90 years.
> > > > >>
> > > > >>  David Lang
> > > > > Hmm... Really,it is damn too much time to wait! It's really better 
> > idea
> > > > to
> > > > > replace the code of this person as said before instead of waiting 
> > such
> > > > 90+
> > > > > years!
> > > >
> > > > exactly, however as others are pointing out, there are a lot of active
> > > > developers who do not agree with some of the key points of the GPLv3
> > > > (including Linus), so until you convince them that the GPLv3 is better 
> > it
> > >
> > >  Last heard, Linus was quite impressed with the toned down version of
> > >  the final draft of GPLv3. I think Linus, and other major developers
> > >  should make their stand on this issue clear so that the kernel
> > >  community can discuss the future steps.
> >
> > "future steps"?  Hah.
> >
> > My code is going to stay GPLv2 as the v3 license is horrible for kernel
> > code for all of the reasons I have said in the past, plus a few more
> > (what, I can make an "industrial" product but not a commercial one?
> 
>      ^^^^^^^^^^^^^
>  What exactly in GPLv3 forbids you from making a commercial product?

Nothing "forbids" me, it's just the artifical distinstion of the two is,
in my opinion, stupid and foolish.  You are trying to define use-cases
to justify their notion that you must give up the hardware keys for one
type of device, yet not for another.

Even the people that feel that v2 says you need to give up the keys
think this is dumb.  But we've been through all of that before (see
previous long thread about v3 and why the kernel developers hate it, it
all still applys to the final draft.)

greg k-h

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 12, 2007, Greg KH <greg@kroah.com> wrote:

> (see previous long thread about v3 and why the kernel developers
> hate it, it all still applys to the final draft.)

You mean all the misunderstandings? ;-)

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bernd Paysan]

On Wednesday 13 June 2007 06:53, Alexandre Oliva wrote:
> You mean all the misunderstandings? ;-)

My impression as well is that there are many misunderstandings, even 
concerning the status of Linux itself. Linus is much better at kernel 
hacking than at license issues, and that's true for most other kernel 
hackers, too - that's why we have Eben Moglen to hack the license.

I want to add my two cents on what I think the legal status of the 
individual contributions to Linux are. The thing in question is not the 
GPLv2 itself (which is pretty clear that code without explicit statements 
is under "any", and if you make an explicit statement, it should read "GPL 
version two, or (at your option) any later version"), it's this text on the 
top of /usr/src/linux/COPYING:

"Also note that the only valid version of the GPL as far as the kernel
 is concerned is _this_ particular version of the license (ie v2, not
 v2.2 or v3.x or whatever), unless explicitly otherwise stated.

                        Linus Torvalds"

This text was added in or around 2.4.0-test9, but without asking for 
permission (neither from the FSF, which has the copyright of the GPL, nor 
from the other authors of the Linux kernel), and with some controversion 
afterwards.

This particular comment to how the GPL is applied to the Linux kernel 
therefore doesn't change the GPL as such (it can't without breaking 
copyright), neither does it change the licensing conditions the original 
authors put on their contribution (it can't without breaking copyright, 
either), but may only provide interpretations downstream (for the user). 
Linus is also entitled to make clairifications there, which the first 
paragraph obviously does (i.e. the text Linus added is not a change of the 
license, but a comment on it).

Again: What Linus is entitled to do is to *select* the license under which 
he redistributes the code downstream. What he can't do is to *change* the 
intention of the original author. So if you can choose what this somewhat 
ambiguous message means, and restrict yourself to reasoning that doesn't go 
into nonsense or copyright infringement, you'll pretty much come to the 
conclusion that the only thing Linus could have done back then without 
asking for permission is the license condition how he *redistributes* the 
compiled work called "Linux kernel" (it's a "compiled work", because it 
consists of a compilation of individual files from many authors).

The GPLv2 however is very clear how the end user gets the license: from the 
original author. Not from the man in the middle, from a distributor or 
kernel maintainer, who can neither add nor drop restrictions/permissions 
(and thus the special rights of a compilation editor are void). The author 
can only speak for himself, not by behalf of somebody else, as well as the 
compilation editor. That's why the FSF is so strict about having each 
author stating copyright and the license conditions on the top of the 
file - nobody else can.

So my conclusion is: If you, as contributor to the Linux kernel, want to 
make clear that your work really is GPLv2 only, you have to do that 
yourself, you have to add a notice like above to files where you 
exclusively own copyright. Very few have done that in the past, most people 
who *did* explicitely declare what versions of the GPL they want their work 
under, did choose the default text from the GPLv2, which sais "GPLv2 or 
later" (most use the GPL text template). The rest (the majority) did not 
choose to say anything, which under the GPL regime means "any"; and nobody 
but the author himself can change that (by adding a specific version). 
Linus can't change the GPL regime, because he can't change the GPL.

So, IMHO and IANAL, technically, there are only a few files in Linux which 
really can't work in a GPLv2+GPLv3 compiled Linux, and a few files wouldn't 
be a problem.

>From a practical point of view, I fully agree with Linus that there's no 
point in switching over to the GPLv3 next month unless there's some 
valueable contribution out there that's only available under GPLv3 (maybe 
from OpenSolaris), or the Linux kernel developers understand the GPLv3 
better. I don't see this point in the near future.

But what I want to say: The route to GPLv3 is not as blocked as it appears. 
And the GPLv2 is even better than you think: It paves that road as well: In 
practice, only projects that have a thight authorship control can really 
make their project GPLv2 only (like MySQL), and those projects have no 
problem to change their mind later.

What I don't understand about the GPLv3 with keys is why that depends on the 
use case. As user of commercial devices like company routers, firewalls and 
such, which often are Linux based, I don't want them sealed by the vendor, 
as well. An explicit statement is even worse than an implicit one (as in 
the GPLv2, which has been tested in a German court by Harald Welte - 
Siemens had to turn in the keys).

And now flame me to death ;-).

-- 
Bernd Paysan
"If you want it done right, you have to do it yourself"
http://www.jwdt.com/~paysan/

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Krzysztof Halasa]

Bernd Paysan <bernd.paysan@gmx.de> writes:

> I want to add my two cents on what I think the legal status of the 
> individual contributions to Linux are. The thing in question is not the 
> GPLv2 itself (which is pretty clear that code without explicit statements 
> is under "any",

That's not exactly true. A work without explicit statements is not
licenced at all.

> This particular comment to how the GPL is applied to the Linux kernel 
> therefore doesn't change the GPL as such

Of course.

> Again: What Linus is entitled to do is to *select* the license under which 
> he redistributes the code downstream.

Sure.

> The GPLv2 however is very clear how the end user gets the license: from the 
> original author.

I'd be surprised if it's for GPL to decide.

> Not from the man in the middle, from a distributor or 
> kernel maintainer, who can neither add nor drop restrictions/permissions 
> (and thus the special rights of a compilation editor are void). The author 
> can only speak for himself, not by behalf of somebody else, as well as the 
> compilation editor.

How about derived works?
Am I free to get BSD source, incorporate it in GPL project, and release
the whole under GPL?
Sure, the original source stays BSD but I don't distribute it.

> That's why the FSF is so strict about having each 
> author stating copyright and the license conditions on the top of the 
> file - nobody else can.

I'm not sure the copyright laws define "files".

> So my conclusion is: If you, as contributor to the Linux kernel, want to 
> make clear that your work really is GPLv2 only, you have to do that 
> yourself, you have to add a notice like above to files where you 
> exclusively own copyright.

I don't think the law works like that.
By default you have no rights to someone's work (file or project).
The only licence I can find with Linux is GPL v2, isn't it? And even
that wasn't stated explicite until that 2.4.0something (though there
is a consensus that the COPYING file was indeed a licence for the
whole kernel).

Then you may have additional rights, such as those given in various
source files.

> The rest (the majority) did not 
> choose to say anything, which under the GPL regime means "any";

What exactly is the "GPL regime" and how is it defined by copyright
law and/or the GPL licence itself (or will of copyright holders etc.)?
-- 
Krzysztof Halasa

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bernd Paysan]

[-- Attachment #1: Type: text/plain, Size: 6225 bytes --]

On Wednesday 13 June 2007 15:11, Krzysztof Halasa wrote:
> > The GPLv2 however is very clear how the end user gets the license: from
> > the original author.
>
> I'd be surprised if it's for GPL to decide.

If you choose the GPL as license, the text of the GPL are the conditions. 
Otherwise, the GPL would be pure nonsense (as would be any other license).

> > Not from the man in the middle, from a distributor or
> > kernel maintainer, who can neither add nor drop
> > restrictions/permissions (and thus the special rights of a compilation
> > editor are void). The author can only speak for himself, not by behalf
> > of somebody else, as well as the compilation editor.
>
> How about derived works?
> Am I free to get BSD source, incorporate it in GPL project, and release
> the whole under GPL?
> Sure, the original source stays BSD but I don't distribute it.

Derivated work is a product of several authors, therefore each author may 
put different conditions on his part of the work - as long as they are 
compatible, it's ok. A derivated work originally under BSD, now with a 
patch under GPL can only be distributed under GPL, but not under BSD 
(because GPL requires redistribution under GPL, whereas BSD doesn't care). 
If you take out the patch, and revert the work to the BSD one, you are free 
to redistribute it under BSD.

There's no point of discussing that the Linux kernel *as a whole* (as 
compilation) currently is under GPLv2 only, since it sais so, and a few 
files also explicitely say so. The whole combination is GPLv2 only, but 
most parts aren't.

> > That's why the FSF is so strict about having each
> > author stating copyright and the license conditions on the top of the
> > file - nobody else can.
>
> I'm not sure the copyright laws define "files".

Copyright law defines "work", and in terms of computer programs, source 
code "work" goes into files. Or patches, but patches end up distributed 
over several files.

The nice thing about this is that you can make automatic checks about the 
license you actually have to fulfill. E.g. if you compile a hypothetical 
Linux 2.8.15.3 without ZFS and dtrace in 2009, you may end up with 
compiling only GPLv2-compatible code, and therefore can tivoize your system 
(unless you sell it to Germany, where the GPLv2 outlaws tivoizing by 
intent), but if you add either ZFS or dtrace, you can't.

> > So my conclusion is: If you, as contributor to the Linux kernel, want
> > to make clear that your work really is GPLv2 only, you have to do that
> > yourself, you have to add a notice like above to files where you
> > exclusively own copyright.
>
> I don't think the law works like that.
> By default you have no rights to someone's work (file or project).
> The only licence I can find with Linux is GPL v2, isn't it?

Yes, and the GPLv2 sais "if the FSF releases a new version of the GPL, you 
may update" (section 9):

"  9. The Free Software Foundation may publish revised and/or new versions
of the General Public License from time to time.  Such new versions will
be similar in spirit to the present version, but may differ in detail to
address new problems or concerns.

Each version is given a distinguishing version number.  If the Program
specifies a version number of this License which applies to it and "any
later version", you have the option of following the terms and conditions
either of that version or of any later version published by the Free
Software Foundation.  If the Program does not specify a version number of
this License, you may choose any version ever published by the Free Software
Foundation."

This is normal contract law (you have to say "yes" to a new M$ EULA every 
few months or so if you are an unlucky Windows user, and like to use their 
patches), contracts can change over time. The FSF is rather nice here, they 
say users and contributors can make choices to which contract they use.

Given that the license you find with Linux is GPLv2, anyway, the comment 
from Linus seems to be superfluous. The license already says which version 
it is. But it has this upgrade option, and one possible interpretation of 
Linus' comment is "no, it doesn't have this update option".

> And even 
> that wasn't stated explicite until that 2.4.0something (though there
> is a consensus that the COPYING file was indeed a licence for the
> whole kernel).
>
> Then you may have additional rights, such as those given in various
> source files.
>
> > The rest (the majority) did not
> > choose to say anything, which under the GPL regime means "any";
>
> What exactly is the "GPL regime" and how is it defined by copyright
> law and/or the GPL licence itself (or will of copyright holders etc.)?

If I use GPL as license, I'm under "GPL regime", i.e. the terms of the GPL 
apply. The GPL requires that I need to speak up explicitely if I want to 
limit the choice of licenses - if I don't say anything, it defaults to "any 
GPL". This is a restriction that goes from author to user, sinde the GPL 
cuts away all middle-men (restrictions applied by middle-men can be 
reverted).

If I decide to build my own compilation of Linux kernel patches* (i.e. a -bp 
kernel, like the -mm kernel is a different compilation of Linux kernel 
patches as the mainline kernel), I'm free to choose under which conditions 
I redistribute it, given that it's compatible with the conditions the 
original authors have chosen. Most of them have said nothing (other than 
implicitely that it's ok for them to put it under GPL, as they haven't 
opposed to inclusion into the Linux kernel), some have said GPLv2 or later, 
some say GPLv1.1 or later (e.g. the parport driver) and a few have 
said "GPLv2 only". Now, I may rewrite those few "GPLv2 only" files, and 
then I have a GPLv2-or later compatible linux-some.version-bp kernel. And 
into this kernel, I can add code under GPLv3 (once the GPLv3 is ready and 
there's code worth to add under GPLv3), which limits me to redistribute the 
whole thing under GPLv3.

-- 
Bernd Paysan
"If you want it done right, you have to do it yourself"
http://www.jwdt.com/~paysan/

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Jan Harkes]

On Wed, Jun 13, 2007 at 04:24:36PM +0200, Bernd Paysan wrote:
> There's no point of discussing that the Linux kernel *as a whole* (as 
> compilation) currently is under GPLv2 only, since it sais so, and a few 
> files also explicitely say so. The whole combination is GPLv2 only, but 
> most parts aren't.

You claim that any source files without a notices are 'any version of
the GPL'. But I read the license and you are totally wrong about that.

The GPL applies to "the Program" which in this case is the Linux kernel
as a whole and it in fact does indicate a specific version. All code
submitted and included in this program has has been submitted with the
understanding that the work as a whole is specifically licensed as
GPLv2. Some authors have granted additional rights, such as dual BSD/GPL
or GPLv2 and later and explicitly added such a notice.

All other code is simply copyrighted, and the only available license is
the GPLv2. Take for example fs/inode.c. Notice how it doesn't have GPL
boilerplate, but it is clearly indicating that it is copyrighted. So
taking that file by itself out of the context of the kernel and then
distributing it would clearly be a copyright violation. The only one
reason you can distribute that code is because of the GPLv2 that covers
the Linux kernel (i.e. "the Program").

> > > So my conclusion is: If you, as contributor to the Linux kernel, want
> > > to make clear that your work really is GPLv2 only, you have to do that
> > > yourself, you have to add a notice like above to files where you
> > > exclusively own copyright.

The kernel is explicitly licensed as GPLv2, any contributions (source
files/parts of the work) that wish to grant additional rights have to
specify so explicitly, and not the other way around however much you'd
like that.

> patches as the mainline kernel), I'm free to choose under which conditions 
> I redistribute it, given that it's compatible with the conditions the 
> original authors have chosen. Most of them have said nothing (other than 
> implicitely that it's ok for them to put it under GPL, as they haven't 
> opposed to inclusion into the Linux kernel), some have said GPLv2 or later, 

Reread section 9 and consider that "the Program" is the Linux kernel,
which does explicitly state a version and does not include the "and any
later" option. Any source that does not explicitly specify additional
rights is GPLv2.

Jan

RE: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Schwartz]

> The GPL applies to "the Program" which in this case is the Linux kernel
> as a whole and it in fact does indicate a specific version. All code
> submitted and included in this program has has been submitted with the
> understanding that the work as a whole is specifically licensed as
> GPLv2. Some authors have granted additional rights, such as dual BSD/GPL
> or GPLv2 and later and explicitly added such a notice.

Since the Linux kernel as a whole does not have a single author, it is
impossible to license it as a whole. Nobody has the authority to do that.
(The GPL is not a copyright assignment type license.)

Fortunately, the GPL clears this up:

"Each time you redistribute the Program (or any work based on the
Program), the recipient automatically receives a license from the
original licensor to copy, distribute or modify the Program subject to
these terms and conditions.  You may not impose any further
restrictions on the recipients' exercise of the rights granted herein.
You are not responsible for enforcing compliance by third parties to
this License."

Linus cannot impose any further restrictions on the recipients' exercise of
the rights granted.

When you download a copy of the Linux kernel, you do not receive one license
because nobody could grant you one license. You receive a logically separate
license from each original licensor. You receive from Linus only a license
to his contributions.

Note that you cannot take a GPLv2+ work and redistribute it as GPLv3 only.
You can license your contributions as GPLv3 only of course. However, each
recipient still receives a GPLv2+ license to the parts that were originally
licensed that way. The people you distribute the work from receive licenses
from the original licensors to those parts, and you have no right to modify
that license. (See GPL section 6, quoted above.)

DS

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Rob Landley]

On Thursday 14 June 2007 15:28:34 David Schwartz wrote:
> > The GPL applies to "the Program" which in this case is the Linux kernel
> > as a whole and it in fact does indicate a specific version. All code
> > submitted and included in this program has has been submitted with the
> > understanding that the work as a whole is specifically licensed as
> > GPLv2. Some authors have granted additional rights, such as dual BSD/GPL
> > or GPLv2 and later and explicitly added such a notice.
>
> Since the Linux kernel as a whole does not have a single author, it is
> impossible to license it as a whole. Nobody has the authority to do that.
> (The GPL is not a copyright assignment type license.)

Actually, Linus Torvalds, as maintainer, probably has a compilation copyright.  
See "compilations and abridgements" in 
http://www.copyright.gov/circs/circ14.html

> Fortunately, the GPL clears this up:
>
> "Each time you redistribute the Program (or any work based on the
> Program), the recipient automatically receives a license from the
> original licensor to copy, distribute or modify the Program subject to
> these terms and conditions.  You may not impose any further
> restrictions on the recipients' exercise of the rights granted herein.
> You are not responsible for enforcing compliance by third parties to
> this License."
>
> Linus cannot impose any further restrictions on the recipients' exercise of
> the rights granted.

If you combine dual licensed code (such as MPL + GPL) with code under only one 
of those licenses (MPL only), the resulting derived work cannot be 
distributed under the dual license, only under one license.  The giant 
derived work knows as Linux has only been distributable under exactly one 
license (GPLv2, the complete text of which is included in the source tarball 
and it's harder to be more explicit than that about which license you mean) 
since version 0.12.

By the way, this entire "oh no, we can use it GPLv3 no matter what you say" 
line of argument is rude.  Linus and most of his lieutenants have explicitly 
said "our contributions are GPLv2 only".  Linus said this explicitly seven 
years ago:
http://www.uwsg.iu.edu/hypermail/linux/kernel/0009.1/0096.html

He confirmed and elaborated his position when people first started pestering 
about v3:
http://lwn.net/Articles/169825/

In James Bottomley's position paper last year, a number of prominent kernel 
developers stated their objection and that their contributions were GPLv2 
only:
  http://lkml.org/lkml/2006/9/22/217

  James E.J. Bottomley             Mauro Carvalho Chehab
  Thomas Gleixner            Christoph Hellwig           Dave Jones
  Greg Kroah-Hartman              Tony Luck           Andrew Morton
  Trond Myklebust             David Woodhouse

Let me translate this into simpler terms:

<lolcats>
GPLv3: Does not want!
</lolcats>

The _reason_ it's rude to go on about it is that several people have chosen to 
see this entire debate as an interesting intellectual exercise, "how much 
code could a GPLv3 licensed project lift from the Linux kernel".   Yet if you 
substitute "BSD Licensed" in there, it's easy to recognize how obnoxious the 
pestering is, despite much of the code in Linux having come from BSD sources.

You don't take Linux kernel code and stick it into a BSD project, even though 
some of it was BSD originally, because Linux (every line of it) is GPLv2.  If 
you want the code under a different license, you go to a differently licensed 
upstream source, such as the original author or the project we adapted it 
from.  If you're not to lift code from Linux to BSD license it, lifting code 
from Linux to GPLv3 it is morally and legally no different.

Linus made his decision, most of his lieutenants explicitly confirmed that 
decision.  Please admit to yourselves that you're arguing that they should 
all change their minds because you don't like their decision, not because 
they didn't have the right to make it or that there's some loophole that 
invalidates it.  What's your argument here, developers who are now 
saying "GPLv2" _accidentally_ gave permission to distribute their code under 
other licenses?  Go ahead and take that to court buddy: you will lose.

If you want to create a GPLv3 fork and can trace back specific files to 
authors who are ok with GPLv3, go create your fork.  If you want to go work 
on Solaris, go do that.  (But if you want to transplant Linux code into that 
thing, talk to Sun's lawyers first.  And IBM's, and Red Hat's, and...)

If you want each and every Linux developer who has ever stated a GPLv2 only 
position to either publicly reverse said position or to be ejected from the 
project and their code tracked down and removed from the kernel via forensic 
analysis (which is the only way the Linux kernel itself could ever go GPLv3), 
then do us a favor and shut up.

> When you download a copy of the Linux kernel, you do not receive one
> license because nobody could grant you one license.

Yes you do, you receive GPLv2.  It's in the file "LICENSE" at the top level of 
the directory.  This is the one and only license you receive.

You receive this one license applied to multiple copyrights, but if you're 
confusing a copyright with a license I can't help you. 

> You receive a logically 
> separate license from each original licensor. You receive from Linus only a
> license to his contributions.

A) Look up "compilation copyright".

B) The whole point of the GPL is that the license applies to the entire 
derived work, as a whole.  You can either distribute the whole thing under 
the GPL, or you cannot distribute period.  Therefore, you're saying you can 
take code that was distributed to you under GPLv2 (and only GPLv2), and 
redistribute it under another set of license terms.  In the name of GPLv3, 
you're trying to weasel around GPLv2.  Congratulations, you have achieved 
hypocrisy.

Rob

RE: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Schwartz]

> > Since the Linux kernel as a whole does not have a single author, it is
> > impossible to license it as a whole. Nobody has the authority
> > to do that.
> > (The GPL is not a copyright assignment type license.)

> Actually, Linus Torvalds, as maintainer, probably has a
> compilation copyright.
> See "compilations and abridgements" in
> http://www.copyright.gov/circs/circ14.html

It doesn't matter. He can license you his compilation, but that doesn't
license you the underlying elements.

I can make a compilation CD of great works of Rock N' Roll. I can hold a
copilation copyright in the compilation. I can license that compilation
copryight. That doesn't mean you can make, copy, or sell a CD with my
compilation on it, because you are also copying and distributing the
original works.

> If you combine dual licensed code (such as MPL + GPL) with code
> under only one
> of those licenses (MPL only), the resulting derived work cannot be
> distributed under the dual license, only under one license.

That is a common simplification. The GPL is clear that it applies
automatically with distribution. If you distribute a GPL'd work (or elements
that are GPL'd inside a larger work), those elements are relicensed under
the GPL automatically. You *cannot* prevent this from happening.

If I take the Linux kernel, modify it, and then give you a copy, you get a
license under GPLv2 from Linus to all of those elements that he placed under
the GPL. I cannot stop or modify this. It applies even if I get separate
permission from Linus to distribute his contributions under some other
conditions.

> The giant
> derived work knows as Linux has only been distributable under exactly one
> license (GPLv2, the complete text of which is included in the
> source tarball
> and it's harder to be more explicit than that about which license
> you mean)
> since version 0.12.

No, not true. I don't have the court citations handy, but it is well-settled
law that a right to distribute a derivative work is useless without also
having the right to distribute the original work from which the derivative
was made.

> By the way, this entire "oh no, we can use it GPLv3 no matter
> what you say"
> line of argument is rude.  Linus and most of his lieutenants have
> explicitly
> said "our contributions are GPLv2 only".  Linus said this
> explicitly seven
> years ago:
> http://www.uwsg.iu.edu/hypermail/linux/kernel/0009.1/0096.html

Huh? I have never argued that any contribution made by Linus could be or had
been licensed under GPLv3. Linus has clearly indicated, along with the works
that he distributes, that the code is only offered under GPLv2. However,
Linus cannot remove rights that other people grant to their code, even if he
modifies that code.

> You don't take Linux kernel code and stick it into a BSD project,
> even though
> some of it was BSD originally, because Linux (every line of it)
> is GPLv2.  If
> you want the code under a different license, you go to a
> differently licensed
> upstream source, such as the original author or the project we adapted it
> from.  If you're not to lift code from Linux to BSD license it,
> lifting code
> from Linux to GPLv3 it is morally and legally no different.
>
> Linus made his decision, most of his lieutenants explicitly
> confirmed that
> decision.  Please admit to yourselves that you're arguing that
> they should
> all change their minds because you don't like their decision, not because
> they didn't have the right to make it or that there's some loophole that
> invalidates it.  What's your argument here, developers who are now
> saying "GPLv2" _accidentally_ gave permission to distribute their
> code under
> other licenses?  Go ahead and take that to court buddy: you will lose.
>
> If you want to create a GPLv3 fork and can trace back specific files to
> authors who are ok with GPLv3, go create your fork.  If you want
> to go work
> on Solaris, go do that.  (But if you want to transplant Linux
> code into that
> thing, talk to Sun's lawyers first.  And IBM's, and Red Hat's, and...)
>
> If you want each and every Linux developer who has ever stated a
> GPLv2 only
> position to either publicly reverse said position or to be
> ejected from the
> project and their code tracked down and removed from the kernel
> via forensic
> analysis (which is the only way the Linux kernel itself could
> ever go GPLv3),
> then do us a favor and shut up.

I don't know who you are talking to or what you are talking about. I haven't
seen anybody doing what you claim in this thread or anywhere else and I
certainly am not.

> > When you download a copy of the Linux kernel, you do not receive one
> > license because nobody could grant you one license.

> Yes you do, you receive GPLv2.  It's in the file "LICENSE" at the
> top level of
> the directory.  This is the one and only license you receive.

No, not true. Please read and understand GPLv2 section 6. If a work is
available under GPLv2+, and you receive that work (even if it's as part of
another work) you get it under GPLv2+. Nobody can take rights away.

> > You receive a logically
> > separate license from each original licensor. You receive from
> > Linus only a
> > license to his contributions.

> A) Look up "compilation copyright".

Do you seriously not understand that a compilation right in "Great Works of
Fiction 2002-2006" doesn't give you the right to actually distribute the
works in it? You also receive Linus' compilation copyright under GPLv2, but
that has no effect on any case except when you wish to distribute the
compilation. It isn't enough by itself to distribute the original works.

> B) The whole point of the GPL is that the license applies to the entire
> derived work, as a whole.

Unfortunately, copyright law makes that impossible without assignment. If I
write a work and you add on to it, the resulting work is one we both hold
copyright to. Absent any assignment, you would need a license from each of
us to distribute the resulting combination.

> You can either distribute the whole
> thing under
> the GPL, or you cannot distribute period.

Again, wishful thinking. Copyright law does not work that way and that's why
GPLv2 section 6 makes it clear that the license does *not* come from the
distributor but from the original author. It follows necessarily that if
there are multiple authors, there are multiple licenses.

> Therefore, you're
> saying you can
> take code that was distributed to you under GPLv2 (and only GPLv2), and
> redistribute it under another set of license terms.

No, that's not what I am saying. I am saying that if I distribute works
under GPLv2+ and you redistribute those works or works derived from them,
everyone you distribute those works to automatically receives a GPLv2+
license from me. That is what GPLv2 section 6 says, and you cannot modify
this process. You cannot take code I licensed under GPLv2+ and distribute it
"under GPLv2 only". You cannot interfere with or control the license's grant
of rights directly from me to the people you distribute to.

> In the name
> of GPLv3,
> you're trying to weasel around GPLv2.  Congratulations, you have achieved
> hypocrisy.

I'm not trying to weasel around anything. I'm trying to explain how GPLv2
actually operates. Complying with it and understanding it are not weaseling
around.

DS

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Rob Landley]

On Thursday 14 June 2007 18:24:42 David Schwartz wrote:
> I don't know who you are talking to or what you are talking about. I
> haven't seen anybody doing what you claim in this thread or anywhere else
> and I certainly am not.

I'm asking what is the _point_ of the discussion?

Linux, the project, is available under GPLv2 only.   It is not available under 
GPLv3, and its maintainers (both Linus, his lieutenants, and numerous other 
contributors) have expressed an explicit desire NOT to license it as such.

So what are the people talking about GPLv3 trying to accomplish?  Are they:

A) Trying to unanimously change the mind of Linus, his lieutentants, and all 
the other contributors who have spoken up in favor of GPLv2 only, so that 
future versions of Linux grew a new license?  (Doesn't matter if this new 
license is GPlv3, MPL, or BSD.  It's a new license Linux is not currently 
distributed under.  Bits of Linux are separately distributed under other 
licenses such as BSD, but Linux is not and won't be any time soon.)

B) Proposing the creation of a fork of Linux which identifies and replaces all 
the code that can't be licensed under GPLv3?

C) Moving to another codebase (Solaris?  The Hurd) and trying to identify 
Linux code that can be ported to that other OS under another license?

D) Blowing smoke to no actual purpose?

Right now, it's looking like D.  Is there an E that I'm not seeing?

Rob

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Jan Harkes]

On Thu, Jun 14, 2007 at 12:28:34PM -0700, David Schwartz wrote:
> > The GPL applies to "the Program" which in this case is the Linux kernel
> > as a whole and it in fact does indicate a specific version. All code
> > submitted and included in this program has has been submitted with the
> > understanding that the work as a whole is specifically licensed as
> > GPLv2. Some authors have granted additional rights, such as dual BSD/GPL
> > or GPLv2 and later and explicitly added such a notice.
> 
> Since the Linux kernel as a whole does not have a single author, it is
> impossible to license it as a whole. Nobody has the authority to do that.
> (The GPL is not a copyright assignment type license.)
> 
> Fortunately, the GPL clears this up:
> 
> "Each time you redistribute the Program (or any work based on the
> Program), the recipient automatically receives a license from the
> original licensor to copy, distribute or modify the Program subject to
> these terms and conditions.  You may not impose any further
> restrictions on the recipients' exercise of the rights granted herein.
> You are not responsible for enforcing compliance by third parties to
> this License."
> 
> Linus cannot impose any further restrictions on the recipients' exercise of
> the rights granted.
> 
> When you download a copy of the Linux kernel, you do not receive one license
> because nobody could grant you one license. You receive a logically separate
> license from each original licensor. You receive from Linus only a license
> to his contributions.
> 
> Note that you cannot take a GPLv2+ work and redistribute it as GPLv3 only.
> You can license your contributions as GPLv3 only of course. However, each
> recipient still receives a GPLv2+ license to the parts that were originally
> licensed that way. The people you distribute the work from receive licenses
> from the original licensors to those parts, and you have no right to modify
> that license. (See GPL section 6, quoted above.)

You have a good point. It can be argued that contributions before
2.4.0-test8 were in fact GPLv2+, but anything after that point has
clearly been contributed as GPLv2 only.

So now we have a bunch of pre-2.4.0-test8 code that may possibly be v2+
and files that explicitly state v2+ in their boiler plate. However many
of these files may have had additional contributions from other authors
which (unless otherwise specified) were GPLv2-only. And because v2 and
v3 are incompatible, all those files with v2-only contributions will
become v2-only when version 3 is released. Of course it may be that all
those copyright owners do not mind re-releasing their copyrighted code
as v2+, but they will have to be contacted.

Several maintainers did pay attention to such details. I once submitted
a patch that among others touched reiserfs, and I promptly got a
friendly email from Hans asking me to sign off any rights he needed to
re-release the related code under a different license, so he made sure
the combined work wouldn't end up GPLv2 only.

Jan

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Krzysztof Halasa]

Bernd Paysan <bernd.paysan@gmx.de> writes:

> If you choose the GPL as license, the text of the GPL are the conditions. 
> Otherwise, the GPL would be pure nonsense (as would be any other license).

The licence can't redefine the copyright laws. It doesn't make it pure
nonsense BTW.

> Each version is given a distinguishing version number.  If the Program
> specifies a version number of this License which applies to it and "any
> later version", you have the option of following the terms and conditions
> either of that version or of any later version published by the Free
> Software Foundation.  If the Program does not specify a version number of
> this License, you may choose any version ever published by the Free Software
> Foundation."

That would be the case if "the Program" (the whole or individual file(s))
contained something like:
"you can redistribute it and/or modify it under the terms of the GNU
General Public License as published by the Free Software Foundation".

Obviously, you could then choose any version including the first one.

> This is normal contract law (you have to say "yes" to a new M$ EULA every 
> few months or so if you are an unlucky Windows user, and like to use their 
> patches),

This is very different (though unrelated) - patches are new work and
I hope you're free to use your MS Windows under old conditions if you
don't need the new work.

> contracts can change over time. The FSF is rather nice here, they 
> say users and contributors can make choices to which contract they use.

FSF has exactly nothing to say here (except that they've created
a useful licence). The author can choose whatever conditions he/she
likes.

> Given that the license you find with Linux is GPLv2, anyway, the comment 
> from Linus seems to be superfluous. The license already says which version 
> it is.

It seems so.

> But it has this upgrade option, and one possible interpretation of 
      ^^
> Linus' comment is "no, it doesn't have this update option".

It? What "it"?
I don't get it. If you say the licence is v2 only, then how can it have
options?

> If I use GPL as license, I'm under "GPL regime", i.e. the terms of the GPL 
> apply.

First, the local and international laws apply. It's not like selling your
soul to the devil.

> The GPL requires that I need to speak up explicitely if I want to 
> limit the choice of licenses - if I don't say anything, it defaults to "any 
> GPL". This is a restriction that goes from author to user, sinde the GPL 
> cuts away all middle-men (restrictions applied by middle-men can be 
> reverted).

... theoretically, by removing their work, perhaps. Back to reality...

> If I decide to build my own compilation of Linux kernel patches* (i.e. a -bp 
> kernel, like the -mm kernel is a different compilation of Linux kernel 
> patches as the mainline kernel), I'm free to choose under which conditions 
> I redistribute it, given that it's compatible with the conditions the 
> original authors have chosen. Most of them have said nothing (other than 
> implicitely that it's ok for them to put it under GPL, as they haven't 
> opposed to inclusion into the Linux kernel),

There is no assumption of "GPL", you can only assume GPL v2 as the kernel
is v2. And it's not left for assumptions anymore, see "signed-off-by" and
licence tags (though the tags often specify "GPL" when the actual
licence, as indicated by the author, is GPL v2).

> Now, I may rewrite those few "GPLv2 only" files, and 
> then I have a GPLv2-or later compatible linux-some.version-bp kernel.

Sure, you can rewrite all non "GPLv2 or later" code and have v3 Linux.
The problem is you think only "few" files are v2.
-- 
Krzysztof Halasa

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bernd Paysan]

[-- Attachment #1: Type: text/plain, Size: 2685 bytes --]

On Wednesday 13 June 2007 22:14, Krzysztof Halasa wrote:
> It seems so.
>
> > But it has this upgrade option, and one possible interpretation of
>
>       ^^
>
> > Linus' comment is "no, it doesn't have this update option".
>
> It? What "it"?
> I don't get it. If you say the licence is v2 only, then how can it have
> options?

By section 9. The license is v2, and basically allows to update the 
license - and it makes this a choice of the user (who also has rights to 
change stuff and redistribute it).

> > If I use GPL as license, I'm under "GPL regime", i.e. the terms of the
> > GPL apply.
>
> First, the local and international laws apply. It's not like selling your
> soul to the devil.

Contract law means that first and foremost the contract itself defines the 
rules, and only if it is not or contradicts the law, the law jumps in. The 
GPL is not really a contract, it's a license, but the law is not much 
different here, especially once you accept the GPL. If you put your code 
under GPL, the text in the GPL is the deal. The law is only the framework 
under which the deal works.

If you accept the M$ EULA, international law still applies, yet you are 
selling your soul to the devil (because the EULA sais so).

> > Now, I may rewrite those few "GPLv2 only" files, and
> > then I have a GPLv2-or later compatible linux-some.version-bp kernel.
>
> Sure, you can rewrite all non "GPLv2 or later" code and have v3 Linux.
> The problem is you think only "few" files are v2.

Because only few files say so, and they must say what they mean, because GPL 
is rather clear that if you put a file which doesn't say which version 
applies under GPL, it's "any GPL". Why is it so difficult to grok section 9 
of the current GPLv2, which people claim is well understood?

A number of kernel hacker deliberately want their work under GPLv2 only 
(like Al Viro), and they are fully entitled to do that - but they must 
announce it in a propper place (not lkml or lwn.org), and a comment in 
COPYING signed by Linus Torvalds doesn't seem to be propper to me, 
especially when the GPLv2 gives a procedure how to do it (look for the 
appendix: "How to Apply These Terms to Your New Programs").

There are good reasons to follow the advice there, and those who did follow 
the advice in the Linux kernel in the vast majority said "GPLv2 or later". 
Verbatim copy without understanding? Or is it rather that the other people 
who didn't follow the advice didn't read the GPL, and therefore understand 
it even less ;-)?

-- 
Bernd Paysan
"If you want it done right, you have to do it yourself"
http://www.jwdt.com/~paysan/

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Al Viro]

On Thu, Jun 14, 2007 at 10:23:20AM +0200, Bernd Paysan wrote:
> A number of kernel hacker deliberately want their work under GPLv2 only 
> (like Al Viro), and they are fully entitled to do that - but they must 
> announce it in a propper place (not lkml or lwn.org)

A court deposition if somebody tries to do relicensing.  At that point
I believe that I made myself sufficiently clear, so I really doubt that
"all files without explicit license get the license defendant would like
and not the one located in the tree" would fly.  But you are welcome to
test that, of course - will make for nice punitive damages.  Just make
sure to test it yourself - giving somebody else an advice that will land
them in trouble is not nice...

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bernd Paysan]

[-- Attachment #1: Type: text/plain, Size: 3963 bytes --]

On Thursday 14 June 2007 17:39, Al Viro wrote:
> On Thu, Jun 14, 2007 at 10:23:20AM +0200, Bernd Paysan wrote:
> > A number of kernel hacker deliberately want their work under GPLv2 only
> > (like Al Viro), and they are fully entitled to do that - but they must
> > announce it in a propper place (not lkml or lwn.org)
>
> A court deposition if somebody tries to do relicensing.  At that point
> I believe that I made myself sufficiently clear, so I really doubt that
> "all files without explicit license get the license defendant would like
> and not the one located in the tree" would fly.  But you are welcome to
> test that, of course - will make for nice punitive damages.

If I test it, it would be in Germany, and I really doubt that relicensing 
one copyleft to another can ever cause puntative damages here. You are only 
entitled to collect damages here when you have actual losses (that's why 
Harald Welte never gets a dime except for his defense expenses and 
voluntary donations to the FSF), but you can demand compliance. That would 
basically mean that the hypothetical linux-something.subversion-bp with 
GPLv3 parts in it can't be shipped further, because I can't fulfill all my 
obligations.

It's probably completely hypothetical, but if I really liked to be nasty, I 
could release the blackfin sound driver I've written for our digital 
amplifyer under GPLv3 or later. The code I've modified is explicitely under 
GPLv2 or later.

> Just make 
> sure to test it yourself - giving somebody else an advice that will land
> them in trouble is not nice...

As a non-lawyer, I can't give anybody legal advice in Germany, and I'd like 
to extend that to the rest of the world. This is my opinion, my 
interpretation of the GPLv2 and what's my logical reasoning what these 
three lines on top of /usr/src/linux/COPYING really mean. And there are 
only two possibilities:

* Either it means what it says, then it's quite likely a copyright 
infingement done by Linus to all those authors of linux-2.4.0-test8 and 
before, and you all may need to stop distributing Linux*, since you can't 
meet your obligations (and restart from linux-2.4.0-test8, which is the 
last legal version), or

* it does not exactly mean what it says, then you still can distribute 
Linux, but you can't really stop anyone who's updating it to GPLv3 - except 
for those few files that have explicit version numbers assigned.

BTW: If I grep through Linux, I find two files where you have noted your 
copyright and the release conditions (GPL v2), and I think last time I did 
the same thing, I found two GPLv2-files, as well - all other files with "Al 
Viro" in it apparently have multiple authors. These two files may be the 
same ones, or maybe there are two other files, making it four in total (or 
some further I missed, the text of v2 only is not as normed as the text 
for "v2 or later", but in general it's rare). These files clearly have to 
be rewritten or premission has to be asked when updating COPYING to GPLv3. 
But that's not a show-stopper.

Example, to test the legal issue:

/usr/src/linux/drivers/net/bmac.c

is Copyright (C) 1998 Randy Gobbel, and was modified by you in 1999, and has 
no specific GPL license. That's both before 2.4.0-test9, so without the 
implied "v2 only" of the "tree root". If I take this file up to GPLv3, what 
can you do against it? Nothing. What can Randy Gobbel do about it? Nothing, 
either. If you claim, it's "v2 only", Randy Gobbel could do something (e.g. 
asking puntative damages from you in the US, or denying you the right to 
redistribute Linux further, because you don't fulfill your obligations).

*) e.g. Microsoft lawyers will hunt down all pre-2.4.0-test9 copyright 
holders and pay the one who's willing to stop Linux a billion.

-- 
Bernd Paysan
"If you want it done right, you have to do it yourself"
http://www.jwdt.com/~paysan/

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Al Viro]

On Thu, Jun 14, 2007 at 06:32:57PM +0200, Bernd Paysan wrote:

> BTW: If I grep through Linux, I find two files where you have noted your 
> copyright and the release conditions (GPL v2), and I think last time I did 
> the same thing, I found two GPLv2-files, as well - all other files with "Al 
> Viro" in it apparently have multiple authors. These two files may be the 
> same ones, or maybe there are two other files, making it four in total (or 
> some further I missed, the text of v2 only is not as normed as the text 
> for "v2 or later", but in general it's rare). These files clearly have to 
> be rewritten or premission has to be asked when updating COPYING to GPLv3. 
> But that's not a show-stopper.
 
Rot.  "Multiple authors" doesn't get you out of that.  If you take a code
available under GPLv2 or later and combine it with code under specific
version of GPL, result is under than specific version of GPL.  If you want
to argue against that, make sure to Cc RMS on that, I would really like to
hear his opinion.

Multiple authors == need permission from each author with enough
contributions to that file to make the contributions in question
copyrightable.

And in my case (and case of gregkh, and...) that would be considerably
more than a couple of files.  Really.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Dmitry Torokhov]

On 6/14/07, Al Viro <viro@ftp.linux.org.uk> wrote:
> On Thu, Jun 14, 2007 at 06:32:57PM +0200, Bernd Paysan wrote:
>
> > BTW: If I grep through Linux, I find two files where you have noted your
> > copyright and the release conditions (GPL v2), and I think last time I did
> > the same thing, I found two GPLv2-files, as well - all other files with "Al
> > Viro" in it apparently have multiple authors. These two files may be the
> > same ones, or maybe there are two other files, making it four in total (or
> > some further I missed, the text of v2 only is not as normed as the text
> > for "v2 or later", but in general it's rare). These files clearly have to
> > be rewritten or premission has to be asked when updating COPYING to GPLv3.
> > But that's not a show-stopper.
>
> Rot.  "Multiple authors" doesn't get you out of that.  If you take a code
> available under GPLv2 or later and combine it with code under specific
> version of GPL, result is under than specific version of GPL.  If you want
> to argue against that, make sure to Cc RMS on that, I would really like to
> hear his opinion.
>
> Multiple authors == need permission from each author with enough
> contributions to that file to make the contributions in question
> copyrightable.
>
> And in my case (and case of gregkh, and...) that would be considerably
> more than a couple of files.  Really.

I would expect that if you contribute to a file that explicitely says
"GPL v2 or later" and you do not change that wording then you agree
GPL v2 or later for that particular contribution. So for example
drivers/net/plip.c could be changed to GPL v3 even though you
contributed to it.

-- 
Dmitry

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Al Viro]

On Thu, Jun 14, 2007 at 01:01:20PM -0400, Dmitry Torokhov wrote:
> >Multiple authors == need permission from each author with enough
> >contributions to that file to make the contributions in question
> >copyrightable.
> >
> >And in my case (and case of gregkh, and...) that would be considerably
> >more than a couple of files.  Really.
> 
> I would expect that if you contribute to a file that explicitely says
> "GPL v2 or later" and you do not change that wording then you agree
> GPL v2 or later for that particular contribution. So for example
> drivers/net/plip.c could be changed to GPL v3 even though you
> contributed to it.

After you exclude such cases it's still more than a couple of files...

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Dmitry Torokhov]

On 6/14/07, Al Viro <viro@ftp.linux.org.uk> wrote:
> On Thu, Jun 14, 2007 at 01:01:20PM -0400, Dmitry Torokhov wrote:
> > >Multiple authors == need permission from each author with enough
> > >contributions to that file to make the contributions in question
> > >copyrightable.
> > >
> > >And in my case (and case of gregkh, and...) that would be considerably
> > >more than a couple of files.  Really.
> >
> > I would expect that if you contribute to a file that explicitely says
> > "GPL v2 or later" and you do not change that wording then you agree
> > GPL v2 or later for that particular contribution. So for example
> > drivers/net/plip.c could be changed to GPL v3 even though you
> > contributed to it.
>
> After you exclude such cases it's still more than a couple of files...
>

Undoubtedly. I was just responding to neet to contact multiple authors point.

-- 
Dmitry

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Paulo Marques]

Al Viro wrote:
> On Thu, Jun 14, 2007 at 01:01:20PM -0400, Dmitry Torokhov wrote:
>>> Multiple authors == need permission from each author with enough
>>> contributions to that file to make the contributions in question
>>> copyrightable.
>>>
>>> And in my case (and case of gregkh, and...) that would be considerably
>>> more than a couple of files.  Really.
>> I would expect that if you contribute to a file that explicitely says
>> "GPL v2 or later" and you do not change that wording then you agree
>> GPL v2 or later for that particular contribution. So for example
>> drivers/net/plip.c could be changed to GPL v3 even though you
>> contributed to it.
> 
> After you exclude such cases it's still more than a couple of files...

FWIW,

$ find -name "*.c" | xargs grep "any later version" | wc -l
3138
$ find -name "*.c" | wc -l
9482

Watching the output of the first grep without "wc -l" shows that, 
although it is not 100% accurate, it is still ok just to get a rough 
estimate.

So yes, ~6300 files are definitely more than a couple ;)

-- 
Paulo Marques - www.grupopie.com

"God is love. Love is blind. Ray Charles is blind. Ray Charles is God."

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Paulo Marques <pmarques@grupopie.com> wrote:

> $ find -name "*.c" | xargs grep "any later version" | wc -l
> 3138
> $ find -name "*.c" | wc -l
> 9482

How many of these don't mention version 2?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Dmitry Torokhov]

On 6/14/07, Alexandre Oliva <aoliva@redhat.com> wrote:
> On Jun 14, 2007, Paulo Marques <pmarques@grupopie.com> wrote:
>
> > $ find -name "*.c" | xargs grep "any later version" | wc -l
> > 3138
> > $ find -name "*.c" | wc -l
> > 9482
>
> How many of these don't mention version 2?
>

It does not matter. GPL v2 and later can be reduced to v2 by
recepient. Linus did just that so unless individual source file
explicitely carries "and later" it is v2.

-- 
Dmitry

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Krzysztof Halasa]

"Dmitry Torokhov" <dmitry.torokhov@gmail.com> writes:

> It does not matter. GPL v2 and later can be reduced to v2 by
> recepient. Linus did just that so unless individual source file
> explicitely carries "and later" it is v2.

Well, if it said "licenced under GPL" it would mean any GPL.
Though it's probably uncommon.
-- 
Krzysztof Halasa

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bernd Paysan]

[-- Attachment #1: Type: text/plain, Size: 2239 bytes --]

On Thursday 14 June 2007 20:55, Dmitry Torokhov wrote:
> It does not matter. GPL v2 and later can be reduced to v2 by
> recepient.

And expanded by the next recipient to GPLv2 or later, as long as the first 
recipient does not make a substantial modification ("substantial" is a 
copyright term - there is no precise definition how much must be modified, 
but a line or two may not count as "substantial"). This is because you 
receive the license from the original author, not from the man in the 
middle.

What's still open is how you can change the conditions if you do make 
substantial changes. My position is: If you modify work (i.e. work with 
multiple licensors), you are not in a position to change the conditions, 
since you have to pass on the rights you have (and that included "you may 
use any GPL" or "you may use GPLv2 or later"). If you create work, you are 
the only licensor, so you can choose (the created work needs to be 
sufficiently independent, which e.g. a ZFS from OpenSolaris clearly would 
be). If you combine work, you can ship the combined work only under a GPL 
version that matches the common subset, but you cannot change the license 
of the parts. By adding stuff under GPLv2 only, and then combining the work 
to a larger work, you may achieve the effect that the larger work is then 
GPLv2 only. You cannot achieve that people take out the GPLv2 only work, 
and recombine it for themselves - these people then can choose other 
license, and combine it e.g. with GPLv3 code.

If you distribute work under multiple possible license, you can also choose 
which conditions you want to fulfill. But that's not imposing restrictions 
to the next recipient, so the next recipient can choose again.

It's so simple: Only the author can impose restrictions, everybody else, 
when using the GPL, has to pass on all the rights he got. If you get a 
court verdict depends on the law system, and in an anglo-saxonian (roman) 
system, you might get away by exploiting loopholes, but in a Code Napoleon 
system, you don't, because exploiting loopholes is not "good faith".

-- 
Bernd Paysan
"If you want it done right, you have to do it yourself"
http://www.jwdt.com/~paysan/

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Dmitry Torokhov]

On 6/15/07, Bernd Paysan <bernd.paysan@gmx.de> wrote:
> On Thursday 14 June 2007 20:55, Dmitry Torokhov wrote:
> > It does not matter. GPL v2 and later can be reduced to v2 by
> > recepient.
>
> And expanded by the next recipient to GPLv2 or later, as long as the first
> recipient does not make a substantial modification ("substantial" is a
> copyright term - there is no precise definition how much must be modified,
> but a line or two may not count as "substantial"). This is because you
> receive the license from the original author, not from the man in the
> middle.
>

No, you do receive the license from the person or entity you received
the program. You have an _option_ to go to the original author and get
copy of original code with original license (or maybe other license).

-- 
Dmitry

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alan Cox]

> No, you do receive the license from the person or entity you received
> the program. You have an _option_ to go to the original author and get
> copy of original code with original license (or maybe other license).

You receive the licence from the original author. The GPL contains no
text allowing a third party to grant new licences.

You may well receive the COPYING file from a different party. The licence
may well place duties on the person who supplied you but the actual right
to use comes from the original author.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Dmitry Torokhov]

On 6/15/07, Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:
> > No, you do receive the license from the person or entity you received
> > the program. You have an _option_ to go to the original author and get
> > copy of original code with original license (or maybe other license).
>
> You receive the licence from the original author. The GPL contains no
> text allowing a third party to grant new licences.
>

GPL itself does not. But the author(s) may when they specify "any
later version", "dual GPL/BSD", etc. In this case (IMHO) distributor
in fact relicenses the code and may reduce license to sipmply BSD or
simply GPL, or "GPL v3 from now on". To "restore" license you would
need to go upstream and get the code from there.

-- 
Dmitry

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alan Cox]

O> GPL itself does not. But the author(s) may when they specify "any
> later version", "dual GPL/BSD", etc. In this case (IMHO) distributor
> in fact relicenses the code and may reduce license to sipmply BSD or
> simply GPL, or "GPL v3 from now on". To "restore" license you would
> need to go upstream and get the code from there.

I don't see anything in the GPL that permits a redistributor to change
the licence a piece of code is distributed under. If my code is GPL v2 or
later you cannot take away the "or later" unless explicitly granted
powers by the author to vary the licence.

What you most certainly can do is modify it and decide your modifications
are GPLv3 only thus creating a derived work which is GPLv3 only. However
anyone receiving your modified version and reverting the modifications is
back at v2 or later.

Alan

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Dmitry Torokhov]

On 6/15/07, Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:
> O> GPL itself does not. But the author(s) may when they specify "any
> > later version", "dual GPL/BSD", etc. In this case (IMHO) distributor
> > in fact relicenses the code and may reduce license to sipmply BSD or
> > simply GPL, or "GPL v3 from now on". To "restore" license you would
> > need to go upstream and get the code from there.
>
> I don't see anything in the GPL that permits a redistributor to change
> the licence a piece of code is distributed under. If my code is GPL v2 or
> later you cannot take away the "or later" unless explicitly granted
> powers by the author to vary the licence.
>
> What you most certainly can do is modify it and decide your modifications
> are GPLv3 only thus creating a derived work which is GPLv3 only. However
> anyone receiving your modified version and reverting the modifications is
> back at v2 or later.
>

Yes, I agree. When I am saying "distributor" it is someone like RedHat
or TiVO who do modify the code, not merely use it in ints original
form.

-- 
Dmitry

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Forrest]

On Fri, Jun 15, 2007 at 02:19:23PM +0100, Alan Cox wrote:
> 
> > GPL itself does not. But the author(s) may when they specify "any
> > later version", "dual GPL/BSD", etc. In this case (IMHO) distributor
> > in fact relicenses the code and may reduce license to sipmply BSD or
> > simply GPL, or "GPL v3 from now on". To "restore" license you would
> > need to go upstream and get the code from there.
> > 
> I don't see anything in the GPL that permits a redistributor to
> change the licence a piece of code is distributed under. If my code
> is GPL v2 or later you cannot take away the "or later" unless
> explicitly granted powers by the author to vary the licence.
> 
> What you most certainly can do is modify it and decide your
> modifications are GPLv3 only thus creating a derived work which is
> GPLv3 only. However anyone receiving your modified version and
> reverting the modifications is back at v2 or later.

But that begs the question: How do you know what has been modified so
you can revert the modifications?  There won't necessarily be any
indication of which files have been modified.

So I think Dmitry's point is valid.  Don't you need to go upstream at
least far enough to verify that you have unmodified code?

And how does the copyright work for kernel patches?  Consider a dual
licensed (i.e. anything beyond GPLv2 only) file.  Someone supplies
patches to Linus, he applies them, the resulting file is distributed
with the kernel as GPLv2.  What precisely has to happen for someone
to get that same file with equivalent patches applied that can be
distributed with the original dual license?

Somehow it seems to me that Linus would have to take the dual licensed
files from his kernel repository and copy them to a separate archive
and people would have to copy from there to keep the dual license.
Aren't the files you extract from a linux tarball only licensed to you
under the terms of GPLv2?

-- 
Dan

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alan Cox]

> But that begs the question: How do you know what has been modified so
> you can revert the modifications?  There won't necessarily be any
> indication of which files have been modified.

There will

2a) You must cause the modified files to carry prominent notices
    stating that you changed the files and the date of any change.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Dmitry Torokhov]

On 6/15/07, Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:
> > But that begs the question: How do you know what has been modified so
> > you can revert the modifications?  There won't necessarily be any
> > indication of which files have been modified.
>
> There will
>
> 2a) You must cause the modified files to carry prominent notices
>    stating that you changed the files and the date of any change.
>

So how do you deduce what to revert (without going upstream for a
pristine copy) if you see the following notice (I assume it is
prominent enough):
:

/*
 * ATTENTION!!!
 * This file was modified from it's original version on 30-FEB-2345 by XYZ
 */

?

-- 
Dmitry

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bernd Paysan]

[-- Attachment #1: Type: text/plain, Size: 1635 bytes --]

On Thursday 14 June 2007 19:20, Paulo Marques wrote:
> Watching the output of the first grep without "wc -l" shows that,
> although it is not 100% accurate, it is still ok just to get a rough
> estimate.
>
> So yes, ~6300 files are definitely more than a couple ;)

Most of them don't say anything, so they are "any GPL" by the author. When 
do you people accept that Linus can't change the GPL, he can only add 
comments of what he thinks is the case! His interpretation of the GPLv2 
might be that not saying anything about the version means "v2 only", but if 
he does so, he's simply wrong. He was wrong in the module case, as well, 
and dropped this comment a while ago. He might drop this comment in future, 
as well. In fact, anybody can drop this comment, as it's just a comment.

The kernel *as a whole* is clearly under GPLv2 only from Linus' comment, 
which is in fact true, since the common subset of GPL versions from all 
authors is indeed GPLv2 (by virtue of some files from Al Viro, and maybe 
some other explicit GPL v2 files). The author must specify the version 
himself, there simply is no other way. If you don't specify any, it's "any 
version", because I can license all patches straight from the authors. The 
way the GPLv2 allows you to explicitely specify "any version" is by not 
saying anything about the version at all. Linus isn't in the positition to 
change that unless he does a substantial change to the file, and also adds 
a comment that this file is now GPLv2 only.

-- 
Bernd Paysan
"If you want it done right, you have to do it yourself"
http://www.jwdt.com/~paysan/

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Paulo Marques]

Bernd Paysan wrote:
> On Thursday 14 June 2007 19:20, Paulo Marques wrote:
>> Watching the output of the first grep without "wc -l" shows that,
>> although it is not 100% accurate, it is still ok just to get a rough
>> estimate.
>>
>> So yes, ~6300 files are definitely more than a couple ;)

I knew I shouldn't post into the yearly GPL flame-fest... :(

> Most of them don't say anything, so they are "any GPL" by the author. 

I've contributed some code for the kernel (unlike yourself, AFAICT), and 
believe me, I did so under GPL v2. The COPYING file is pretty much self 
explanatory, so I didn't need to add any explicit license statement to 
my code.

> When 
> do you people accept that Linus can't change the GPL, he can only add 
> comments of what he thinks is the case! His interpretation of the GPLv2 
> might be that not saying anything about the version means "v2 only", but if 
> he does so, he's simply wrong. He was wrong in the module case, as well, 
> and dropped this comment a while ago. He might drop this comment in future, 
> as well. In fact, anybody can drop this comment, as it's just a comment.

Linus can't and is not _changing_ the GPL. He can however use whatever 
license he sees fit for _his_ code just like all the other kernel 
developers do.

People seem to forget that the kernel license in COPYING *never had* the 
"v2 or later" clause. Never. Period.

The only change in license was from the previous hand-made one from 
Linus into GPL v2 only. And that is perfectly fine since the previous 
license was even more permissive than GPL v2.

> The kernel *as a whole* is clearly under GPLv2 only from Linus' comment, 
> which is in fact true, since the common subset of GPL versions from all 
> authors is indeed GPLv2 (by virtue of some files from Al Viro, and maybe 
> some other explicit GPL v2 files). The author must specify the version 
> himself, there simply is no other way. If you don't specify any, it's "any 
> version", because I can license all patches straight from the authors. 

No, it is not "any version". It is the license specified in COPYING and 
nothing else.

> The way the GPLv2 allows you to explicitely specify "any version" is by not 
> saying anything about the version at all. Linus isn't in the positition to 
> change that unless he does a substantial change to the file, and also adds 
> a comment that this file is now GPLv2 only.

Man, I sure ain't a lawyer, but people in these discussions seem to not 
understand the basics at all.

And the basics are: "people who write the code decide the license to 
give it". And that's just it.

And people who write kernel code are perfectly aware that the kernel 
license is GPL v2 only, and always has been (except for the initial 
linus license).

So don't go around saying that because people don't put explicit license 
statements they don't care about the license. I care very much about the 
license, and would have never contributed to the kernel if it had a BSD 
license of some sort.

Putting a license statement in _every_ file in the kernel tree would 
just be idiotic when there is such a clear COPYING file in the root of 
the kernel tree.

-- 
Paulo Marques - www.grupopie.com

"Oh dear, I think you'll find reality's on the blink again."
Marvin The Paranoid Android

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bernd Paysan]

[-- Attachment #1: Type: text/plain, Size: 2359 bytes --]

On Friday 15 June 2007 13:49, Paulo Marques wrote:
> I've contributed some code for the kernel (unlike yourself, AFAICT), and
> believe me, I did so under GPL v2. The COPYING file is pretty much self
> explanatory, so I didn't need to add any explicit license statement to
> my code.

It's not, it's a personal comment from a misunderstanding of the GPL text. 
It's as valid as the "closed source kernel modules are legal" comment that 
was there some years ago.

> People seem to forget that the kernel license in COPYING *never had* the
> "v2 or later" clause. Never. Period.

It's there in section 9.

> The only change in license was from the previous hand-made one from
> Linus into GPL v2 only. And that is perfectly fine since the previous
> license was even more permissive than GPL v2.

??? Linus changed his own less permissive license (which 
excluded "commercial use", and certainly the TiVO device is commercial) in 
0.0x time-frame. He added and deleted comments on top of COPYING in later 
years, some simply wrong like the assertion on proprietary kernel modules. 
He added his interpretation about the version issue in 2.4.0-test9, and he 
may delete it any time he wants. It's *his* interpretation.

> No, it is not "any version". It is the license specified in COPYING and
> nothing else.

COPYING says in section 9 that there may be other versions, and if you as 
author don't specify the version, it's "any version".

> And the basics are: "people who write the code decide the license to
> give it". And that's just it.

Yo, then fucking do it! Write it in the files you contribute! If you don't, 
you haven't! You decide, not Linus Torvalds. Make it clear you have 
decided.

> And people who write kernel code are perfectly aware that the kernel
> license is GPL v2 only, and always has been (except for the initial
> linus license).

Wrong.

> Putting a license statement in _every_ file in the kernel tree would
> just be idiotic when there is such a clear COPYING file in the root of
> the kernel tree.

It's a personal comment from Linus, and not clear in any way. Do it the way 
the file COPYING itself suggests. It's not "idiotic", it's the most obvious 
way to do it.

-- 
Bernd Paysan
"If you want it done right, you have to do it yourself"
http://www.jwdt.com/~paysan/

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Dmitry Torokhov]

On 6/15/07, Bernd Paysan <bernd.paysan@gmx.de> wrote:
> On Friday 15 June 2007 13:49, Paulo Marques wrote:
>
> > No, it is not "any version". It is the license specified in COPYING and
> > nothing else.
>
> COPYING says in section 9 that there may be other versions, and if you as
> author don't specify the version, it's "any version".
>

Please read this sentence over and over until it sinks:

"If the Program specifies a version number of this License which
applies to it and "any later version", you have the option of
following the terms and conditions
either of that version or of any later version published by the Free
Software Foundation."

Pay most attention to this part "If the Program specifies xxx and "any
later version"", especially word "specifies".

-- 
Dmitry

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 15, 2007, "Dmitry Torokhov" <dmitry.torokhov@gmail.com> wrote:

> On 6/15/07, Bernd Paysan <bernd.paysan@gmx.de> wrote:
>> On Friday 15 June 2007 13:49, Paulo Marques wrote:
>> 
>> > No, it is not "any version". It is the license specified in COPYING and
>> > nothing else.
>> 
>> COPYING says in section 9 that there may be other versions, and if you as
>> author don't specify the version, it's "any version".

> Please read this sentence over and over until it sinks:

I believe he was talking about the sentence just after the one you
quoted:

  If the Program does not specify a version number of this License,
  you may choose any version ever published by the Free Software
  Foundation.

Linux files don't all specify version 2, but Linus, Al Viro and other
authors very clearly mean their contributions to be version 2 only,
while others very clearly mean their contributions to be v2+.

The moment anyone makes copyrightable changes to any such files, and
offers them under GPLv2 only (if that's at all possible; I used to
believe so, but I've read interesting, even if surprising, arguments
indicating it might not be), the result of the modification is GPLv2
only.

So there's no doubt that the whole of the kernel is meant to be under
GPLv2 only, even if some individual authors may choose to make their
contributions available under other licenses, and be willing to make
such offers when they are legally entitled to do so.


I don't quite understand what this fuss is all about.  Even if a
majority of the Linux authors had chosen GPLv2+, or GPLvany, if any
single author makes a contribution under GPLv2 only, and that
contribution is integrated, that's a veto for distributing the whole
under any other license.  This single contributor could dictate his
choice upon others, as long as his contribution was present.

IANAL, but I believe that's how it works.  And this means Linux is
under GPLv2, no matter how much of the code in it is available under
any other versions of the GPL, or even different (but compatible)
licenses.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Dmitry Torokhov]

On 6/15/07, Alexandre Oliva <aoliva@redhat.com> wrote:
> On Jun 15, 2007, "Dmitry Torokhov" <dmitry.torokhov@gmail.com> wrote:
>
> > On 6/15/07, Bernd Paysan <bernd.paysan@gmx.de> wrote:
> >> On Friday 15 June 2007 13:49, Paulo Marques wrote:
> >>
> >> > No, it is not "any version". It is the license specified in COPYING and
> >> > nothing else.
> >>
> >> COPYING says in section 9 that there may be other versions, and if you as
> >> author don't specify the version, it's "any version".
>
> > Please read this sentence over and over until it sinks:
>
> I believe he was talking about the sentence just after the one you
> quoted:
>
>  If the Program does not specify a version number of this License,
>  you may choose any version ever published by the Free Software
>  Foundation.
>

My response to this is that by including an entire copy of specific
version of GPL in the release the version number was specified. You
can't say that inclusion of copy of GPL is enough to specify class of
licenses (all GPL) but not specific version.

-- 
Dmitry

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 15, 2007, "Dmitry Torokhov" <dmitry.torokhov@gmail.com> wrote:

> On 6/15/07, Alexandre Oliva <aoliva@redhat.com> wrote:
>> On Jun 15, 2007, "Dmitry Torokhov" <dmitry.torokhov@gmail.com> wrote:
>> 
>> > On 6/15/07, Bernd Paysan <bernd.paysan@gmx.de> wrote:
>> >> On Friday 15 June 2007 13:49, Paulo Marques wrote:
>> >>
>> >> > No, it is not "any version". It is the license specified in COPYING and
>> >> > nothing else.
>> >>
>> >> COPYING says in section 9 that there may be other versions, and if you as
>> >> author don't specify the version, it's "any version".
>> 
>> > Please read this sentence over and over until it sinks:
>> 
>> I believe he was talking about the sentence just after the one you
>> quoted:
>> 
>> If the Program does not specify a version number of this License,
>> you may choose any version ever published by the Free Software
>> Foundation.

> My response to this is that by including an entire copy of specific
> version of GPL in the release the version number was specified.

It's not that simple.  Including a copy of the license is a license
requirement for any redistributor, yes.

But if you, a sole copyright holder, were to distribute your program,
without any copy of the GPL, claiming "it's under the GPL", you're not
a violator.

Then, any redistributor adds a copy of any version of the GPL (because
you didn't specify a version number).  At this point, is the program
licensed by *you* only under this specific license?

Now, if you picked one of the various versions of the license, to make
things easier for redistributors, does it mean you're choosing that
particular version of the license, even though the license itself
says otherwise?

> You can't say that inclusion of copy of GPL is enough to specify
> class of licenses (all GPL) but not specific version.

I can't say either of these, indeed.  Or rather, I can, but I wouldn't
know whether I was right ;-)

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Dmitry Torokhov]

On Friday 15 June 2007 17:08, Alexandre Oliva wrote:
> On Jun 15, 2007, "Dmitry Torokhov" <dmitry.torokhov@gmail.com> wrote:
> 
> > On 6/15/07, Alexandre Oliva <aoliva@redhat.com> wrote:
> >> On Jun 15, 2007, "Dmitry Torokhov" <dmitry.torokhov@gmail.com> wrote:
> >> 
> >> > On 6/15/07, Bernd Paysan <bernd.paysan@gmx.de> wrote:
> >> >> On Friday 15 June 2007 13:49, Paulo Marques wrote:
> >> >>
> >> >> > No, it is not "any version". It is the license specified in COPYING and
> >> >> > nothing else.
> >> >>
> >> >> COPYING says in section 9 that there may be other versions, and if you as
> >> >> author don't specify the version, it's "any version".
> >> 
> >> > Please read this sentence over and over until it sinks:
> >> 
> >> I believe he was talking about the sentence just after the one you
> >> quoted:
> >> 
> >> If the Program does not specify a version number of this License,
> >> you may choose any version ever published by the Free Software
> >> Foundation.
> 
> > My response to this is that by including an entire copy of specific
> > version of GPL in the release the version number was specified.
> 
> It's not that simple.  Including a copy of the license is a license
> requirement for any redistributor, yes.
> 
> But if you, a sole copyright holder, were to distribute your program,
> without any copy of the GPL, claiming "it's under the GPL", you're not
> a violator.
>

Distributing a copy of GPL is not a requirement for me as a licensor
however I did chose to include a copy of a specific version of a
specific license and did not make any other statements that is the
license for the work.
 
> Then, any redistributor adds a copy of any version of the GPL (because
> you didn't specify a version number).  At this point, is the program
> licensed by *you* only under this specific license?
>

If they did not make any changes then they have to include the earliest
version of GPL that applies. If they did modifications and chose GPLv4
they will have to include GPL v4 (if such requirement is in GPLv4) because
what good any other license will do me?
 
> Now, if you picked one of the various versions of the license, to make
> things easier for redistributors, does it mean you're choosing that
> particular version of the license,

Yes, this is my opinion.

> even though the license itself 
> says otherwise?
>

License does not say otherwise. License says that if there is an
_additional_ stipulation my the licensor then some other license
(non existing yet license) may be used. They had to use this wording
because these licensed do not exist yet. If GPL would say:

"If the Program specifies a version number of this License which
applies to it and "or BSD license", you have the option of following
the terms and conditions either of that version or of BSD license"

would you still say that BSD is allowed by default by GPL? "GPL v2
and later versions" is not different from "GPL v2 or BSD" or
"GPLv2 or CDDL".  

> > You can't say that inclusion of copy of GPL is enough to specify
> > class of licenses (all GPL) but not specific version.
> 
> I can't say either of these, indeed.  Or rather, I can, but I wouldn't
> know whether I was right ;-)
> 

-- 
Dmitry

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 16, 2007, Dmitry Torokhov <dtor@insightbb.com> wrote:

> On Friday 15 June 2007 17:08, Alexandre Oliva wrote:
>> >> If the Program does not specify a version number of this License,
>> >> you may choose any version ever published by the Free Software
>> >> Foundation.

> Distributing a copy of GPL is not a requirement for me as a licensor
> however I did chose to include a copy of a specific version of a
> specific license and did not make any other statements that is the
> license for the work.
 
And the copy you chose to include says the above.

Are you absolutely sure you could terminate the license of a
distributor that refrains to pass on a patent license it obtained, if
you included a copy of the GPLv2 without any other indication that
you're choosing GPLv2 and no other version of the GPL, in spite of the
above?

Would it change anything if you had released the program back when
GPLv3 wasn't under discussion, and GPLv1 was long forgotten, so most
people (yourself included) only referred to it as GPL, or GNU GPL?

>> Then, any redistributor adds a copy of any version of the GPL (because
>> you didn't specify a version number).  At this point, is the program
>> licensed by *you* only under this specific license?

> If they did not make any changes then they have to include the earliest
> version of GPL that applies.

Why?  Why does it have to be the earliest?

>> even though the license itself 
>> says otherwise?

> License does not say otherwise. License says that if there is an
> _additional_ stipulation my the licensor then some other license
> (non existing yet license) may be used.

No, you're referring to the portion you quoted, but I'm referring to
another portion, that I quoted above.

> "If the Program specifies a version number of this License which
> applies to it and "or BSD license", you have the option of following
> the terms and conditions either of that version or of BSD license"

> would you still say that BSD is allowed by default by GPL? "GPL v2
> and later versions" is not different from "GPL v2 or BSD" or
> "GPLv2 or CDDL".

Agreed.  But this is not what this is about.  This is about the
license saying something like:

  If the program does not specify a license version number, then
  you're permitted to relicense the program under the BSD license.

Since the license file itself is not part of the program (if it were,
a program under the GPL would require the GPL itself to be under the
GPL, right?), I claim the program does not specify a license version
number.

Now what?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Dmitry Torokhov]

On Friday 15 June 2007 23:51, Alexandre Oliva wrote:
> On Jun 16, 2007, Dmitry Torokhov <dtor@insightbb.com> wrote:
> 
> > On Friday 15 June 2007 17:08, Alexandre Oliva wrote:
> >> >> If the Program does not specify a version number of this License,
> >> >> you may choose any version ever published by the Free Software
> >> >> Foundation.
> 
> > Distributing a copy of GPL is not a requirement for me as a licensor
> > however I did chose to include a copy of a specific version of a
> > specific license and did not make any other statements that is the
> > license for the work.
>  
> And the copy you chose to include says the above.
> 
> Are you absolutely sure you could terminate the license of a
> distributor that refrains to pass on a patent license it obtained, if
> you included a copy of the GPLv2 without any other indication that
> you're choosing GPLv2 and no other version of the GPL, in spite of the
> above?
> 

You never sure about anything taht involves lawyers ;)

> Would it change anything if you had released the program back when
> GPLv3 wasn't under discussion, and GPLv1 was long forgotten, so most
> people (yourself included) only referred to it as GPL, or GNU GPL?
> 
> >> Then, any redistributor adds a copy of any version of the GPL (because
> >> you didn't specify a version number).  At this point, is the program
> >> licensed by *you* only under this specific license?
> 
> > If they did not make any changes then they have to include the earliest
> > version of GPL that applies.
> 
> Why?  Why does it have to be the earliest?

Earliest is wrong I suppose. What I meant is post permissive. Otherwise it
does not make any sense. And what about if there is a version of GPL that
does not require passing a copy of the license along with the program?

I guess it does not matter because somewhere it would still state
"this program is released under GPL" (as you said there is no version
number) so receient can look up what versions of GPL were ever released.

This is different from attaching a specific license.

> 
> >> even though the license itself 
> >> says otherwise?
> 
> > License does not say otherwise. License says that if there is an
> > _additional_ stipulation my the licensor then some other license
> > (non existing yet license) may be used.
> 
> No, you're referring to the portion you quoted, but I'm referring to
> another portion, that I quoted above.
> 
> > "If the Program specifies a version number of this License which
> > applies to it and "or BSD license", you have the option of following
> > the terms and conditions either of that version or of BSD license"
> 
> > would you still say that BSD is allowed by default by GPL? "GPL v2
> > and later versions" is not different from "GPL v2 or BSD" or
> > "GPLv2 or CDDL".
> 
> Agreed.  But this is not what this is about.  This is about the
> license saying something like:
> 
>   If the program does not specify a license version number, then
>   you're permitted to relicense the program under the BSD license.
> 
> Since the license file itself is not part of the program (if it were,
> a program under the GPL would require the GPL itself to be under the
> GPL, right?), I claim the program does not specify a license version
> number.
> 

Why don't you claim that actually the program is in public domain and
the license file just got there by mistake? Attaching a specific license
(and GPL v2 is a distinctive license, not a bumped up version of other
license) places work under this (and only this) license. In my book
this is different form just saying "the program is under GPL".

I guess we'll have to agree to disagree.

-- 
Dmitry

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 16, 2007, Dmitry Torokhov <dtor@insightbb.com> wrote:

>> >> Then, any redistributor adds a copy of any version of the GPL (because
>> >> you didn't specify a version number).  At this point, is the program
>> >> licensed by *you* only under this specific license?
>> 
>> > If they did not make any changes then they have to include the earliest
>> > version of GPL that applies.
>> 
>> Why?  Why does it have to be the earliest?

> Earliest is wrong I suppose. What I meant is post permissive.

Again, why?  In the absence of a version number, why wouldn't the
redistributor choose any one he liked?

> I guess it does not matter because somewhere it would still state
> "this program is released under GPL" (as you said there is no version
> number) so receient can look up what versions of GPL were ever released.

Yes.  The initial recipient knows that, because he received the
announcement by e-mail, where the "released under GPL" was.  But how
about downstream recipients?  (yeah, I'm filling in blanks and making
this up on the fly, I hope you don't mind)

> This is different from attaching a specific license.

How can the downstream recipient tell this case from the case in which
you attached one specific version of the license and didn't write
anywhere that only that version applied?

> Why don't you claim that actually the program is in public domain and
> the license file just got there by mistake?

License file there by mistake is a possibility, but this wouldn't make
the program public domain, it would rather turn actions controlled by
copyright law into copyright infringement, but as long as the
recipient acted within the unclear intent of the licensor, the
licensor probably wouldn't enforce the license anyway.  And then, if
he did, there'd be a number of defenses available for the
licensee/infringer.  But IANAL.

> Attaching a specific license (and GPL v2 is a distinctive license,
> not a bumped up version of other license) places work under this
> (and only this) license.

We'll see if that works when someone tries to takes advantage of any
of the holes in GPLv1 that GPLv2 plugged and you try to enforce
GPLv2.  I'm not sure whether to hope it will (such that this implied
v2 gets better freedom protection than v1) or won't (such that I could
redistribute under v3 ;-)

> In my book this is different form just saying "the program is under
> GPL".

I hope you've consulted a lawyer about this.  If not, it might be
safer to state your intentions more explicitly, like Linus did.

> I guess we'll have to agree to disagree.

Works for me ;-)

Best regards,

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Paulo Marques]

Bernd Paysan wrote:
> On Friday 15 June 2007 13:49, Paulo Marques wrote:
>> I've contributed some code for the kernel (unlike yourself, AFAICT), and
>> believe me, I did so under GPL v2. The COPYING file is pretty much self
>> explanatory, so I didn't need to add any explicit license statement to
>> my code.
> 
> It's not, it's a personal comment from a misunderstanding of the GPL text. 
> It's as valid as the "closed source kernel modules are legal" comment that 
> was there some years ago.

These are not changes to the license text. These are just clarifications 
to help people understand the license. They don't change what the 
license already said.

>> People seem to forget that the kernel license in COPYING *never had* the
>> "v2 or later" clause. Never. Period.
> 
> It's there in section 9.

The section 9 is meant to explain how you select one version of the 
license in a program without having to copy the entire license text to 
it, i.e., in simple programs you can just put the small text, suggested 
by FSF at the bottom of the gpl, and have the version number there, and 
that should be enough to reference the entire text.

But COPYING *is* the entire text and starts with: "
		    GNU GENERAL PUBLIC LICENSE
		       Version 2, June 1991"

so there is no confusion about the version.

>>[...]
>> And people who write kernel code are perfectly aware that the kernel
>> license is GPL v2 only, and always has been (except for the initial
>> linus license).
> 
> Wrong.

Why do you say "Wrong"? Have you contributed some code to the kernel 
thinking that the kernel was "v2 or later", only to find out later that 
it wasn't?

In case you haven't followed previous discussions, here's a pointer:

http://lkml.org/lkml/2006/9/22/176

The major kernel developers (and probably most of the total number of 
developers) are perfectly aware of the kernel license and chose GPL v2.

I'm getting pretty tired of listening to people that just _know_ what I 
should do with _my_ code. And people who treat kernel developers as 
morons who can't read a license.

We definitely need more Al Viro style comments on this thread ;)

-- 
Paulo Marques - www.grupopie.com

"The Mexicans have the Chupacabra. We have Al Viro. If you hear him 
roar, just _pray_ he's about to dissect somebody elses code than yours.. 
There is no point in running."

Linus Torvalds

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alan Cox]

> But COPYING *is* the entire text and starts with: "
> 		    GNU GENERAL PUBLIC LICENSE
> 		       Version 2, June 1991"
> 
> so there is no confusion about the version.

The version of the COPYING file (and the licence document), not of the
licence on the code.

> > Wrong.
> 
> Why do you say "Wrong"? Have you contributed some code to the kernel 
> thinking that the kernel was "v2 or later", only to find out later that 
> it wasn't?

A fair bit of the kernel is probably v2 or later but not all of it and
that shouldn't really matter as regards the kernel anyway, the GPLv2 only
bits (if v2 only is a valid status) anchor it.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Paulo Marques]

Alan Cox wrote:
>> But COPYING *is* the entire text and starts with: "
>> 		    GNU GENERAL PUBLIC LICENSE
>> 		       Version 2, June 1991"
>>
>> so there is no confusion about the version.
> 
> The version of the COPYING file (and the licence document), not of the
> licence on the code.
> 
>>> Wrong.
>> Why do you say "Wrong"? Have you contributed some code to the kernel 
>> thinking that the kernel was "v2 or later", only to find out later that 
>> it wasn't?
> 
> A fair bit of the kernel is probably v2 or later but not all of it and
> that shouldn't really matter as regards the kernel anyway, the GPLv2 only
> bits (if v2 only is a valid status) anchor it.

So we are violently agreeing, then?

This sub-thread started by me showing that:

> $ find -name "*.c" | xargs grep "any later version" | wc -l
> 3138
> $ find -name "*.c" | wc -l
> 9482

This is a somewhat crude measure but it shows that only about 30% of the 
kernel is "v2 or later" and those pieces could be used on some other "v2 
or later" project (including v3). But the kernel as a whole is v2 and my 
point was that the claim that there are just a few "v2 only" files was 
bogus.

-- 
Paulo Marques - www.grupopie.com

"As far as we know, our computer has never had an undetected error."
Weisert

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Dmitry Torokhov]

On 6/15/07, Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:
> > But COPYING *is* the entire text and starts with: "
> >                   GNU GENERAL PUBLIC LICENSE
> >                      Version 2, June 1991"
> >
> > so there is no confusion about the version.
>
> The version of the COPYING file (and the licence document), not of the
> licence on the code.
>

Using this logic one can say that Linux kernel is BSD or even public
domain and COPYING is there just for kicks.

-- 
Dmitry

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Jesper Juhl]

On 15/06/07, Dmitry Torokhov <dmitry.torokhov@gmail.com> wrote:
> On 6/15/07, Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:
> > > But COPYING *is* the entire text and starts with: "
> > >                   GNU GENERAL PUBLIC LICENSE
> > >                      Version 2, June 1991"
> > >
> > > so there is no confusion about the version.
> >
> > The version of the COPYING file (and the licence document), not of the
> > licence on the code.
> >
>
> Using this logic one can say that Linux kernel is BSD or even public
> domain and COPYING is there just for kicks.
>
No. Only the original author can specify the license. If no license at
all is specified only the author has any rights to the work, other
people don't have any right to distribute, modify or whatever.
So if the COPYING file doesn't specify the license for work without a
license clause directly in the file, then only the author has any
rights, you can't just then move in and assign an arbitrary license.
But I think you would find it very hard to argue that files
contributed to the Linux kernel without an explicit license notice
does not fall under the terms set forth in the COPYING document.

-- 
Jesper Juhl <jesper.juhl@gmail.com>
Don't top-post  http://www.catb.org/~esr/jargon/html/T/top-post.html
Plain text mails only, please      http://www.expita.com/nomime.html

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alan Cox]

On Fri, 15 Jun 2007 10:52:27 -0400
"Dmitry Torokhov" <dmitry.torokhov@gmail.com> wrote:

> On 6/15/07, Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:
> > > But COPYING *is* the entire text and starts with: "
> > >                   GNU GENERAL PUBLIC LICENSE
> > >                      Version 2, June 1991"
> > >
> > > so there is no confusion about the version.
> >
> > The version of the COPYING file (and the licence document), not of the
> > licence on the code.
> >
> 
> Using this logic one can say that Linux kernel is BSD or even public
> domain and COPYING is there just for kicks.

Not really because
	1. The file is called COPYING which rather suggests its purpose
	2. There is a note at the top of it
	3. Lots of the code contains GPL headers

Any sane Judge is going to come to the conclusion that this was the
intended licence of the code. The fact people have said so also settles
the matter pretty much.

One of the big differences between law (at least UK/US law) and code is
that the legal process seeks in part to figure out the intention of a
licence or contract. Civil law is a dispute resolution process. Not a
very good one, not a very cheap one, but as the previous system involved
sending large blokes around to the opponents HQ with swords and axes it
was found to have distinct benefit.

Alan

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Carlo Wood]

This was actually a part of a larger reply - but I decided not to sent
that. Didn't want to throw away the work I did to get accurate numbers
though ;)

   I investigated:

   find -name "*.c" | wc --lines
   11100
   find -name "*.c" | xargs egrep "(version 2.*([Ll]icense|redistribute|released|GNU|GPL|any later version|Foundation))|(([Ll]icense|redistribute|released|GNU|GPL|any later version|Foundation).*version 2)" | sed -e 's/:.*//' | sort -u | wc --lines
   4042
   find -name "*.c" | xargs egrep "(version 2.*([Ll]icense|redistribute|released|GNU|GPL|any later version|Foundation))|(([Ll]icense|redistribute|released|GNU|GPL|any later version|Foundation).*version 2)" | egrep -v '(either version 2(\.[0-9])*|version 2(\.[0-9])*( of the License)*,* or)' | sed -e 's/:.*//' | sort -u | wc --lines
   1377
   find -name "*.c" | xargs egrep "(version 2.*([Ll]icense|redistribute|released|GNU|GPL|any later version|Foundation))|(([Ll]icense|redistribute|released|GNU|GPL|any later version|Foundation).*version 2)" | egrep -v '(either version 2(\.[0-9])*|version 2(\.[0-9])*( of the License)*,* or)' | sed -e 's/:.*//' | sort -u | xargs grep 'any later version' | wc --lines
   4

   Hence there are 1377 - 4 = 1373 .c files that explicitely say 'version 2',
   and 4042 - 1373 = 2669 .c files that say 'version 2 or later'.
   There are 11100 - 1373 - 2669 = 7058 files that do not say anything.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Al Viro]

On Fri, Jun 15, 2007 at 02:03:55PM +0200, Bernd Paysan wrote:
> > I've contributed some code for the kernel (unlike yourself, AFAICT), and
> > believe me, I did so under GPL v2. The COPYING file is pretty much self
> > explanatory, so I didn't need to add any explicit license statement to
> > my code.
> 
> It's not, it's a personal comment from a misunderstanding of the GPL text. 
> It's as valid as the "closed source kernel modules are legal" comment that 
> was there some years ago.

You forgot something.  Namely, that file *without* any mentioning of
GPL or other license is either illegal to distribute at all, under
any license, *OR* inherits the default license of the project.  Which
is to say, what is stated in COPYING.

Take your pick.  For what it's worth, it would be interesting to hear
the opinion of RMS - both on how much of the kernel is possible to
distribute under v3 according to him and on the morality of your position.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 15, 2007, Al Viro <viro@ftp.linux.org.uk> wrote:

> *OR* inherits the default license of the project.

You got any case law for this?  Seriously, I could use this for
FSFLA's IRPF2007-Livre project.
http://fsfla.org/svnwiki/blogs/lxo/pub/freeing-the-lion

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Al Viro]

On Fri, Jun 15, 2007 at 03:18:24PM -0300, Alexandre Oliva wrote:
> On Jun 15, 2007, Al Viro <viro@ftp.linux.org.uk> wrote:
> 
> > *OR* inherits the default license of the project.
> 
> You got any case law for this?  Seriously, I could use this for
> FSFLA's IRPF2007-Livre project.
> http://fsfla.org/svnwiki/blogs/lxo/pub/freeing-the-lion

Umm...  What other license choices are there?  Seriously, if file
*does* get a license from somewhere (and if it doesn't, it can't
be distributed at all), where else would that license come from?

I can see one arguing that it shouldn't be distributed at all (and
we obviously don't want that), I can see one arguing that copyright
statement floating in root of tree in file called "COPYING" and
not tied to specific parts of that tree should apply, but I don't
see how one would argue that some other license he happens to like
should apply here.

No specific case law, but I'd expect serious [eventual] trouble for
somebody trying to slap some different license in such case.  Not
sure if anybody actually ever tried that...

IIRC, the usual argument for slapping copyright into every file is
along the lines of "making sure that it doesn't get lost when
file is lifted into another project", not "it's free for grabbing
by anyone" or "it can't be distributed at all"...

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 15, 2007, Al Viro <viro@ftp.linux.org.uk> wrote:

> On Fri, Jun 15, 2007 at 03:18:24PM -0300, Alexandre Oliva wrote:
>> On Jun 15, 2007, Al Viro <viro@ftp.linux.org.uk> wrote:
>> 
>> > *OR* inherits the default license of the project.
>> 
>> You got any case law for this?  Seriously, I could use this for
>> FSFLA's IRPF2007-Livre project.
>> http://fsfla.org/svnwiki/blogs/lxo/pub/freeing-the-lion

> Umm...  What other license choices are there?

Where does it say that there must be one?

> No specific case law, but I'd expect serious [eventual] trouble for
> somebody trying to slap some different license in such case.

Consider this (to make the freeing-the-lion story short):

  Jar file with .class files, with a copy of LGPL in the root of the
  tree.  No other license anywhere to be seen.  Is it safe to assume
  the whole thing is under the LGPL?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Al Viro]

On Fri, Jun 15, 2007 at 06:04:33PM -0300, Alexandre Oliva wrote:
> > No specific case law, but I'd expect serious [eventual] trouble for
> > somebody trying to slap some different license in such case.
> 
> Consider this (to make the freeing-the-lion story short):
> 
>   Jar file with .class files, with a copy of LGPL in the root of the
>   tree.  No other license anywhere to be seen.  Is it safe to assume
>   the whole thing is under the LGPL?

It certainly sounds like a reasonable first assumption; unless you are
aware of couterexamples, you probably would be able at least to prove
that you've acted in good faith if somebody starts to complain.  IANAL,
obviously, so ask FSF lawyers.  Really.  Especially if you are doing that
for a text associated with FSF-LA in any way.  That's what they are for.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 15, 2007, Al Viro <viro@ftp.linux.org.uk> wrote:

> On Fri, Jun 15, 2007 at 06:04:33PM -0300, Alexandre Oliva wrote:
>> > No specific case law, but I'd expect serious [eventual] trouble for
>> > somebody trying to slap some different license in such case.
>> 
>> Consider this (to make the freeing-the-lion story short):
>> 
>> Jar file with .class files, with a copy of LGPL in the root of the
>> tree.  No other license anywhere to be seen.  Is it safe to assume
>> the whole thing is under the LGPL?

> It certainly sounds like a reasonable first assumption; unless you are
> aware of couterexamples, you probably would be able at least to prove
> that you've acted in good faith if somebody starts to complain.  IANAL,
> obviously, so ask FSF lawyers.  Really.  Especially if you are doing that
> for a text associated with FSF-LA in any way.  That's what they are for.

I've covered my grounds and talked to lawyers in Brazil, where this
all happened.  But it wouldn't hurt me to have cases of law abroad,
which is why I asked.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Glauber de Oliveira Costa]

> > And the basics are: "people who write the code decide the license to
> > give it". And that's just it.
>
> Yo, then fucking do it! Write it in the files you contribute! If you don't,
> you haven't! You decide, not Linus Torvalds. Make it clear you have
> decided.
>
> > And people who write kernel code are perfectly aware that the kernel
> > license is GPL v2 only, and always has been (except for the initial
> > linus license).
>
> Wrong.
>
> > Putting a license statement in _every_ file in the kernel tree would
> > just be idiotic when there is such a clear COPYING file in the root of
> > the kernel tree.
>
> It's a personal comment from Linus, and not clear in any way. Do it the way
> the file COPYING itself suggests. It's not "idiotic", it's the most obvious
> way to do it.

If it wasn't clear someday, it seems to be utterly clear now, after
all this discussion (that is not the first related, AFAIK). Why is the
"spirit" so important for the GPLv{2,3} understanding, and not
important at all here? The intention that the kernel is gplv2 is very
clearly stated.

-- 
Glauber de Oliveira Costa.
"Free as in Freedom"
http://glommer.net

"The less confident you are, the more serious you have to act."

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Dmitry Torokhov]

On 6/15/07, Bernd Paysan <bernd.paysan@gmx.de> wrote:
> On Thursday 14 June 2007 19:20, Paulo Marques wrote:
> > Watching the output of the first grep without "wc -l" shows that,
> > although it is not 100% accurate, it is still ok just to get a rough
> > estimate.
> >
> > So yes, ~6300 files are definitely more than a couple ;)
>
> Most of them don't say anything, so they are "any GPL" by the author.

Woah! Stop right there. Since when a work without a license spelled
out becomes "any GPL"?

> When
> do you people accept that Linus can't change the GPL, he can only add
> comments of what he thinks is the case! His interpretation of the GPLv2
> might be that not saying anything about the version means "v2 only", but if
> he does so, he's simply wrong. He was wrong in the module case, as well,
> and dropped this comment a while ago. He might drop this comment in future,
> as well. In fact, anybody can drop this comment, as it's just a comment.
>
> The kernel *as a whole* is clearly under GPLv2 only from Linus' comment,
> which is in fact true, since the common subset of GPL versions from all
> authors is indeed GPLv2 (by virtue of some files from Al Viro, and maybe
> some other explicit GPL v2 files). The author must specify the version
> himself, there simply is no other way. If you don't specify any, it's "any
> version", because I can license all patches straight from the authors.

Yes, you can. In this case you get _different_ software, maybe even
under different license. I bet if you go straight to the authors you
can get XFS or JFS under a commercial license, not GPL at all.

>  The
> way the GPLv2 allows you to explicitely specify "any version" is by not
> saying anything about the version at all.

Let me quote GPLv2 for you:

"If the Program specifies a version number of this License which
applies to it and "any later version", you have the option of
following the terms and conditions
either of that version or of any later version published by the Free
Software Foundation."

As you can see for "ant later version" to apply the program must
explicitely say "This program is distributed under GPLv2 and any later
version". Supplying text of GPL v2 with the program does not
automatically add that clause. IOW, if copying file would litrally
read "Distributed under GPL v2. For the text of GPL v2 go to
www.fsf.org" woudl you still argue that it is "GPLv2 and later"?

-- 
Dmitry

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Carlo Wood]

On Fri, Jun 15, 2007 at 10:14:44AM +0200, Bernd Paysan wrote:
>                                           Linus isn't in the positition to 
> change that unless he does a substantial change to the file, and also adds 
> a comment that this file is now GPLv2 only.

Which would only have effect on future additions, not the current
content of the file - of course.

Now - what if someone would write a patch for such a file (that was
'any version' before and then had a header added saying 'just version 2')
under the license 'version 2 or later'? This patch could be transformed
to 'just 2', and then applied to said source file - but, it could also
be added to the previous version of that file (without the new header)
as 'version 2 or later'.  Hence, the file can still be constructed at
any moment (provided the header is removed) as license 'version 2 or
later', UNLESS someone adds a *crucial* patch (that cannot be removed
as well, along with the header) that is explicitely made version 2
ONLY by its author.

Bottom line - adding a header to those files with "version 2 only"
by Linus is pointless.

-- 
Carlo Wood <carlo@alinoe.com>

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 13, 2007, Bernd Paysan <bernd.paysan@gmx.de> wrote:

> What I don't understand about the GPLv3 with keys is why that depends on the 
> use case. As user of commercial devices like company routers, firewalls and 
> such, which often are Linux based, I don't want them sealed by the vendor, 
> as well. An explicit statement is even worse than an implicit one (as in 
> the GPLv2, which has been tested in a German court by Harald Welte - 
> Siemens had to turn in the keys).

My personal guess as to the reasoning behind this decision is that
consumer devices are the ones that require most attention, mainly
because the home users are the ones with least (individual) power to
demand respect for their freedoms.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Wed, 13 Jun 2007, Alexandre Oliva wrote:

> On Jun 12, 2007, Greg KH <greg@kroah.com> wrote:
> 
> > (see previous long thread about v3 and why the kernel developers
> > hate it, it all still applys to the final draft.)
> 
> You mean all the misunderstandings? ;-)

I see the smiley, but I hate it how the FSF thinks others are morons and 
cannot read or think for themselves.

Any time you disagree with the FSF, you "misunderstand" (insert 
condescending voice) the issue. 

_Please_ don't continue that idiocy. Disagreement and thinking that the 
FSF is controlling and putting its fingers where they don't belong is 
_not_ misunderstanding. It's just not "blind and unquestioning obedience".

			Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 13, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:

> On Wed, 13 Jun 2007, Alexandre Oliva wrote:

>> On Jun 12, 2007, Greg KH <greg@kroah.com> wrote:
>> 
>> > (see previous long thread about v3 and why the kernel developers
>> > hate it, it all still applys to the final draft.)
>> 
>> You mean all the misunderstandings? ;-)

> I see the smiley, but I hate it how the FSF thinks others are morons and 
> cannot read or think for themselves.

Look, there was room for misunderstandings in earlier drafts of the
license.  Based on the public comments, the wording was improved.  I'd
like to think the issues that arose from misunderstandings of the
earlier drafts are no longer an issue.  Is it not so?

Keeping on making false claims about the license drafts can be one of
two things: misunderstandings, out of ambiguity in the text or
preconceptions, or ill intentions.  I'd rather believe it's the
former.

Now, of course you can look at the licenses and decide that you never
agreed with the spirit of the GPL in the first place, and that GPLv2
models better your intentions than GPLv3.

Your assessment about sharing of code between Linux and OpenSolaris
very much makes it seem like that the spirit of sharing, of letting
others run, study, modify and share the code as long as they respect
others' freedoms, has never been what moved you.  Rather, you seem to
perceive the GPL as demanding some form of payback, of contribution,
rather than the respect for others' freedoms that it requires.  In
fact, you said something along these lines yourself many months ago.

With this different frame of mind, it is not surprising at all that
you don't find GPLv3 a better license.  With different goals in mind,
reasonable people can reach different conclusions.  But claiming that
GPLv3 is changing the spirit of the license, or that it prohibits
certain kinds of software, is plain false.  In fact, the spirit has
always been described in its preamble, and it didn't change at all:
it's all about respecting others' freedoms.

Sure, this evokes a number of other nice behaviors in various players,
and it's clear to me that it's in these other nice behaviors that you
seek when you choose GPLv2.  There's nothing inherently wrong in that.

However, it seems to me that GPLv3 would do an even better job at
serving these goals than GPLv2, even if the holes v3 plugs that
enabled players to disrespect others' freedoms might steer away the
participants who are not willing to contribute, to really be part of
your community.  It's not like you lose much.

But the new defenses against disrespect for freedoms introduced in
GPLv3 may turn out to be very helpful, not only in protecting your
community from external threats, but also in strengthening
participation, as the benefits of participation outweight the
perceived costs of respecting others' freedoms.

It sure seems to me that trading some threats and non-contributors for
some more-committed participants is a good idea.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Lennart Sorensen]

On Wed, Jun 13, 2007 at 05:11:16PM -0300, Alexandre Oliva wrote:
> On Jun 13, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:
> Now, of course you can look at the licenses and decide that you never
> agreed with the spirit of the GPL in the first place, and that GPLv2
> models better your intentions than GPLv3.

I believe a number of people don't think the GPL v3 is in the same
spirit as the GPL v2.  I guess it comes down to what people thought the
spirit of the GPL v2 was.  There certainly seems to be a variety of
opinions on that, and I am not sure the FSF's opinion on it agrees with
what most others believe, but that would be rather difficult to
determine.

--
Len Sorensen

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 13, 2007, lsorense@csclub.uwaterloo.ca (Lennart Sorensen) wrote:

> I believe a number of people don't think the GPL v3 is in the same
> spirit as the GPL v2.  I guess it comes down to what people thought the
> spirit of the GPL v2 was.

So let's go back to the preamble, that provides motivations and some
guidance as to the interpretation of the legal text (i.e., the spirit
of the license):

  [...] the GNU General Public License is intended to guarantee your
  freedom to share and change free software--to make sure the software
  is free for all its users. [...]

  [...] Our General Public Licenses are designed to make sure that you have
  the freedom to distribute copies of free software (and charge for
  this service if you wish), that you receive source code or can get
  it if you want it, that you can change the software or use pieces of
  it in new free programs; and that you know you can do these things.

  To protect your rights, we need to make restrictions that forbid
  anyone to deny you these rights or to ask you to surrender the
  rights.  These restrictions translate to certain responsibilities
  for you if you distribute copies of the software, or if you modify
  it.

  [...] if you distribute copies of such a program, whether gratis or
  for a fee, you must give the recipients all the rights that you have


Can anyone show me how any of the provisions of GPLv3 fails to meet
this spirit?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Wed, 13 Jun 2007, Alexandre Oliva wrote:
> 
>   [...] Our General Public Licenses are designed to make sure that you have
>   the freedom to distribute copies of free software (and charge for
>   this service if you wish), that you receive source code or can get
>   it if you want it, that you can change the software or use pieces of
>   it in new free programs; and that you know you can do these things.
> 
>   To protect your rights, we need to make restrictions that forbid
>   anyone to deny you these rights or to ask you to surrender the
>   rights.  These restrictions translate to certain responsibilities
>   for you if you distribute copies of the software, or if you modify
>   it.
> 
>   [...] if you distribute copies of such a program, whether gratis or
>   for a fee, you must give the recipients all the rights that you have
> 
> 
> Can anyone show me how any of the provisions of GPLv3 fails to meet
> this spirit?

What kind of logic is that? It sounds like "Can you prove that God doesn't 
exist?"

The fact is, Tivo didn't take those rights away from you, yet the FSF says 
that what Tivo did was "against the spirit". That's *bullshit*.

So the whole "to protect these rights, we take away other rigths" argument 
hinges on the false premise that the new language in GPLv3 is somehow 
needed. It's not. You still had the right to distribute the software (and 
modify it), even if the *hardware* is limited to only one version.

In other words, GPLv3 restricts rights that do not need to be restricted, 
and yes, I think that violates the spirit of the GPLv2 preamble!

Think of it this way: what if the GPLv3 had an addition saying "You can 
not use this software to make a weapon". Do you see the problem? It 
restricts peoples rights, would you agree? Would you _also_ agree that it 
doesn't actually follow that "To protect your rights" logic AT ALL?

And this is exactly where the GPLv3 *diverges* from the above logic. If I 
build hardware, and sell it with software installed, you can still copy 
and modify the software. You may not do so within the confines of the 
hardware I built, but the hardware was never under the license in the 
first place.

In other words, GPLv3 *restricts* peoples freedoms more than it protects 
them. It does *not* cause any additional stated freedoms - quite the 
reverse. It tries to free up stuff that was never mentioned in the first 
place.

And then the FSF has the gall to call themselves the "protector of 
freedoms", and claim that everybody else is evil. What a crock. 

In other words, if you want to argue for the changes in GPLv3, you need to 
CHANGE THE PREAMBLE TOO! You should change:

	When we speak of free software, we are referring to freedom, not
	price.  Our General Public Licenses are designed to make sure that you
	have the freedom to distribute copies of free software (and charge for
	this service if you wish), that you receive source code or can get it
	if you want it, that you can change the software or use pieces of it
	in new free programs, that you can do so in place on your devices, 
        even if those devices weren't licensed under the GPL;  and that 
        you know you can do these things.

where I added the "that you can do so in place on your devices, even if 
those devices weren't licensed under the GPL".

That wasn't there in the original. Yet it's what the GPLv3 tries to shove 
down our throats in the name of "freedom".

I don't know if you've followed US politics very much over the last six 
years, but there's been a lot of "protecting our freedoms" going on. And 
it's been ugly. Maybe you could realize that sometimes "protecting your 
freedom" is *anything*but*!

		Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 13, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:

> On Wed, 13 Jun 2007, Alexandre Oliva wrote:
>> 
>> [...] Our General Public Licenses are designed to make sure that you have
>> the freedom to distribute copies of free software (and charge for
>> this service if you wish), that you receive source code or can get
>> it if you want it, that you can change the software or use pieces of
>> it in new free programs; and that you know you can do these things.
>> 
>> To protect your rights, we need to make restrictions that forbid
>> anyone to deny you these rights or to ask you to surrender the
>> rights.  These restrictions translate to certain responsibilities
>> for you if you distribute copies of the software, or if you modify
>> it.
>> 
>> [...] if you distribute copies of such a program, whether gratis or
>> for a fee, you must give the recipients all the rights that you have
>> 
>> 
>> Can anyone show me how any of the provisions of GPLv3 fails to meet
>> this spirit?

> What kind of logic is that? It sounds like "Can you prove that God doesn't 
> exist?"

By this reasoning, it sounds like you've been claiming that "God does
exist", even though you can't prove it.

It shouldn't be anywhere that difficult to show that the GPLv3 fails
to meet the spirit of the GPLs.  You just have to show a single
counter-example.  Since there are so many objections to the changes,
it shouldn't be that hard.  Can you at least try?

> The fact is, Tivo didn't take those rights away from you, yet the FSF says 
> that what Tivo did was "against the spirit". That's *bullshit*.

Oh, good, let's take this one.

  if you distribute copies of such a program, [...]
  you must give the recipients all the rights that you have

So, TiVo includes a copy of Linux in its DVR.  

TiVo retains the right to modify that copy of Linux as it sees fit.

It doesn't give the recipients the same right.

Oops.

Sounds like a violation of the spirit to me.

Sounds like plugging this hole would retain the same spirit.

> In other words, GPLv3 restricts rights that do not need to be restricted, 

That's correct.  They don't need to be restricted.  The whole idea of
copyleft, implemented through the GPL, is not based on needs, but
rather on the wish to defend the freedoms established in the preamble
from those who would rather not respect them.

Do you deny that TiVo prevents you (or at least a random customer)
from modifying the copy of Linux that they ship in their DVR?

Do you deny that they can still do it themselves?

> Think of it this way: what if the GPLv3 had an addition saying "You can 
> not use this software to make a weapon".

This would make GPLv3 a non-Free Software license.

But the GPLv3 last call draft doesn't say anything along these lines.

You can use the software as much as you like, even in DVRs, and even
to implement DRM in them, as long as you respect the users' freedoms
to change and share the software.  Per the GPLv3 (paraphrased), if it
is possible to install modified versions of the covered program in the
device, you must tell the recipient how to do it.  Otherwise, the
freedom to modify the program is being too severely limited.

And, in the particular case of TiVo, it's a case of distributing
incomplete source code, of refraining from including functional
portions of the source code.

> In other words, GPLv3 *restricts* peoples freedoms more than it
> protects them.

While you look at it from the point of view of TiVo, who wants to be
free to prohibit people from modifying the workings of the device it
sells while it can still modify it itself, and it does that in order
to prohibit people from removing locks that stop them from doing
things they're legally entitled to do, I see a lot more prohibitions
than freedoms in what TiVo does.  I don't understand why you'd stand
up for it.  Is it more important that a single company be allowed to
impose prohibitions on others in order for its business model to work,
than to maintain the spirit of hacking and sharing that enabled Free
Software and Linux to flourish?

Do you expect Linux would have flourished if computers had locks that
stopped people from modifying Linux in them?

> where I added the "that you can do so in place on your devices, even if 
> those devices weren't licensed under the GPL".

You're mistakenly focusing on the device.  As you say, the device is
not under the license.

What's under the license is the software in it.  And that license
spirit requires the distributor to pass on the right to modify the
software.

> I don't know if you've followed US politics very much over the last
> six years, but there's been a lot of "protecting our freedoms" going
> on. And it's been ugly. Maybe you could realize that sometimes
> "protecting your freedom" is *anything*but*!

Is this why you're overreacting?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bongani Hlope]

On Thursday 14 June 2007 01:49:23 Alexandre Oliva wrote:
> On Jun 13, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:
> > The fact is, Tivo didn't take those rights away from you, yet the FSF
> > says that what Tivo did was "against the spirit". That's *bullshit*.
>
> Oh, good, let's take this one.
>
>   if you distribute copies of such a program, [...]
>   you must give the recipients all the rights that you have
>
> So, TiVo includes a copy of Linux in its DVR.
>

And they give you the same right that they had, which is obtain free software 
that you can modify and redistribute. There's nothing in there that says they 
should give you the tools they used after they received the software, which 
is what you seem to be looking for.

> TiVo retains the right to modify that copy of Linux as it sees fit.
>
> It doesn't give the recipients the same right.
>

It does, can't you modify their kernel source? Where does it say you should be 
able to run you modifications on the same hardware?

> Oops.
>
> Sounds like a violation of the spirit to me.
>
> Sounds like plugging this hole would retain the same spirit.

The only fear that I have with the whole Tivo saga, is that companies like 
Dell can use the same thing to say: "Our hardware will only run Company's X 
distribution of Linux". 

Do we just hope users won't buy those Dell machines, or do you modify your 
software license to force Dell to allow custom distributions to run on their 
machines? Then where do we draw the line of "Software Licenses".

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Adrian Bunk]

On Thu, Jun 14, 2007 at 02:15:53AM +0200, Bongani Hlope wrote:
> On Thursday 14 June 2007 01:49:23 Alexandre Oliva wrote:
> > On Jun 13, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:
> > > The fact is, Tivo didn't take those rights away from you, yet the FSF
> > > says that what Tivo did was "against the spirit". That's *bullshit*.
> >
> > Oh, good, let's take this one.
> >
> >   if you distribute copies of such a program, [...]
> >   you must give the recipients all the rights that you have
> >
> > So, TiVo includes a copy of Linux in its DVR.
> 
> And they give you the same right that they had, which is obtain free software 
> that you can modify and redistribute. There's nothing in there that says they 
> should give you the tools they used after they received the software, which 
> is what you seem to be looking for.
>...

Wrong, the GPLv2 says:

"For an executable work, complete source code means all the source code 
for all modules it contains, plus any associated interface definition 
files, plus the scripts used to control compilation and installation of 
the executable."

The question is whether this includes private keys.
Different people have different opinions regarding this issue.

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 13, 2007, Bongani Hlope <bhlope@mweb.co.za> wrote:

> On Thursday 14 June 2007 01:49:23 Alexandre Oliva wrote:

>> if you distribute copies of such a program, [...]
>> you must give the recipients all the rights that you have

>> So, TiVo includes a copy of Linux in its DVR.

> And they give you the same right that they had, which is obtain free software 
> that you can modify and redistribute. There's nothing in there that says they 
> should give you the tools they used after they received the software, which 
> is what you seem to be looking for.

Can they modify the software in their device?

Do they pass this right on?

>> TiVo retains the right to modify that copy of Linux as it sees fit.

>> It doesn't give the recipients the same right.

> It does, can't you modify their kernel source?

It's not the kernel source.  That's not where the TiVo anti-tampering
machinery blocks modifications.

It's about that copy of the kernel that ships in the device in object
code.  That's the one that TiVo customers ought to be entitled to
modify, if TiVo can modify it itself.

> Where does it say you should be able to run you modifications on the
> same hardware?

Where it says that you should pass on all the rights that you have.

While TiVo retains the ability to replace, upgrade, fix, break or make
any other change in the GPLed software in the device, it ought to pass
it on to its customers.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Wednesday 13 June 2007 20:55:52 Alexandre Oliva wrote:
> On Jun 13, 2007, Bongani Hlope <bhlope@mweb.co.za> wrote:
> > On Thursday 14 June 2007 01:49:23 Alexandre Oliva wrote:
> >> if you distribute copies of such a program, [...]
> >> you must give the recipients all the rights that you have
> >>
> >> So, TiVo includes a copy of Linux in its DVR.
> >
> > And they give you the same right that they had, which is obtain free
> > software that you can modify and redistribute. There's nothing in there
> > that says they should give you the tools they used after they received
> > the software, which is what you seem to be looking for.
>
> Can they modify the software in their device?
>
> Do they pass this right on?

But this *ISN'T* a right that the GPLv2 *REQUIRES* be passed on. 

> >> TiVo retains the right to modify that copy of Linux as it sees fit.
> >>
> >> It doesn't give the recipients the same right.
> >
> > It does, can't you modify their kernel source?
>
> It's not the kernel source.  That's not where the TiVo anti-tampering
> machinery blocks modifications.
>
> It's about that copy of the kernel that ships in the device in object
> code.  That's the one that TiVo customers ought to be entitled to
> modify, if TiVo can modify it itself.

The GPLv2 makes no real provision for *DIRECTLY* modifying object code. What 
provisions the GPLv2 has apply to the source code.

And no, the end user *SHOULD* *NOT* be entitled to run whatever kernel they 
like on a TiVO. It was designed with the "install new kernel" functionality 
so that the TiVO corporation could update the kernel running on the hardware 
when security problems came up, when bugs were fixed or even when the new 
version gives better performance.

> > Where does it say you should be able to run you modifications on the
> > same hardware?
>
> Where it says that you should pass on all the rights that you have.
>
> While TiVo retains the ability to replace, upgrade, fix, break or make
> any other change in the GPLed software in the device, it ought to pass
> it on to its customers.

It *DOES* *NOT* say "All rights that you have". It says "All rights that are 
granted you by this license". If every piece of software released under the 
GPL had *ALL* rights passed on, then *ANYONE* could do the "I'm granting 
company X the right to use this software outside the GPL for $50,000USD." 
instead of just the *creator* of the software.

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 13, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> On Wednesday 13 June 2007 20:55:52 Alexandre Oliva wrote:
>> On Jun 13, 2007, Bongani Hlope <bhlope@mweb.co.za> wrote:
>> > On Thursday 14 June 2007 01:49:23 Alexandre Oliva wrote:
>> >> if you distribute copies of such a program, [...]
>> >> you must give the recipients all the rights that you have

>> >> So, TiVo includes a copy of Linux in its DVR.

>> Can they modify the software in their device?

>> Do they pass this right on?

> But this *ISN'T* a right that the GPLv2 *REQUIRES* be passed on. 

You may be right.  The spirit says it should, but the legalese may
have missed the mark.  So GPLv3 makes it clear that it is.

>> It's about that copy of the kernel that ships in the device in object
>> code.  That's the one that TiVo customers ought to be entitled to
>> modify, if TiVo can modify it itself.

> The GPLv2 makes no real provision for *DIRECTLY* modifying object
> code.

Sure.  And that's not what I'm talking about.

What I'm talking about is being able to replace, upgrade, fix, tweak,
hack, and otherwise modify the program on the machine in the same way
that the vendor still can.

> What provisions the GPLv2 has apply to the source code.

This is too narrow a view of the GPLv2 provisions.

> And no, the end user *SHOULD* *NOT* be entitled to run whatever kernel they 
> like on a TiVO. It was designed with the "install new kernel" functionality 
> so that the TiVO corporation could update the kernel running on the hardware 
> when security problems came up, when bugs were fixed or even when the new 
> version gives better performance.

I.e., it was designed such that TiVo could modify the installed
kernel, but the user couldn't.  That's an outright violation of the
spirit of the GPL.

>> > Where does it say you should be able to run you modifications on the
>> > same hardware?

>> Where it says that you should pass on all the rights that you have.

>> While TiVo retains the ability to replace, upgrade, fix, break or make
>> any other change in the GPLed software in the device, it ought to pass
>> it on to its customers.

> It *DOES* *NOT* say "All rights that you have". It says "All rights
> that are granted you by this license".

I suggest you to reboot into memtest ;-)  The preamble of GPLv2 says:

  For example, if you distribute copies of such a program, whether
  gratis or for a fee, you must give the recipients
  all the rights that you have.
  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
 
> If every piece of software released under the GPL had *ALL* rights
> passed on, then *ANYONE* could do the "I'm granting company X the
> right to use this software outside the GPL for $50,000USD."

The requirement above applies to licensees, not to the licensor.  The
licensor doesn't have to pass on all the rights s/he has, s/he only
decides to respect the licensee's freedoms, conditioned to the respect
of others' freedoms by means of passing on all rights the licensee has
over the software.

Arguably, one could use this argument to state that any authors of
derived works ought to pass on the right to choose the license for the
derived work under the GPL, but since (a) the above is not in the
legal terms, and (b) the downstream recipients would be bound by the
terms of the GPL anyway, and that requires the use of the GPL itself,
this would make no difference whatsoever.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Dmitry Torokhov]

On Wednesday 13 June 2007 21:59, Alexandre Oliva wrote:
> > It *DOES* *NOT* say "All rights that you have". It says "All rights
> > that are granted you by this license".
> 
> I suggest you to reboot into memtest ;-)  The preamble of GPLv2 says:
> 
>   For example, if you distribute copies of such a program, whether
>   gratis or for a fee, you must give the recipients
>   all the rights that you have.
>   ^^^^^^^^^^^^^^^^^^^^^^^^^^^^

So if I am a sole author of a program and I chose to distribute it under
GPL then all recepients will get _all_ my rights, including right to
re-license the program under BSD or a proprietory license? Yeah, riiight...

Thankfully it is just preamble and not the actual license text.
 
-- 
Dmitry

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Dmitry Torokhov <dtor@insightbb.com> wrote:

> On Wednesday 13 June 2007 21:59, Alexandre Oliva wrote:
>>   For example, if you distribute copies of such a program, whether
>>   gratis or for a fee, you must give the recipients
>>   all the rights that you have.
>>   ^^^^^^^^^^^^^^^^^^^^^^^^^^^^

> So if I am a sole author of a program and I chose to distribute it under
> GPL

then you're not a licensee, you're a licensor, and these terms don't
apply to you.  Already covered upthread BTW.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Dmitry Torokhov]

On 6/14/07, Alexandre Oliva <aoliva@redhat.com> wrote:
> On Jun 14, 2007, Dmitry Torokhov <dtor@insightbb.com> wrote:
>
> > On Wednesday 13 June 2007 21:59, Alexandre Oliva wrote:
> >> For example, if you distribute copies of such a program, whether
> >> gratis or for a fee, you must give the recipients
> >> all the rights that you have.
> >> ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
>
> > So if I am a sole author of a program and I chose to distribute it under
> > GPL
>
> then you're not a licensee, you're a licensor, and these terms don't
> apply to you.

Heh. When you change a GPLed program and pass your changes you are the
licensor for the new code. You still have a right and license pieces
of the code you wrote under different license but you do not pass that
right to recepient of modified work.

-- 
Dmitry

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, "Dmitry Torokhov" <dmitry.torokhov@gmail.com> wrote:

> On 6/14/07, Alexandre Oliva <aoliva@redhat.com> wrote:
>> On Jun 14, 2007, Dmitry Torokhov <dtor@insightbb.com> wrote:
>> 
>> > On Wednesday 13 June 2007 21:59, Alexandre Oliva wrote:
>> >> For example, if you distribute copies of such a program, whether
>> >> gratis or for a fee, you must give the recipients
>> >> all the rights that you have.
>> >> ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
>> 
>> > So if I am a sole author of a program and I chose to distribute it under
>> > GPL
>> 
>> then you're not a licensee, you're a licensor, and these terms don't
>> apply to you.

> Heh. When you change a GPLed program and pass your changes you are the
> licensor for the new code. You still have a right and license pieces
> of the code you wrote under different license but you do not pass that
> right to recepient of modified work.

You are the author of the change, and you can license them however you
like.  Your change itself is not bound by the terms of the GPL, it is
only if it is a derived work of the GPLed work.

If your change is not a derived work, you're not bound by the terms of
the GPL as far as the change is concerned, so the GPL has no say
whatsoever as to how you must release it.  If you choose the GPL, then
you're a licensor, and the requirements to pass on all the rights you
have do not apply.

If it *is* a derived work, then you're constrained by the terms of the
license, and you can only distribute it under the same license.  You
don't have a right to offer it under a different license in the first
place, so you can't pass this right on.

Derived work or not, when you combine that change with the program,
then you're bound by the terms of the license, and then you cannot
change the licensing terms of the whole program, so you can't pass
this right on either.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Dmitry Torokhov]

On 6/14/07, Alexandre Oliva <aoliva@redhat.com> wrote:
> On Jun 14, 2007, "Dmitry Torokhov" <dmitry.torokhov@gmail.com> wrote:
>
> > On 6/14/07, Alexandre Oliva <aoliva@redhat.com> wrote:
> >> On Jun 14, 2007, Dmitry Torokhov <dtor@insightbb.com> wrote:
> >>
> >> > On Wednesday 13 June 2007 21:59, Alexandre Oliva wrote:
> >> >> For example, if you distribute copies of such a program, whether
> >> >> gratis or for a fee, you must give the recipients
> >> >> all the rights that you have.
> >> >> ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> >>
> >> > So if I am a sole author of a program and I chose to distribute it under
> >> > GPL
> >>
> >> then you're not a licensee, you're a licensor, and these terms don't
> >> apply to you.
>
> > Heh. When you change a GPLed program and pass your changes you are the
> > licensor for the new code. You still have a right and license pieces
> > of the code you wrote under different license but you do not pass that
> > right to recepient of modified work.
>
> You are the author of the change, and you can license them however you
> like.
[... skip...]
>
> Derived work or not, when you combine that change with the program,
> then you're bound by the terms of the license, and then you cannot
> change the licensing terms of the whole program, so you can't pass
> this right on either.
>

Ok, consider non-derived work. Because I am distributing whole program
I have to do it under GPL. However I still have the right to
distribute just the portion that is written by me under whatevel
license I want but you as a recepient of GPLed whole do not get this
right. IOW I am not passing all the rights _I have_.

-- 
Dmitry

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, "Dmitry Torokhov" <dmitry.torokhov@gmail.com> wrote:

> Ok, consider non-derived work.

I did, you snipped it out:

>> If your change is not a derived work, you're not bound by the terms
>> of the GPL as far as the change is concerned, so the GPL has no say
>> whatsoever as to how you must release it.  If you choose the GPL,
>> then you're a licensor, and the requirements to pass on all the
>> rights you have do not apply.

> Because I am distributing whole program
> I have to do it under GPL. However I still have the right to
> distribute just the portion that is written by me under whatevel
> license I want but you as a recepient of GPLed whole do not get this
> right. IOW I am not passing all the rights _I have_.

I see what you mean.  IANAL, but I don't think that's how it works.

When your work is not a derived work, the GPL that applies to the rest
of the program does not make you a licensee, and it only covers your
work if you choose to license it that way.  And then, you're the sole
licensor of that piece of the work.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Dmitry Torokhov]

On 6/14/07, Alexandre Oliva <aoliva@redhat.com> wrote:
> On Jun 14, 2007, "Dmitry Torokhov" <dmitry.torokhov@gmail.com> wrote:
>
> > Ok, consider non-derived work.
>
> I did, you snipped it out:
>
> >> If your change is not a derived work, you're not bound by the terms
> >> of the GPL as far as the change is concerned, so the GPL has no say
> >> whatsoever as to how you must release it.  If you choose the GPL,
> >> then you're a licensor, and the requirements to pass on all the
> >> rights you have do not apply.
>

Yes, I did, thank you for putting the text back in.

> > Because I am distributing whole program
> > I have to do it under GPL.

Please notice this sentence. GPL still influences the way I release
stuff (if I want to release the work as whole) but it does not mean
passing all rigths I could possibly have.

> > However I still have the right to
> > distribute just the portion that is written by me under whatevel
> > license I want but you as a recepient of GPLed whole do not get this
> > right. IOW I am not passing all the rights _I have_.
>
> I see what you mean.  IANAL, but I don't think that's how it works.
>
> When your work is not a derived work, the GPL that applies to the rest
> of the program does not make you a licensee, and it only covers your
> work if you choose to license it that way.  And then, you're the sole
> licensor of that piece of the work.

So, with regard to TIVO, why are you saying that GPL shoudl affect
their hardware (I assume that key check/enforce is done in firmware
taht is separate from kernel image)?

-- 
Dmitry

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, "Dmitry Torokhov" <dmitry.torokhov@gmail.com> wrote:

> So, with regard to TIVO, why are you saying that GPL shoudl affect
> their hardware

I'm not.

I'm just saying that TiVO, as a licensee of Linux, agreed that it
wouldn't impose further restrictions on recipients of Linux on the
exercise of the rights granted by the license.

So, just like it couldn't use a patent to stop people from modifying
or sharing Linux, it can't use the hardware to do that.

And if they fail to supply portions of the functional source code in
order to prevent modified versions to run, they are infringing the
spirit and quite possibly the letter of the license.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Thu, 14 Jun 2007, Alexandre Oliva wrote:
> 
> I see what you mean.  IANAL, but I don't think that's how it works.

Why the hell do you keep saying that?

There *are* lawyers who have said that what Tivo did was legal. They were 
the FSF's own lawyers. So now you're saying "I am not a lawyer, but that's 
not right".

So you're trying to state some legal point, admitting that you're not a 
lawyer, and admitting that actual real-life lawyers disagree with you?

So please explain why the *hell* you would expect us to take your points 
seriously?

		Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:

> On Thu, 14 Jun 2007, Alexandre Oliva wrote:
>> 
>> I see what you mean.  IANAL, but I don't think that's how it works.

> There *are* lawyers who have said that what Tivo did was legal.

What I wrote above had ZERO to do with TiVO.  Please re-read the
message you responded to, and the two previous messages in that
sub-thread for the context you snipped out.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bongani Hlope]

On Thursday 14 June 2007 02:55:52 Alexandre Oliva wrote:
> On Jun 13, 2007, Bongani Hlope <bhlope@mweb.co.za> wrote:
> > On Thursday 14 June 2007 01:49:23 Alexandre Oliva wrote:
> >> if you distribute copies of such a program, [...]
> >> you must give the recipients all the rights that you have
> >>
> >> So, TiVo includes a copy of Linux in its DVR.
> >
> > And they give you the same right that they had, which is obtain free
> > software that you can modify and redistribute. There's nothing in there
> > that says they should give you the tools they used after they received
> > the software, which is what you seem to be looking for.
>
> Can they modify the software in their device?
>
> Do they pass this right on?
>
> >> TiVo retains the right to modify that copy of Linux as it sees fit.
> >>
> >> It doesn't give the recipients the same right.
> >
> > It does, can't you modify their kernel source?
>
> It's not the kernel source.  That's not where the TiVo anti-tampering
> machinery blocks modifications.
>
> It's about that copy of the kernel that ships in the device in object
> code.  That's the one that TiVo customers ought to be entitled to
> modify, if TiVo can modify it itself.
>
> > Where does it say you should be able to run you modifications on the
> > same hardware?
>
> Where it says that you should pass on all the rights that you have.
>
> While TiVo retains the ability to replace, upgrade, fix, break or make
> any other change in the GPLed software in the device, it ought to pass
> it on to its customers.

So according to your logic, I can go to Sharp's website and download the GPL 
source code for their Zaurus. But I don't own a Sharp Zaurus; to keep with 
your interpretation of the spirit of GPL, they have to give me a Zaurus so 
that I can run my modifications on the same hardware?

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Bongani Hlope <bhlope@mweb.co.za> wrote:

> On Thursday 14 June 2007 02:55:52 Alexandre Oliva wrote:

>> While TiVo retains the ability to replace, upgrade, fix, break or make
>> any other change in the GPLed software in the device, it ought to pass
>> it on to its customers.

> So according to your logic, I can go to Sharp's website and download the GPL 
> source code for their Zaurus. But I don't own a Sharp Zaurus; to keep with 
> your interpretation of the spirit of GPL, they have to give me a Zaurus so 
> that I can run my modifications on the same hardware?

Sharp can modify the copy of the code in your Zaurus as much as you
do, when you don't have a Zaurus.  I don't see how you can get to the
conclusion that they have to give you a Zaurus, when all you're
getting is software.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Michael Gerdau]

[-- Attachment #1: Type: text/plain, Size: 2704 bytes --]

> > > The fact is, Tivo didn't take those rights away from you, yet the FSF
> > > says that what Tivo did was "against the spirit". That's *bullshit*.
> >
> > Oh, good, let's take this one.
> >
> >   if you distribute copies of such a program, [...]
> >   you must give the recipients all the rights that you have
> >
> > So, TiVo includes a copy of Linux in its DVR.
> >
> 
> And they give you the same right that they had, which is obtain free software 
> that you can modify and redistribute. There's nothing in there that says they 
> should give you the tools they used after they received the software, which 
> is what you seem to be looking for.

IANAL so I won't comment on the legal aspects of TiVo's doing.

However it definitely is against _MY_ understanding of the spirit
of the GPL. At least to me that's quite obvious. I'm sure you all
know the story of the printer driver RMS couldn't fix that reportedly
made him start the whole FSF business.

Looking at what TiVo did I realize glaring similarities.

<disclaimer>
I'm in no way related with the FSF. I hereby state I'm not parroting
anyone's else position but have come to this conclusion solely on
my own.
</disclaimer>

> > TiVo retains the right to modify that copy of Linux as it sees fit.
> >
> > It doesn't give the recipients the same right.
> 
> It does, can't you modify their kernel source? Where does it say you should be 
> able to run you modifications on the same hardware?

Come on! The whole idea of software is to have it run on some HW.
Why would I want to change it in the first place if I can't run it ?

If what they did is actually allowed by the wording of the legal phrases
of the GPLv2 then that IMO is a loophole w/r to the spirit (as I understand)
it and IMO should be plugged.

> The only fear that I have with the whole Tivo saga, is that companies like 
> Dell can use the same thing to say: "Our hardware will only run Company's X 
> distribution of Linux". 

Would not such a restriction voilate the spirit of the GPL ?

Anyway, my simplistic view is:
Once it is under the GPL I could change it and actually make the
changes work as I see fit.

That's what I think my freedom as of the GPL is about.

Now all that needs to be done is make sure the legal phrases are such
that they convince the judges they actually mean this in court too.

Best wishes,
Michael
-- 
 Technosis GmbH, Geschäftsführer: Michael Gerdau, Tobias Dittmar
 Sitz Hamburg; HRB 89145 Amtsgericht Hamburg
 Vote against SPAM - see http://www.politik-digital.de/spam/
 Michael Gerdau       email: mgd@technosis.de
 GPG-keys available on request or at public keyserver

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Al Viro]

On Thu, Jun 14, 2007 at 06:44:25AM +0200, Michael Gerdau wrote:
> > able to run you modifications on the same hardware?
                                         ^^^^
> Come on! The whole idea of software is to have it run on some HW.
                                                           ^^^^
> Why would I want to change it in the first place if I can't run it ?


See the difference?

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Michael Gerdau]

[-- Attachment #1: Type: text/plain, Size: 2170 bytes --]

> > > able to run you modifications on the same hardware?
>                                          ^^^^
> > Come on! The whole idea of software is to have it run on some HW.
>                                                            ^^^^
> > Why would I want to change it in the first place if I can't run it ?
> 
> See the difference?

Forgive my poor mastery of the english language and me letting slip
this inconsistency.

The first sentence you cited was a general remark IMO valid outside
of this context and possibly ill placed as it was.

The second sentence pertains the key msg I was trying to deliver and
apparently I did a poor job in phrasing it so let me redo it:
Why would I want to change the SW targetted for some HW if I can't run
the changed version on said HW ?

[note that for the TiVo case I possibly would not own or be able to own
similar HW being able to run my modified SW; so even some HW would not
be triggered either]                                 ^^^^

Remember I'm discussing my understanding of the spirit of the GPL,
not whether the legal part actually does give me that right enforceable
in court.

Here is another stmt which is valid outside of this context AFAIAC:
If the GPLv2 does not legally give me the right that I think its spirit
gives me then the legal phrases should be changed to achieve that.

Whether or not others share my view of what the spirit of the GPL
implies is completely theirs to decide and if they differ they likely
won't agree on my previous stmt either. Fine with me.

And this leads to another observation:
IMO this thread is partly fueled by a fundamental mixing of PoVs.
Some argue based on their perceived view of the spirit of the GPL
and some based on the actual legal phrases in GPLv2 and GPLv3 and
whether or how they reflect the perceived spirit.

Best wishes,
Michael
-- 
 Technosis GmbH, Geschäftsführer: Michael Gerdau, Tobias Dittmar
 Sitz Hamburg; HRB 89145 Amtsgericht Hamburg
 Vote against SPAM - see http://www.politik-digital.de/spam/
 Michael Gerdau       email: mgd@technosis.de
 GPG-keys available on request or at public keyserver

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Valdis.Kletnieks]

[-- Attachment #1: Type: text/plain, Size: 1160 bytes --]

On Thu, 14 Jun 2007 06:44:25 +0200, Michael Gerdau said:

> > > TiVo retains the right to modify that copy of Linux as it sees fit.
> > >
> > > It doesn't give the recipients the same right.
> > It does, can't you modify their kernel source? Where does it say you should be
> > able to run you modifications on the same hardware?
> 
> Come on! The whole idea of software is to have it run on some HW.
> Why would I want to change it in the first place if I can't run it ?

Maybe this quote will summarize the situation:

Judith: [on Stan's desire to be a mother] Here! I've got an idea: Suppose you
	agree that he can't actually have babies, not having a womb - which is
	nobody's fault, not even the Romans' - but that he can have the *right*
	to have babies.
Francis: Good idea, Judith. We shall fight the oppressors for your right to
	have babies, brother... sister, sorry.
Reg: What's the *point*?
Francis: What?
Reg: What's the point of fighting for his right to have babies, when he can't
	have babies?
Francis: It is symbolic of our struggle against oppression.
Reg: It's symbolic of his struggle against reality.

		-- Monty Python's "Life of Brian"



[-- Attachment #2: Type: application/pgp-signature, Size: 226 bytes --]

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Wednesday 13 June 2007 19:49:23 Alexandre Oliva wrote:
<snip>
> > The fact is, Tivo didn't take those rights away from you, yet the FSF
> > says that what Tivo did was "against the spirit". That's *bullshit*.
>
> Oh, good, let's take this one.
>
>   if you distribute copies of such a program, [...]
>   you must give the recipients all the rights that you have
>
> So, TiVo includes a copy of Linux in its DVR.
>
> TiVo retains the right to modify that copy of Linux as it sees fit.
>
> It doesn't give the recipients the same right.
>
> Oops.
>
> Sounds like a violation of the spirit to me.
>
> Sounds like plugging this hole would retain the same spirit.

Are you an idiot, or do you just choose to ignore all proof that doesn't fit 
your preconceived beliefs? TiVO gives you every right to the Linux kernel 
that they recieved. What they don't give you the right to do is use modified 
versions on their *HARDWARE* - which they have *NEVER* given you any rights 
to, except for "normal use". (And no, it isn't legal to put those 200G hard 
drives in your TiVO no matter what you think) 

> > In other words, GPLv3 restricts rights that do not need to be restricted,
>
> That's correct.  They don't need to be restricted.  The whole idea of
> copyleft, implemented through the GPL, is not based on needs, but
> rather on the wish to defend the freedoms established in the preamble
> from those who would rather not respect them.
>
> Do you deny that TiVo prevents you (or at least a random customer)
> from modifying the copy of Linux that they ship in their DVR?

Exactly. They don't. What TiVO prevents is using that modified version on 
their hardware. And they have that right, because the Hardware *ISN'T* 
covered by the GPL.

Do you understand that, or do I need to break out the finger-puppets next ?

> Do you deny that they can still do it themselves?
>
> > Think of it this way: what if the GPLv3 had an addition saying "You can
> > not use this software to make a weapon".
>
> This would make GPLv3 a non-Free Software license.
>
> But the GPLv3 last call draft doesn't say anything along these lines.
>
> You can use the software as much as you like, even in DVRs, and even
> to implement DRM in them, as long as you respect the users' freedoms
> to change and share the software.  Per the GPLv3 (paraphrased), if it
> is possible to install modified versions of the covered program in the
> device, you must tell the recipient how to do it.  Otherwise, the
> freedom to modify the program is being too severely limited.

And this unnaturally restricts the freedom of hardware manufacturers. If they 
add a custom, internal connector so a repair shop can restore the hardware to 
its *FACTORY* state then it is "possible to install modified versions", 
provided the person doing it has the specialized hardware needed.

And this is what the FSF, RMS and yes, *YOU*, Alexandre, fail to realize - the 
GPL covers *ONLY* the software. It has *ZERO* legal standing when applied to 
hardware. Not even the most draconian of MS EULA's tries to apply itself to 
the hardware.

In the case of 99% of the hardware targeted by the clause of the GPLv3 you 
elucidate on, the "ability to install modified versions of the software" was 
*NOT* intended for that use, nor was it intended for *ANYONE* *EXCEPT* 
trained service personell to have *ACCESS* to that functionality. Arguing 
otherwise is just idiotic - I have never found a piece of "high tech" 
hardware (like a TiVO) that was designed for the end-user to modify. (yes, 
installing a new version of the linux kernel is "modifying" the system)

> And, in the particular case of TiVo, it's a case of distributing
> incomplete source code, of refraining from including functional
> portions of the source code.

And? They distribute the kernel source - as they recieved it - in compliance 
with the GPL. Their additions - whether they be "modules" or just the UI - do 
not, necessarily, fall under the GPL. (Yes, there have been discussions about 
whether a kernel module is a derived work, but most of the time those 
discussions ended "Legally they aren't, even though I feel they should be")

> > In other words, GPLv3 *restricts* peoples freedoms more than it
> > protects them.
>
> While you look at it from the point of view of TiVo, who wants to be
> free to prohibit people from modifying the workings of the device it
> sells while it can still modify it itself, and it does that in order
> to prohibit people from removing locks that stop them from doing
> things they're legally entitled to do, I see a lot more prohibitions
> than freedoms in what TiVo does.  I don't understand why you'd stand
> up for it.  Is it more important that a single company be allowed to
> impose prohibitions on others in order for its business model to work,
> than to maintain the spirit of hacking and sharing that enabled Free
> Software and Linux to flourish?

What "Legally Entitled" things?

And... You do realize that almost every difference between the GPLv2 and the 
GPLv3 is going to cause a hell of a lot of problems? The fact that the GPLv3 
is designed to prevent things that RMS *PERSONALLY* finds distasteful - DRM 
and the like - is a big turn-off for a *LOT* of people. (Personally I don't 
like *ANY* version of the GPL, because there are chunks I have problems with)

> Do you expect Linux would have flourished if computers had locks that
> stopped people from modifying Linux in them?

But you aren't talking about a "computer" here. You're talking about a 
mass-market device that must comply with both US and International copyright 
law - and that's just a TiVO. Other devices have other laws they have to 
comply with - in the US the FCC's regulations control all radio devices, so 
if you upload a modified linux kernel to your wireless router that gives it a 
2000 foot range, you've just broken the law *AND* violated the license on the 
hardware which states that you "won't modify it or the controlling 
software" - in most cases "the controlling software" is just the firmware, 
but with modern wireless hardware, the firmware is loaded by the OS.

> > where I added the "that you can do so in place on your devices, even if
> > those devices weren't licensed under the GPL".
>
> You're mistakenly focusing on the device.  As you say, the device is
> not under the license.

But he isn't. The GPLv3 says, and I'll quote you here - "Per the GPLv3 
(paraphrased), if it is possible to install modified versions of the covered 
program in the device, you must tell the recipient how to do it."

>From the latest version of the GPLv3:
"Installation Information" for a User Product means any methods, procedures, 
authorization keys, or other information required to install and execute 
modified versions of a covered work in that User Product from a modified 
version of its Corresponding Source. The information must suffice to ensure 
that the continued functioning of the modified object code is in no case 
prevented or interfered with solely because modification has been made.

and:

If you convey an object code work under this section in, or with, or 
specifically for use in, a User Product, and the conveying occurs as part of 
a transaction in which the right of possession and use of the User Product is 
transferred to the recipient in perpetuity or for a fixed term (regardless of 
how the transaction is characterized), the Corresponding Source conveyed 
under this section must be accompanied by the Installation Information. But 
this requirement does not apply if neither you nor any third party retains 
the ability to install modified object code on the User Product (for example, 
the work has been installed in ROM).

So it's not just a designed in ability to run modified code - ie: running of 
modified code is a feature meant for the user to take advantage of - but even 
things like the connectors used to upload the operating software at the 
factory that people now cannot have in a device that runs GPL(v3) covered 
software unless they ship the related "Installation Information". In other 
words, companies are no longer allowed to have a completely separate license  
for hardware and software.

That, to me, reads like RMS got mad about TiVO and said "I don't like it, lets 
add a clause making it wrong to the next GPL". Hell, that *IS* what happened, 
and nothing the FSF or Eben Moglen says will convince me otherwise. It's the 
same for the bits that were added after Novell signed their agreement with 
MS. 

> What's under the license is the software in it.  And that license
> spirit requires the distributor to pass on the right to modify the
> software.

And since when did they have to enable people to use their hardware in 
violation of the licensing agreement they implicitly agree to when opening 
the package?

There is *NOTHING* stopping you from doing whatever you want with the code 
that runs on a TiVO (or any similar device). You (and everyone that thinks 
like you) are confusing a want to use the *HARDWARE* however you want with 
your GPL granted "right" to do what you want with the *SOFTWARE*.

> > I don't know if you've followed US politics very much over the last
> > six years, but there's been a lot of "protecting our freedoms" going
> > on. And it's been ugly. Maybe you could realize that sometimes
> > "protecting your freedom" is *anything*but*!
>
> Is this why you're overreacting?

No, he's making a point. RMS and the FSF, in drafting GPLv3 to include the 
language and clauses it does, is "protecting your freedom" the way a lot of 
the post 9/11 changes to the US Federal code does it. (ie: by saying "no, you 
can't do that anymore")

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 13, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> On Wednesday 13 June 2007 19:49:23 Alexandre Oliva wrote:

> Exactly. They don't. What TiVO prevents is using that modified version on 
> their hardware. And they have that right, because the Hardware *ISN'T* 
> covered by the GPL.

Indeed, TiVO has this legal right.  But then they must not use
software under the GPLv3 in it.  And, arguably, they must not use
software under the GPLv2 either.

> In the case of 99% of the hardware targeted by the clause of the GPLv3 you 
> elucidate on, the "ability to install modified versions of the software" was 
> *NOT* intended for that use, nor was it intended for *ANYONE* *EXCEPT* 
> trained service personell to have *ACCESS* to that functionality. Arguing 
> otherwise is just idiotic - I have never found a piece of "high tech" 
> hardware (like a TiVO) that was designed for the end-user to modify. (yes, 
> installing a new version of the linux kernel is "modifying" the system)

It's about time for a change for better, wouldn't you think?

In 95% of the desktop computers, you can't make changes to the OS that
runs on it.  Whom is this good for?

> And? They distribute the kernel source - as they recieved it - in
> compliance with the GPL.

This makes it seem like you think that passing on the source code is
enough to comply with the GPL.  Check your assumptions.  It's not.

>> to prohibit people from removing locks that stop them from doing
>> things they're legally entitled to do

> What "Legally Entitled" things?

Time shifting of any shows, creating copies of shows for personal use,
letting others do so.  Think fair use, and how TiVO software and DRM
in general gets in the way.

> And... You do realize that almost every difference between the GPLv2
> and the GPLv3 is going to cause a hell of a lot of problems?

For those who are not willing to abide by the spirit of the license,
yes.  Does it look like I'm concerned about them?  If they're willing
to look for and maybe even find holes in the license to disrespect
users' freedoms, why should I worry about the problems that plugging
these holes is going to cause them?  If they'd taken the spirit of the
GPL for what it is, instead of looking for loopholes, this improved
wording wouldn't be causing them any problems whatsoever.

> The fact that the GPLv3 is designed to prevent things that RMS
> *PERSONALLY* finds distasteful - DRM and the like - is a big
> turn-off for a *LOT* of people.

This is a pretty sad accusation.  2/3s of the Free Software packages
use the GPL with its existing spirit, and you still haven't shown that
any changes proposed in GPLv3 fail to abide by the same spirit.  That
some (many?) people misunderstood or disregarded the spirit is an
unfortunate fact, but trying to pose the patching that's going into
GPLv3 as if it was a matter of personal taste, rather than improved
compliance with the spirit, is unfair and uncalled for.

> (Personally I don't like *ANY* version of the GPL, because there are
> chunks I have problems with)

What are you doing lurking and spreading confusion in a list about a
project that chose to use it, then?

>> Do you expect Linux would have flourished if computers had locks that
>> stopped people from modifying Linux in them?

> But you aren't talking about a "computer" here. You're talking about
> a mass-market device that must comply with both US and International
> copyright law - and that's just a TiVO.

Oh, sorry.  I missed when the meaning of the word computer was
narrowed from "machine with a general-purpose microprocessor, memory
and other peripherals" to whatever you decide it is.

And then, the GPL doesn't talk about computers at all.  It's not about
the hardware, it's about the software, remember? ;-)

> if you upload a modified linux kernel to your wireless router that
> gives it a 2000 foot range, you've just broken the law

At which point, you get punished by the law system.

> *AND* violated the license on the hardware which states that you
> "won't modify it or the controlling software"

Err..  The hardware licensor who includes software under the GPL be
supposed to be a licensee of the software in order to have legal
permission to distribute it, at which point the following provision
kicks in:

  6. Each time you redistribute the Program (or any work based on the
  Program) [...] You may not impose any further restrictions on the
  recipients' exercise of the rights granted herein. [...]

And here's one of the rights granted herein that would be restricted
by this hardware license:

  2. You may modify your copy or copies of the Program or any portion
  of it, thus forming a work based on the Program,

So such a restriction in the hardware license seems to be failure to
comply with the GPL, which means the violator may lose the license.

> even things like the connectors used to upload the operating
> software at the factory that people now cannot have in a device that
> runs GPL(v3) covered software unless they ship the related
> "Installation Information".

This sounds like a reasonable point.  Please bring it up at
gplv3.fsf.org.  If it requires specialized hardware to modify the
software in the device, the hardware manufacturer can't modify the
software without cooperation from the user, and then perhaps it would
be fair for the user to need cooperation from the manufacturer.

> That, to me, reads like RMS got mad about TiVO and said "I don't
> like it, lets add a clause making it wrong to the next GPL". Hell,
> that *IS* what happened, and nothing the FSF or Eben Moglen says
> will convince me otherwise.

If you've already made your mind about this, in spite of not having
the facts, I guess it doesn't make sense for me to waste my time
trying to convince you, does it?

>> What's under the license is the software in it.  And that license
>> spirit requires the distributor to pass on the right to modify the
>> software.

> And since when did they have to enable people to use their hardware in 
> violation of the licensing agreement they implicitly agree to when opening 
> the package?

Since they got permission to distribute the software under the
condition of passing on the freedoms without imposing further
restrictions on their exercise.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[david]

since the latest draft of the GPLv3 now discriminates against some uses 
(industrial vs commercial I think are the terms used) does it even qualify 
as a Open Source lincense anymore by the OSI terms?

David Lang

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 13, 2007, david@lang.hm wrote:

> since the latest draft of the GPLv3 now discriminates against some
> uses (industrial vs commercial I think are the terms used) 

  A "User Product" is either (1) a "consumer product," which means any
  tangible personal property which is normally used for personal,
  family, or household purposes, or (2) anything designed or sold for
  incorporation into a dwelling.

> does it even qualify as a Open Source lincense anymore by the OSI
> terms?

The definition is about the hardware, not the software, so it may
still qualify.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Wednesday 13 June 2007 22:38:05 Alexandre Oliva wrote:
> On Jun 13, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> > On Wednesday 13 June 2007 19:49:23 Alexandre Oliva wrote:
> >
> > Exactly. They don't. What TiVO prevents is using that modified version on
> > their hardware. And they have that right, because the Hardware *ISN'T*
> > covered by the GPL.
>
> Indeed, TiVO has this legal right.  But then they must not use
> software under the GPLv3 in it.  And, arguably, they must not use
> software under the GPLv2 either.

Yes. It can be argued. But I cannot find *ANYTHING* in the GPLv2 that stops 
anyone from doing that, unless you add extra meaning to one specific clause. 
(And I am willing to admit that *MOST* people do give it that extra meaning)

> > In the case of 99% of the hardware targeted by the clause of the GPLv3
> > you elucidate on, the "ability to install modified versions of the
> > software" was *NOT* intended for that use, nor was it intended for
> > *ANYONE* *EXCEPT* trained service personell to have *ACCESS* to that
> > functionality. Arguing otherwise is just idiotic - I have never found a
> > piece of "high tech" hardware (like a TiVO) that was designed for the
> > end-user to modify. (yes, installing a new version of the linux kernel is
> > "modifying" the system)
>
> It's about time for a change for better, wouldn't you think?

I've never had a reason to want to change the way any device like a TiVO 
works. So I can't comment on this.

> In 95% of the desktop computers, you can't make changes to the OS that
> runs on it.  Whom is this good for?

Faulty logic. I have yet to find a computer that I couldn't change the OS on. 
I have run Linux on 3 different Mac's, every x86 machine I've ever owned and 
even had it running on my Palm. Whats more is that I have *never* heard of a 
person that knows what they are doing not being able to change the OS on a 
desktop computer.

> > And? They distribute the kernel source - as they recieved it - in
> > compliance with the GPL.
>
> This makes it seem like you think that passing on the source code is
> enough to comply with the GPL.  Check your assumptions.  It's not.

Hrm. Strange, but thats what most companies think. Hell, it even says that you 
have to do just that in the GPL. If you're talking about the fact that it can 
be argued that they are "distributing" Linux by selling their boxes and its a 
modified version then I'll agree with you. 

> >> to prohibit people from removing locks that stop them from doing
> >> things they're legally entitled to do
> >
> > What "Legally Entitled" things?
>
> Time shifting of any shows, creating copies of shows for personal use,
> letting others do so.  Think fair use, and how TiVO software and DRM
> in general gets in the way.

I thought that time shifting and creating personal copies was what the TiVO 
did already. Or do you mean "transferring the recorded copies off the TiVO 
and on to a different medium"? If that is what you mean by "Creating Copies" 
then, IIRC, you're wrong. DRM, I do agree, gets in the way of "Fair Use".

> > And... You do realize that almost every difference between the GPLv2
> > and the GPLv3 is going to cause a hell of a lot of problems?
>
> For those who are not willing to abide by the spirit of the license,
> yes.  Does it look like I'm concerned about them?  If they're willing
> to look for and maybe even find holes in the license to disrespect
> users' freedoms, why should I worry about the problems that plugging
> these holes is going to cause them?  If they'd taken the spirit of the
> GPL for what it is, instead of looking for loopholes, this improved
> wording wouldn't be causing them any problems whatsoever.

Okay. So you're not concerned that you're potentially pushing companies that 
would otherwise be major consumers of GPL'd software away? That doesn't make 
sense to me.

> > The fact that the GPLv3 is designed to prevent things that RMS
> > *PERSONALLY* finds distasteful - DRM and the like - is a big
> > turn-off for a *LOT* of people.
>
> This is a pretty sad accusation.  2/3s of the Free Software packages
> use the GPL with its existing spirit, and you still haven't shown that
> any changes proposed in GPLv3 fail to abide by the same spirit.  That
> some (many?) people misunderstood or disregarded the spirit is an
> unfortunate fact, but trying to pose the patching that's going into
> GPLv3 as if it was a matter of personal taste, rather than improved
> compliance with the spirit, is unfair and uncalled for.

Why should I repeat Linus' explanation of the ways that GPLv3 violates the 
spirit of GPLv2?

And why shouldn't I pose it as a matter of "Personal Taste"? The biggest and 
most powerful voice in the FSF says "I don't like Tivoization" and "I don't 
like DRM" and when the GPLv3 appears it has language that makes those 
violations of the license. Just like people have started using "GNU/Linux" 
or "GNU+Linux" to refer to Linux - a big voice spoke and said "It should be 
GNU/Linux" and it happens. (Not that I really have anything against that - 
though I feel that "GNU/Linux" gives the wrong impression)

> > (Personally I don't like *ANY* version of the GPL, because there are
> > chunks I have problems with)
>
> What are you doing lurking and spreading confusion in a list about a
> project that chose to use it, then?

Just because I don't like the license doesn't make disqualify me from liking 
something that uses the license. And I doubt I've "confused" anyone - where I 
have someone has caught it and called me on it.

> >> Do you expect Linux would have flourished if computers had locks that
> >> stopped people from modifying Linux in them?
> >
> > But you aren't talking about a "computer" here. You're talking about
> > a mass-market device that must comply with both US and International
> > copyright law - and that's just a TiVO.
>
> Oh, sorry.  I missed when the meaning of the word computer was
> narrowed from "machine with a general-purpose microprocessor, memory
> and other peripherals" to whatever you decide it is.

The word "Computer", in the manner I used it there, means "General Purpose 
Computational Device". A "TiVO" is not, and has never been, a "General 
Purpose Computational Device". 

(Perhaps I should have made the above clear)

> And then, the GPL doesn't talk about computers at all.  It's not about
> the hardware, it's about the software, remember? ;-)

Exactly. And I don't see anything about a TiVO (or any device that, like a 
TiVO, requires binaries that run on it to be digitally signed) that stops you 
from exercising the "freedoms" guaranteed by the GPL. As I said before, what 
it does is stop you from violating the license on the hardware.

> > if you upload a modified linux kernel to your wireless router that
> > gives it a 2000 foot range, you've just broken the law
>
> At which point, you get punished by the law system.

But the GPLv2 gives companies a chance to protect themselves from legal 
actions by people that are sure to follow. With the GPLv3 they have none, 
because it becomes a case of "I was just doing what I have the right to do 
under the license for the software. If I wasn't supposed to do it, they 
shouldn't have made it possible."

> > *AND* violated the license on the hardware which states that you
> > "won't modify it or the controlling software"
>
> Err..  The hardware licensor who includes software under the GPL be
> supposed to be a licensee of the software in order to have legal
> permission to distribute it, at which point the following provision
> kicks in:
>
>   6. Each time you redistribute the Program (or any work based on the
>   Program) [...] You may not impose any further restrictions on the
>   recipients' exercise of the rights granted herein. [...]

Hrm... I'd forgotten that clause entirely. That must be why I can't find a 
WRT-54GL anywhere in the US anymore...

> And here's one of the rights granted herein that would be restricted
> by this hardware license:
>
>   2. You may modify your copy or copies of the Program or any portion
>   of it, thus forming a work based on the Program,
>
> So such a restriction in the hardware license seems to be failure to
> comply with the GPL, which means the violator may lose the license.

My apologies. As I said above I totally forgot about section 6. 

> > even things like the connectors used to upload the operating
> > software at the factory that people now cannot have in a device that
> > runs GPL(v3) covered software unless they ship the related
> > "Installation Information".
>
> This sounds like a reasonable point.  Please bring it up at
> gplv3.fsf.org.  If it requires specialized hardware to modify the
> software in the device, the hardware manufacturer can't modify the
> software without cooperation from the user, and then perhaps it would
> be fair for the user to need cooperation from the manufacturer.

Will do. I'm surprised that this situation never came up before, actually. 

> > That, to me, reads like RMS got mad about TiVO and said "I don't
> > like it, lets add a clause making it wrong to the next GPL". Hell,
> > that *IS* what happened, and nothing the FSF or Eben Moglen says
> > will convince me otherwise.
>
> If you've already made your mind about this, in spite of not having
> the facts, I guess it doesn't make sense for me to waste my time
> trying to convince you, does it?

You've made some well reasoned and supported arguments here. I respect RMS and 
Eben Moglen because of what they've done, but that doesn't mean I have to 
like them. However, you are free to (try) to change my mind about them.

> >> What's under the license is the software in it.  And that license
> >> spirit requires the distributor to pass on the right to modify the
> >> software.
> >
> > And since when did they have to enable people to use their hardware in
> > violation of the licensing agreement they implicitly agree to when
> > opening the package?
>
> Since they got permission to distribute the software under the
> condition of passing on the freedoms without imposing further
> restrictions on their exercise.

Okay. Given the interpretation of GPLv2 you seem to have I will grant you this 
point. Personally, I do not feel that the "spirit" of the GPLv2 has anything 
to do with running modified binaries on a "closed" hardware platform (like a 
TiVO). But, as I said, I'll grant you the point, because our views on 
the "spirit" of the GPL and interpretations of it seem to be completely 
incompatible.

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> I've never had a reason to want to change the way any device like a TiVO 
> works. So I can't comment on this.

Have you never wanted to improve any aspect of the software in your
cell phone?  In your TV, VCR, DVD player, anything?  In the microwave
oven, maybe?

> On Wednesday 13 June 2007 22:38:05 Alexandre Oliva wrote:
>> In 95% of the desktop computers, you can't make changes to the OS that
>> runs on it.  Whom is this good for?

> Faulty logic. I have yet to find a computer that I couldn't change the OS on. 

I was not talking about installing another OS, I was talking about
making changes to the OS.  As in, improving one particular driver,
avoiding a blue screen, stuff like that.

> Or do you mean "transferring the recorded copies off the TiVO 
> and on to a different medium"?

Sure.  Such that I can watch shows while wasting time in public
transportation, in an airplane, whatever.

> DRM, I do agree, gets in the way of "Fair Use".

And the fact that TiVO can be, and has been modified remotely to add
restrictions on what users could do, means nothing you do with it is
safe.  You, and everything you've recorded with the TiVO, are at the
mercy of this one company.

> So you're not concerned that you're potentially pushing companies
> that would otherwise be major consumers of GPL'd software away? That
> doesn't make sense to me.

What would their consuming GPL software buy us, if they won't respect
users' freedoms, which is the very reason behind the GPL?

Heck, if they don't want to play by the rules, that's up to them.  But
then they shouldn't use the software at all.

Yeah, I wish they'd rather play by the rules, but if they don't want
to, too bad, for us and for them.

> Why should I repeat Linus' explanation of the ways that GPLv3 violates the 
> spirit of GPLv2?

Don't worry about parrotting here, he hasn't provided that explanation
yet ;-)  Please give it a try.

BTW, what license is Linux licensed under?  It's GPLv2 plus userland
exception, right?  (There's some additional module exception, right?)

> And why shouldn't I pose it as a matter of "Personal Taste"? The
> biggest and most powerful voice in the FSF says "I don't like
> Tivoization" and "I don't like DRM" and when the GPLv3 appears it
> has language that makes those violations of the license.

Have you ever wondered *why* he doesn't like them?

Could it possibly be because they harm the goal of his life, which is
to enable people to live their digital lives in freedom?

> Just like people have started using "GNU/Linux" or "GNU+Linux" to
> refer to Linux

No, no, you got it wrong.  Linux is the kernel.  GNU was the
nearly-complete operating system it fit in.  GNU+Linux is a complete
operating system.

And you don't have to believe me, believe Linus, the initial author of
Linux:

http://www.kernel.org/pub/linux/kernel/Historic/old-versions/RELNOTES-0.01

  Although linux is a complete kernel

  Sadly, a kernel by itself gets you nowhere. To get a working system
  you need a shell, compilers, a library etc. These are separate parts
  and may be under a stricter (or even looser) copyright. Most of the
  tools used with linux are GNU software and are under the GNU
  copyleft.

> A "TiVO" is not, and has never been, a "General Purpose
> Computational Device".

Err...  Last I looked it was a bunch of general-purpose components,
packaged in a way that made it not look like a general-purpose
computer.  Who gets to decide?  And with what motivations?

> Exactly. And I don't see anything about a TiVO (or any device that, like a 
> TiVO, requires binaries that run on it to be digitally signed) that stops you 
> from exercising the "freedoms" guaranteed by the GPL.

  2. You may modify your copy or copies of the Program or any portion
  of it

>> > if you upload a modified linux kernel to your wireless router that
>> > gives it a 2000 foot range, you've just broken the law

>> At which point, you get punished by the law system.

> But the GPLv2 gives companies a chance to protect themselves from legal 
> actions by people that are sure to follow.

So does the GPLv3.  It might be a bit narrower, to cut on other kinds
of abuses, but all constraints I'm aware of that are mandated by law
can still be achieved.  The point is to forbid disrespecting users'
freedoms to modify the software.  Configuration parameters for the
hardware, needed to comply with regulations, can be easily taken care
of without disrespecting users' freedoms.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Thursday 14 June 2007 01:51:13 Alexandre Oliva wrote:
> On Jun 14, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> > I've never had a reason to want to change the way any device like a TiVO
> > works. So I can't comment on this.
>
> Have you never wanted to improve any aspect of the software in your
> cell phone?  In your TV, VCR, DVD player, anything?  In the microwave
> oven, maybe?

Nope. I've been tempted several times, but decided that the extra bits I'd 
thought about wouldn't add anything to the device.

> > On Wednesday 13 June 2007 22:38:05 Alexandre Oliva wrote:
> >> In 95% of the desktop computers, you can't make changes to the OS that
> >> runs on it.  Whom is this good for?
> >
> > Faulty logic. I have yet to find a computer that I couldn't change the OS
> > on.
>
> I was not talking about installing another OS, I was talking about
> making changes to the OS.  As in, improving one particular driver,
> avoiding a blue screen, stuff like that.

Ah, well... In the case of "Windos" and other proprietary OS's I try to 
educate people and get them to switch. I don't, personally, have any 
computers that run Windows (and I switched my Palm back to PalmOS because it 
wasn't getting the same performance under Linux - which rather surprised me. 
And rather than fight with it I just switched it back.

> > Or do you mean "transferring the recorded copies off the TiVO
> > and on to a different medium"?
>
> Sure.  Such that I can watch shows while wasting time in public
> transportation, in an airplane, whatever.

Under the US Copyright law I'm not sure that making a "second copy" like that 
is legal. IIRC, "Fair Use" only allows for one copy.

> > DRM, I do agree, gets in the way of "Fair Use".
>
> And the fact that TiVO can be, and has been modified remotely to add
> restrictions on what users could do, means nothing you do with it is
> safe.  You, and everything you've recorded with the TiVO, are at the
> mercy of this one company.

As has been noted in their TOS and the licenses for the hardware from the 
start. The FSF itself explicitly reserves the right to change the GPL at any 
time - which is no different. (when you remove all the bits explaining the 
purpose of the license)

> > So you're not concerned that you're potentially pushing companies
> > that would otherwise be major consumers of GPL'd software away? That
> > doesn't make sense to me.
>
> What would their consuming GPL software buy us, if they won't respect
> users' freedoms, which is the very reason behind the GPL?

I'm not referring to companies that are embedding GPL'd software in their 
products. The companies I'm referring to are the ones that would like to use 
GPL'd software internally. A lot of them would probably have private 
modifications that would never be distributed - and under the GPLv2 it is 
clear that you can keep modifications private as long as you don't distribute 
them. "Pushing them away" means that they'd not do that because they would be 
concerned that the license will change under them in such a way that even 
those private modifications need to be released to the public.

(and don't try to argue that even though those modifications are truly private 
(to the company) they should be released anyway to comply with the "spirit" 
of the license. It is made clear that it isn't by the text of the license 
itself)

> Heck, if they don't want to play by the rules, that's up to them.  But
> then they shouldn't use the software at all.
>
> Yeah, I wish they'd rather play by the rules, but if they don't want
> to, too bad, for us and for them.
>
> > Why should I repeat Linus' explanation of the ways that GPLv3 violates
> > the spirit of GPLv2?
>
> Don't worry about parrotting here, he hasn't provided that explanation
> yet ;-)  Please give it a try.

But he has. Whether you have accepted that his explanations are valid or not 
doesn't change the fact.

> BTW, what license is Linux licensed under?  It's GPLv2 plus userland
> exception, right?  (There's some additional module exception, right?)

The kernel itself is GPLv2 (only). Individual components - even individual 
files - have other licenses or retain the "any later version" clause. 
(Someone pointed out, earlier in this thread, that there is GPLv1.1 code in 
the kernel)

> > And why shouldn't I pose it as a matter of "Personal Taste"? The
> > biggest and most powerful voice in the FSF says "I don't like
> > Tivoization" and "I don't like DRM" and when the GPLv3 appears it
> > has language that makes those violations of the license.
>
> Have you ever wondered *why* he doesn't like them?

Not really. I've always figured he had reasons similar to mine for not liking 
DRM. As to his dislike of "Tivoization", well, that I've always attributed to 
the fact that someone at that company managed to outsmart him. (and no, I 
wasn't being serious with that last line)

> Could it possibly be because they harm the goal of his life, which is
> to enable people to live their digital lives in freedom?
>
> > Just like people have started using "GNU/Linux" or "GNU+Linux" to
> > refer to Linux
>
> No, no, you got it wrong.  Linux is the kernel.  GNU was the
> nearly-complete operating system it fit in.  GNU+Linux is a complete
> operating system.

Yet I still find people that insist that the *ENTIRE* system - kernel and 
all - is a GNU project. Not just the common idiot you'd find on almost any 
street in the US, but also educated people in technical fields. The reason is 
the name. 

*AND* you cut out the bit where I said "I have no problems with it"

> And you don't have to believe me, believe Linus, the initial author of
> Linux:
>
> http://www.kernel.org/pub/linux/kernel/Historic/old-versions/RELNOTES-0.01
>
>   Although linux is a complete kernel
>
>   Sadly, a kernel by itself gets you nowhere. To get a working system
>   you need a shell, compilers, a library etc. These are separate parts
>   and may be under a stricter (or even looser) copyright. Most of the
>   tools used with linux are GNU software and are under the GNU
>   copyleft.

Never claimed otherwise. The problem is that using a composite name like that 
*does* confuse a hell of a lot of people. See my statements above about that.

> > A "TiVO" is not, and has never been, a "General Purpose
> > Computational Device".
>
> Err...  Last I looked it was a bunch of general-purpose components,
> packaged in a way that made it not look like a general-purpose
> computer.  Who gets to decide?  And with what motivations?

And so is every game console. But until the original XBox was released nobody 
tried using one as a "General Purpose" machine. The TiVO wasn't designed as a 
general purpose machine - it was designed for a specific purpose. That the 
*easiest* design to produce uses a bunch of general purpose components is an 
economic choice, nothing else.

I will not, however, argue about this anymore. As with other bits, we've 
reached a point where we disagree and no amount of explanation will change 
the others viewpoint. (I hope you understand mine as well as (I think) I 
understand yours)

> > Exactly. And I don't see anything about a TiVO (or any device that, like
> > a TiVO, requires binaries that run on it to be digitally signed) that
> > stops you from exercising the "freedoms" guaranteed by the GPL.
>
>   2. You may modify your copy or copies of the Program or any portion
>   of it
>
> >> > if you upload a modified linux kernel to your wireless router that
> >> > gives it a 2000 foot range, you've just broken the law
> >>
> >> At which point, you get punished by the law system.
> >
> > But the GPLv2 gives companies a chance to protect themselves from legal
> > actions by people that are sure to follow.
>
> So does the GPLv3.  It might be a bit narrower, to cut on other kinds
> of abuses, but all constraints I'm aware of that are mandated by law
> can still be achieved.  The point is to forbid disrespecting users'
> freedoms to modify the software.  Configuration parameters for the
> hardware, needed to comply with regulations, can be easily taken care
> of without disrespecting users' freedoms.

Let me quote Linus here:

But I think the whole thing is totally misguided, because the fact is, the 
GPLv2 doesn't talk about "in place" or "on the same hardware".

In other words, GPLv3 is breaking with its predecessor - it's adding a 
requirement that doesn't exist in previous versions. *AND* it's dictating 
terms for *HARDWARE* when it isn't a hardware license. If I release software 
under the GPL and somebody modifies it to run on a different hardware 
platform I'll be happy, even if they don't send me a patchset for the new 
version. If I create a piece of hardware and run Linux on it, but have it 
locked to a specific version or versions from a specific source (ie: me) and 
release it to the public, I *WILL* release the version of Linux I'm running 
on it. What I won't do is release whatever tools and such that are needed to 
make the hardware run a different version of the kernel. Why? Because: the 
hardware was designed so that a specific version of the kernel runs without 
problems, there is hardware that is very picky and running a customized 
kernel could cause that hardware to fail, etc... There are more reasons than 
the legal protection I previously mentioned. Not a single one of them 
is "Because I want to restrict peoples freedoms re: the GPL'd software". 

Admittedly the two examples I chose aren't very realistic, but those were the 
first two examples that came to mind.

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> On Thursday 14 June 2007 01:51:13 Alexandre Oliva wrote:
>> On Jun 14, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
>> > On Wednesday 13 June 2007 22:38:05 Alexandre Oliva wrote:
>> >> In 95% of the desktop computers, you can't make changes to the OS that
>> >> runs on it.  Whom is this good for?

>> > Faulty logic. I have yet to find a computer that I couldn't change the OS
>> > on.

>> I was not talking about installing another OS, I was talking about
>> making changes to the OS.  As in, improving one particular driver,
>> avoiding a blue screen, stuff like that.

> Ah, well... In the case of "Windos" and other proprietary OS's I try to 
> educate people and get them to switch.

Good.  So I presume you'd tell them to switch away from a
turned-proprietary GNU/Linux operating system as well, right?

So, again, what do we gain if companies abuse the GPL and disrespect
users' rights that we meant them to respect?

>> > Or do you mean "transferring the recorded copies off the TiVO
>> > and on to a different medium"?

>> Sure.  Such that I can watch shows while wasting time in public
>> transportation, in an airplane, whatever.

> Under the US Copyright law I'm not sure that making a "second copy"
> like that is legal. IIRC, "Fair Use" only allows for one copy.

Even if you delete the "first copy"?

Actually, I thought fair use in US entitled you to make a backup copy.
So the copy in your TiVO would be your original, and the external copy
would be your fair-use backup.

> As has been noted in their TOS and the licenses for the hardware from the 
> start.

If it is used to disrespect the inalienable freedoms associated with
the GPL software in the device, it seems like a license violation to
me.

> The FSF itself explicitly reserves the right to change the GPL at any 
> time - which is no different.

Actually, it's completely different.

If the FSF revises the GPL, the old version remains available for
anyone to use for any new software, and all software released under
the old version remains available under that old version.

In contrast, your TiVO may get a software upgrade without your
permission that will take your rights away from that point on, and
there's very little you can do about it, other than unplugging it from
the network to avoid the upgrade if it's not too late already.

> A lot of them would probably have private modifications that would
> never be distributed - and under the GPLv2 it is clear that you can
> keep modifications private as long as you don't distribute them.

Likewise with GPLv3.

> "Pushing them away" means that they'd not do that because they would
> be concerned that the license will change under them in such a way
> that even those private modifications need to be released to the
> public.

This would not only change the spirit of the license, but turn it into
a non-Free Software license.

And then, again, the license can't possibly be changed from under
them.  A new revision of the GPL would only affect software licensed
under that new revision.  If you already got it under an earlier
revision, you know what you got, and nobody can take that away from
you.

> (and don't try to argue that even though those modifications are
> truly private (to the company) they should be released anyway to
> comply with the "spirit" of the license. It is made clear that it
> isn't by the text of the license itself)

How could you possibly come to the conclusion that forcing anyone to
release private modifications would be in compliance with the spirit
of the license?  can != must

>> > Why should I repeat Linus' explanation of the ways that GPLv3 violates
>> > the spirit of GPLv2?

>> Don't worry about parrotting here, he hasn't provided that explanation
>> yet ;-)  Please give it a try.

> But he has. Whether you have accepted that his explanations are
> valid or not doesn't change the fact.

His explanation is based on a reading of the license that doesn't
match what its authors meant.  I guess the authors know better what
they meant the spirit of the license to be than someone else who
studied it a lot but that until very recently couldn't even tell the
spirit from the legal terms.

>> > Just like people have started using "GNU/Linux" or "GNU+Linux" to
>> > refer to Linux

>> No, no, you got it wrong.  Linux is the kernel.  GNU was the
>> nearly-complete operating system it fit in.  GNU+Linux is a complete
>> operating system.

> *AND* you cut out the bit where I said "I have no problems with it"

Referring to Linux as GNU/Linux would be wrong, because Linux is the
kernel, and that's unrelated with the GNU operating system.  It's the
combination of them that forms GNU+Linux.  And it's referring to this
combination as Linux that is wrong.

I'm sorry that I got the impression that you meant the combination
when you wrote "refer to Linux" above.  It looked like you meant the
combination, since I've never seen anyone call the kernel GNU/Linux or
GNU+Linux.

> Never claimed otherwise. The problem is that using a composite name like that 
> *does* confuse a hell of a lot of people.

Pronouncing the '+' or '/' helps a lot.  GNU plus Linux makes a lot of
sense, and so does GNU on Linux.

>> > A "TiVO" is not, and has never been, a "General Purpose
>> > Computational Device".

>> Err...  Last I looked it was a bunch of general-purpose components,
>> packaged in a way that made it not look like a general-purpose
>> computer.  Who gets to decide?  And with what motivations?

> And so is every game console. But until the original XBox was released nobody 
> tried using one as a "General Purpose" machine. The TiVO wasn't designed as a 
> general purpose machine - it was designed for a specific purpose. That the 
> *easiest* design to produce uses a bunch of general purpose components is an 
> economic choice, nothing else.

So, if I put together a general-purpose computer, a general-purpose
operating system, adding a label "not a general-purpose computer" is
enough to make it so, just so that I can escape the obligations to
respect users' freedoms?

> I will not, however, argue about this anymore.

Fair enough.

> Let me quote Linus here:

> But I think the whole thing is totally misguided, because the fact is, the 
> GPLv2 doesn't talk about "in place" or "on the same hardware".

> In other words, GPLv3 is breaking with its predecessor - it's adding a 
> requirement that doesn't exist in previous versions.

No dispute about this.  The requirements are being added to the legal
terms, precisely such that they better reflect the spirit, under the
light of the new threats that appeared since GPLv2 was published.

But the new requirements do abide by the same spirit, and that was a
promise the FSF made WRT revisions of the GPL.

> *AND* it's dictating terms for *HARDWARE* when it isn't a hardware
> license.

Only in as much as you try to use the hardware as means to disrespect
the spirit of the license and escape from the obligation to respect
users' freedoms.

> If I release software under the GPL and somebody modifies it to run
> on a different hardware platform I'll be happy, even if they don't
> send me a patchset for the new version.

Yup.  See the bit about GPL not being tit-for-tat.

> If I create a piece of hardware and run Linux on it, but have it
> locked to a specific version or versions from a specific source (ie:
> me) and release it to the public, I *WILL* release the version of
> Linux I'm running on it.

Good.

> What I won't do is release whatever tools and such that are needed to 
> make the hardware run a different version of the kernel. Why? Because: the 
> hardware was designed so that a specific version of the kernel runs without 
> problems, there is hardware that is very picky and running a customized 
> kernel could cause that hardware to fail, etc...

Why do you care?  It's no longer your hardware, it's theirs.

Why would you refrain from providing information to others such that
they *could* make the software do what *they* want in their hardware
that they got from you?

If you let them change it and they break it, they get to keep all the
pieces.  Your job is done.  Why get out of your way to stop them from
making the best out of *their* hardware?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Thursday 14 June 2007 03:11:45 Alexandre Oliva wrote:
> On Jun 14, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> > On Thursday 14 June 2007 01:51:13 Alexandre Oliva wrote:
> >> On Jun 14, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> >> > On Wednesday 13 June 2007 22:38:05 Alexandre Oliva wrote:
> >> >> In 95% of the desktop computers, you can't make changes to the OS
> >> >> that runs on it.  Whom is this good for?
> >> >
> >> > Faulty logic. I have yet to find a computer that I couldn't change the
> >> > OS on.
> >>
> >> I was not talking about installing another OS, I was talking about
> >> making changes to the OS.  As in, improving one particular driver,
> >> avoiding a blue screen, stuff like that.
> >
> > Ah, well... In the case of "Windos" and other proprietary OS's I try to
> > educate people and get them to switch.
>
> Good.  So I presume you'd tell them to switch away from a
> turned-proprietary GNU/Linux operating system as well, right?

If that happened I'd be lost. I've tried the various BSD's and found they had 
problems with hardware support and getting a new version of the BSD kernel to 
compile and boot is something of a black art.

The point is moot, though. It can never happen.

> So, again, what do we gain if companies abuse the GPL and disrespect
> users' rights that we meant them to respect?
>
> >> > Or do you mean "transferring the recorded copies off the TiVO
> >> > and on to a different medium"?
> >>
> >> Sure.  Such that I can watch shows while wasting time in public
> >> transportation, in an airplane, whatever.
> >
> > Under the US Copyright law I'm not sure that making a "second copy"
> > like that is legal. IIRC, "Fair Use" only allows for one copy.
>
> Even if you delete the "first copy"?
>
> Actually, I thought fair use in US entitled you to make a backup copy.
> So the copy in your TiVO would be your original, and the external copy
> would be your fair-use backup.

Hrm... Perhaps. 

> > As has been noted in their TOS and the licenses for the hardware from the
> > start.
>
> If it is used to disrespect the inalienable freedoms associated with
> the GPL software in the device, it seems like a license violation to
> me.

As much as the US "Declaration of Independence" and other sources want people 
to believe otherwise there is no such thing as "inalienable rights" 
or "inalienable freedoms". In this case I have been unable to find 
this "inalienable freedom" to run custom versions of software "on the same 
machine" that you received the original copy on anywhere before the GPLv3 - 
and even then it isn't explicitly clear. There is no restriction on your 
right to modify, copy, distribute or run the software as provided by versions 
of the GPL prior to version 3. If this "run modified copies on the same 
hardware you received the original on" *IS* the "spirit" of the license, then 
why isn't it stated anywhere before GPLv3? (After all, the FSF has have 20+ 
years to mention it)

> > The FSF itself explicitly reserves the right to change the GPL at any
> > time - which is no different.
>
> Actually, it's completely different.
>
> If the FSF revises the GPL, the old version remains available for
> anyone to use for any new software, and all software released under
> the old version remains available under that old version.

I'll grant you that. But, at this point, where can I find a copy of the GPLv1 
without having to dig around the net ?

> In contrast, your TiVO may get a software upgrade without your
> permission that will take your rights away from that point on, and
> there's very little you can do about it, other than unplugging it from
> the network to avoid the upgrade if it's not too late already.

And because its a device that connects to their network - and TiVO isn't a 
telecommunications company - they have the right to upgrade and configure the 
software inside however they want. (In the US at least)

> > A lot of them would probably have private modifications that would
> > never be distributed - and under the GPLv2 it is clear that you can
> > keep modifications private as long as you don't distribute them.
>
> Likewise with GPLv3.

I can see this, but will a company see this?

> > "Pushing them away" means that they'd not do that because they would
> > be concerned that the license will change under them in such a way
> > that even those private modifications need to be released to the
> > public.
>
> This would not only change the spirit of the license, but turn it into
> a non-Free Software license.

Point. But once again - would a company pay attention to that fact?

> And then, again, the license can't possibly be changed from under
> them.  A new revision of the GPL would only affect software licensed
> under that new revision.  If you already got it under an earlier
> revision, you know what you got, and nobody can take that away from
> you.

True. But that doesn't save them from lawsuits trying to force them to obey 
the terms of the new revision even though they received the software under an 
earlier version.

> > (and don't try to argue that even though those modifications are
> > truly private (to the company) they should be released anyway to
> > comply with the "spirit" of the license. It is made clear that it
> > isn't by the text of the license itself)
>
> How could you possibly come to the conclusion that forcing anyone to
> release private modifications would be in compliance with the spirit
> of the license?  can != must

I was trying to be sarcastic and inject a little humor here. Guess I should 
have used the old <sarcasm> tag :)

> >> > Why should I repeat Linus' explanation of the ways that GPLv3 violates
> >> > the spirit of GPLv2?
> >>
> >> Don't worry about parrotting here, he hasn't provided that explanation
> >> yet ;-)  Please give it a try.
> >
> > But he has. Whether you have accepted that his explanations are
> > valid or not doesn't change the fact.
>
> His explanation is based on a reading of the license that doesn't
> match what its authors meant.  I guess the authors know better what
> they meant the spirit of the license to be than someone else who
> studied it a lot but that until very recently couldn't even tell the
> spirit from the legal terms.

And his interpretation is no less valid than that of anyone else. In fact, 
after a recent conversation with a couple of lawyers that I know, I can state 
that his interpretation isn't that far off from theirs.

> >> > Just like people have started using "GNU/Linux" or "GNU+Linux" to
> >> > refer to Linux
> >>
> >> No, no, you got it wrong.  Linux is the kernel.  GNU was the
> >> nearly-complete operating system it fit in.  GNU+Linux is a complete
> >> operating system.
> >
> > *AND* you cut out the bit where I said "I have no problems with it"
>
> Referring to Linux as GNU/Linux would be wrong, because Linux is the
> kernel, and that's unrelated with the GNU operating system.  It's the
> combination of them that forms GNU+Linux.  And it's referring to this
> combination as Linux that is wrong.
>
> I'm sorry that I got the impression that you meant the combination
> when you wrote "refer to Linux" above.  It looked like you meant the
> combination, since I've never seen anyone call the kernel GNU/Linux or
> GNU+Linux.

Then you're lucky. I've had a lot of people say something similar to the 
following: "Oh, I've heard about that. So which version of the GNU-Linux 
kernel are you running?" 

> > Never claimed otherwise. The problem is that using a composite name like
> > that *does* confuse a hell of a lot of people.
>
> Pronouncing the '+' or '/' helps a lot.  GNU plus Linux makes a lot of
> sense, and so does GNU on Linux.

Yes, it does. While pronouncing the '/' or '+' sounds a bit odd it does get 
the point across that it's the GNU userspace running on top of the Linux 
kernel. (as does "GNU on Linux")

> >> > A "TiVO" is not, and has never been, a "General Purpose
> >> > Computational Device".
> >>
> >> Err...  Last I looked it was a bunch of general-purpose components,
> >> packaged in a way that made it not look like a general-purpose
> >> computer.  Who gets to decide?  And with what motivations?
> >
> > And so is every game console. But until the original XBox was released
> > nobody tried using one as a "General Purpose" machine. The TiVO wasn't
> > designed as a general purpose machine - it was designed for a specific
> > purpose. That the *easiest* design to produce uses a bunch of general
> > purpose components is an economic choice, nothing else.
>
> So, if I put together a general-purpose computer, a general-purpose
> operating system, adding a label "not a general-purpose computer" is
> enough to make it so, just so that I can escape the obligations to
> respect users' freedoms?
>
> > I will not, however, argue about this anymore.
>
> Fair enough.
>
> > Let me quote Linus here:
> >
> > But I think the whole thing is totally misguided, because the fact is,
> > the GPLv2 doesn't talk about "in place" or "on the same hardware".
> >
> > In other words, GPLv3 is breaking with its predecessor - it's adding a
> > requirement that doesn't exist in previous versions.
>
> No dispute about this.  The requirements are being added to the legal
> terms, precisely such that they better reflect the spirit, under the
> light of the new threats that appeared since GPLv2 was published.
>
> But the new requirements do abide by the same spirit, and that was a
> promise the FSF made WRT revisions of the GPL.

As I've stated before - I can find nothing in the history of the GPL or the 
FSF that makes the "on the same hardware" requirement clear and part of 
the "spirit" of "Free Software". The closest anything comes is the "printer 
driver" that was the (in)famous "last straw" for RMS and caused him to create 
the FSF and the GNU Project.

> > *AND* it's dictating terms for *HARDWARE* when it isn't a hardware
> > license.
>
> Only in as much as you try to use the hardware as means to disrespect
> the spirit of the license and escape from the obligation to respect
> users' freedoms.

Shouldn't matter. As I've repeated quite a bit I cannot find a single mention 
that "on the same hardware" has been a goal of the FSF or part of 
the "spirit" of the GPL at any point before the drafting of GPLv3.

> > If I release software under the GPL and somebody modifies it to run
> > on a different hardware platform I'll be happy, even if they don't
> > send me a patchset for the new version.
>
> Yup.  See the bit about GPL not being tit-for-tat.
>
> > If I create a piece of hardware and run Linux on it, but have it
> > locked to a specific version or versions from a specific source (ie:
> > me) and release it to the public, I *WILL* release the version of
> > Linux I'm running on it.
>
> Good.
>
> > What I won't do is release whatever tools and such that are needed to
> > make the hardware run a different version of the kernel. Why? Because:
> > the hardware was designed so that a specific version of the kernel runs
> > without problems, there is hardware that is very picky and running a
> > customized kernel could cause that hardware to fail, etc...
>
> Why do you care?  It's no longer your hardware, it's theirs.

Legal requirements in some countries that require manufacturers to provide 
support for their product for a period of time after it has been purchased.

> Why would you refrain from providing information to others such that
> they *could* make the software do what *they* want in their hardware
> that they got from you?

See above. 

> If you let them change it and they break it, they get to keep all the
> pieces.  Your job is done.  Why get out of your way to stop them from
> making the best out of *their* hardware?

*thinks* Okay - looks like I had a thinko there. Arguments on this subject 
(ie: the theoretical hardware) are withdrawn and the point is conceded.

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> On Thursday 14 June 2007 03:11:45 Alexandre Oliva wrote:
>> On Jun 14, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
>> > Ah, well... In the case of "Windos" and other proprietary OS's I try to
>> > educate people and get them to switch.

>> Good.  So I presume you'd tell them to switch away from a
>> turned-proprietary GNU/Linux operating system as well, right?

> If that happened I'd be lost. I've tried the various BSD's and found they had 
> problems with hardware support and getting a new version of the BSD kernel to 
> compile and boot is something of a black art.

> The point is moot, though. It can never happen.

Look again, it's already happened in the TiVO and other devices.

The software that ships in them is no longer Free Software.


Consider a new microprocessor.

Consider that Linux is ported to it by the microprocessor
manufacturer.

Consider that the manufacturer only sells devices with that
microprocessor with TiVO-like locks.

How exactly can you enjoy the freedoms WRT the GPLed software you got
from the manufacturer?


Now consider that you have a single computer, and that's built by TiVO.

How exactly can you enjoy the freedoms the author meant you to have,
if the TiVO box won't run the program after you modify it?

> If this "run modified copies on the same hardware you received the
> original on" *IS* the "spirit" of the license, then why isn't it
> stated anywhere before GPLv3?

For the same reasons that the pro-DRM laws weren't mentioned before,
and the patent retaliation clauses weren't mention before: these
specific cases hadn't been studied, only the general idea of
respecting users' freedoms was.

> I'll grant you that. But, at this point, where can I find a copy of
> the GPLv1 without having to dig around the net ?

In the program you received under GPLv1.

Hey, you said there was code under GPLv1.1 in the Linux tree.  Then,
there should be a copy of GPLv1.1 in there, otherwise AFAICT the
distribution of that code is copyright infringement.  IANAL.

>> In contrast, your TiVO may get a software upgrade without your
>> permission that will take your rights away from that point on, and
>> there's very little you can do about it, other than unplugging it from
>> the network to avoid the upgrade if it's not too late already.

> And because its a device that connects to their network - and TiVO
> isn't a telecommunications company - they have the right to upgrade
> and configure the software inside however they want. (In the US at
> least)

But do they have the right to not pass this right on, under the GPL?


>> > A lot of them would probably have private modifications that would
>> > never be distributed - and under the GPLv2 it is clear that you can
>> > keep modifications private as long as you don't distribute them.

>> Likewise with GPLv3.

> I can see this, but will a company see this?

In what sense does the GPLv3 make this particular point any less
obscure?

> True. But that doesn't save them from lawsuits trying to force them
> to obey the terms of the new revision even though they received the
> software under an earlier version.

Nothing saves anyone from silly lawsuits.  This one would likely be
laughed out of court in no time.  Anyone worried about this should
also be concerned about their neighbor suing them for copyright
infringment every time they set their stereo loud enough for the
neighbors to listen and be annoyed.  (Hint: only the copyright holder
would stand a chance of winning such a lawsuit)

>> > (and don't try to argue that even though those modifications are
>> > truly private (to the company) they should be released anyway to
>> > comply with the "spirit" of the license. It is made clear that it
>> > isn't by the text of the license itself)
>> 
>> How could you possibly come to the conclusion that forcing anyone to
>> release private modifications would be in compliance with the spirit
>> of the license?  can != must

> I was trying to be sarcastic and inject a little humor here. Guess I should 
> have used the old <sarcasm> tag :)

Aah.  I'm not sure I'd have understood it either.

>> >> > Why should I repeat Linus' explanation of the ways that GPLv3 violates
>> >> > the spirit of GPLv2?
>> >>
>> >> Don't worry about parrotting here, he hasn't provided that explanation
>> >> yet ;-)  Please give it a try.
>> >
>> > But he has. Whether you have accepted that his explanations are
>> > valid or not doesn't change the fact.
>> 
>> His explanation is based on a reading of the license that doesn't
>> match what its authors meant.  I guess the authors know better what
>> they meant the spirit of the license to be than someone else who
>> studied it a lot but that until very recently couldn't even tell the
>> spirit from the legal terms.

> And his interpretation is no less valid than that of anyone else. In
> fact, after a recent conversation with a couple of lawyers that I
> know, I can state that his interpretation isn't that far off from
> theirs.

Interpretation as applied to the legal terms, yes.  As for the spirit
of the license, the authors ought to know better than anyone else what
they meant.  Sure, other interpretations might lead to different
understandings as to what the readers *think* it means, but that
doesn't change what it was *intended* to mean.

> Then you're lucky. I've had a lot of people say something similar to the 
> following: "Oh, I've heard about that. So which version of the GNU-Linux 
> kernel are you running?" 

Oh my.  That's indeed unfortunate and unfair.

> As I've stated before - I can find nothing in the history of the GPL or the 
> FSF that makes the "on the same hardware" requirement clear and part of 
> the "spirit" of "Free Software".

Put the considerations above, about a single computer or a
uniformly-limited computing platform, and you'll see that this "on the
same hardware" argument is just a means to deny people freedom.  If I
could stop you from running modified versions on one piece of
hardware, then I could on two, and 3, and then soon it's all of them,
and we're back to square zero in terms of freedom.

>> > What I won't do is release whatever tools and such that are needed to
>> > make the hardware run a different version of the kernel. Why? Because:
>> > the hardware was designed so that a specific version of the kernel runs
>> > without problems, there is hardware that is very picky and running a
>> > customized kernel could cause that hardware to fail, etc...

>> Why do you care?  It's no longer your hardware, it's theirs.

> Legal requirements in some countries that require manufacturers to provide 
> support for their product for a period of time after it has been purchased.

If you replace a component in the hardware, are you still required to
provide support or offer warranty?  Why should this be different just
because it's a software component?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Lennart Sorensen]

On Thu, Jun 14, 2007 at 02:26:30PM -0300, Alexandre Oliva wrote:
> In the program you received under GPLv1.
> 
> Hey, you said there was code under GPLv1.1 in the Linux tree.  Then,
> there should be a copy of GPLv1.1 in there, otherwise AFAICT the
> distribution of that code is copyright infringement.  IANAL.

So now the copy of the GPL v2 isn't good enough for the GPLv1.1 code?
Maybe that code said 'or later' in the license and hence someone added
it to a GPL v2 project since that sounds perfectly OK.

--
Len Sorensen

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Thu, 14 Jun 2007, Lennart Sorensen wrote:
> 
> So now the copy of the GPL v2 isn't good enough for the GPLv1.1 code?
> Maybe that code said 'or later' in the license and hence someone added
> it to a GPL v2 project since that sounds perfectly OK.

Where did that GPLv1.1 nonsense come from?

There is no GPLv1.1 code in the tree. By the time I selected the GPL for 
the kernel license, the GPLv1.1 had long since been discontinued. The 
kernel was *never* GPLv1.1-only compatible. That's just total nonsense.

There was indeed a kernel license before the GPLv2, but it wasn't the GPL 
at all, it was my own made-up thing. Appended here, for those who are too 
lazy to actually look up and check the original Linux-0.01 announcement.

		Linus

---
This kernel is (C) 1991 Linus Torvalds, but all or part of it may be
redistributed provided you do the following:

	- Full source must be available (and free), if not with the
	  distribution then at least on asking for it.

	- Copyright notices must be intact. (In fact, if you distribute
	  only parts of it you may have to add copyrights, as there aren't
	  (C)'s in all files.) Small partial excerpts may be copied
	  without bothering with copyrights.

	- You may not distibute this for a fee, not even "handling"
	  costs.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Thursday 14 June 2007 14:53:47 Linus Torvalds wrote:
> On Thu, 14 Jun 2007, Lennart Sorensen wrote:
> > So now the copy of the GPL v2 isn't good enough for the GPLv1.1 code?
> > Maybe that code said 'or later' in the license and hence someone added
> > it to a GPL v2 project since that sounds perfectly OK.
>
> Where did that GPLv1.1 nonsense come from?
>
> There is no GPLv1.1 code in the tree. By the time I selected the GPL for
> the kernel license, the GPLv1.1 had long since been discontinued. The
> kernel was *never* GPLv1.1-only compatible. That's just total nonsense.
>
> There was indeed a kernel license before the GPLv2, but it wasn't the GPL
> at all, it was my own made-up thing. Appended here, for those who are too
> lazy to actually look up and check the original Linux-0.01 announcement.
>

A hundred or so messages back someone stated that the parport driver in Linux 
is GPLv1.1 - however, on checking on this statement for myself I've found 
that there is no statement about it being v1.1 and, in fact, given that Linux 
itself is GPLv2 there is no possible way any code covered by GPLv1.1 can 
exist.

DRH

> 		Linus
>
> ---
> This kernel is (C) 1991 Linus Torvalds, but all or part of it may be
> redistributed provided you do the following:
>
> 	- Full source must be available (and free), if not with the
> 	  distribution then at least on asking for it.
>
> 	- Copyright notices must be intact. (In fact, if you distribute
> 	  only parts of it you may have to add copyrights, as there aren't
> 	  (C)'s in all files.) Small partial excerpts may be copied
> 	  without bothering with copyrights.
>
> 	- You may not distibute this for a fee, not even "handling"
> 	  costs.



-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alan Cox]

> A hundred or so messages back someone stated that the parport driver in Linux 
> is GPLv1.1 - however, on checking on this statement for myself I've found 
> that there is no statement about it being v1.1 and, in fact, given that Linux 
> itself is GPLv2 there is no possible way any code covered by GPLv1.1 can 
> exist.

Wrong again.

If a piece of code was merged into the kernel with a GPL v1 "or later"
license then it still has a GPL v1 "or later" license.  The "or later"
makes it compatible with the v2 code but does not change the fundamental
copyright on the original work that was combined. Thus if you could
identify specifically a GPL v1 work within the kernel you could use that
GPL v1 work as per GPL v1 providing you didn't mix it with v2 code.

If I take a public domain book and create a derivative work from it the
original work does not magically become restricted.

Alan

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bernd Paysan]

[-- Attachment #1: Type: text/plain, Size: 470 bytes --]

On Friday 15 June 2007 00:36, Daniel Hazelton wrote:
> A hundred or so messages back someone stated that the parport driver in
> Linux is GPLv1.1

Probably a misinterpretation - there are comments in the parport driver 
mentioning the GFDL version 1.1. If you just grep through, you might think 
it's GPL version 1.1 (but the code is really v2 or later).

-- 
Bernd Paysan
"If you want it done right, you have to do it yourself"
http://www.jwdt.com/~paysan/

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, lsorense@csclub.uwaterloo.ca (Lennart Sorensen) wrote:

> On Thu, Jun 14, 2007 at 02:26:30PM -0300, Alexandre Oliva wrote:
>> In the program you received under GPLv1.
>> 
>> Hey, you said there was code under GPLv1.1 in the Linux tree.  Then,
>> there should be a copy of GPLv1.1 in there, otherwise AFAICT the
>> distribution of that code is copyright infringement.  IANAL.

> So now the copy of the GPL v2 isn't good enough for the GPLv1.1 code?

If it exists and it's 1.1-only, I believe it wouldn't, but IANAL.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Thursday 14 June 2007 13:26:30 Alexandre Oliva wrote:
> On Jun 14, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> > On Thursday 14 June 2007 03:11:45 Alexandre Oliva wrote:
> >> On Jun 14, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> >> > Ah, well... In the case of "Windos" and other proprietary OS's I try
> >> > to educate people and get them to switch.
> >>
> >> Good.  So I presume you'd tell them to switch away from a
> >> turned-proprietary GNU/Linux operating system as well, right?
> >
> > If that happened I'd be lost. I've tried the various BSD's and found they
> > had problems with hardware support and getting a new version of the BSD
> > kernel to compile and boot is something of a black art.
> >
> > The point is moot, though. It can never happen.
>
> Look again, it's already happened in the TiVO and other devices.
>
> The software that ships in them is no longer Free Software.
>

In *YOUR* opinion and by *YOUR* definition of the term. Yes, I have seen some 
evidence that TiVO hasn't made some of the modifications they made public - 
doesn't mean that they won't, just it hasn't *YET* been done. (Not that I'm 
so omniscient I can say, definitively, whether they will or won't - or even 
that they haven't done it already).

By my own definition replacement != modification.

>
> Consider a new microprocessor.
>
> Consider that Linux is ported to it by the microprocessor
> manufacturer.
>
> Consider that the manufacturer only sells devices with that
> microprocessor with TiVO-like locks.
>
> How exactly can you enjoy the freedoms WRT the GPLed software you got
> from the manufacturer?

The same as I would with a TiVO. I have the right to copy, modify, distribute 
and run the code - even if I can't do any of those things on the hardware the 
original binary operates on.

>
>
> Now consider that you have a single computer, and that's built by TiVO.
>
> How exactly can you enjoy the freedoms the author meant you to have,
> if the TiVO box won't run the program after you modify it?

Simple: I don't buy it. Each and every piece of hardware I buy has a rather 
laborious research process before I actually spend the money on it. This 
makes it a certainty that I can use the hardware in the manner I want without 
problems like your hypothetical.

Whats worse - forcing your morals and ideals on someone or giving them the 
same freedom of choice you had?

Before you answer remember that that is *EXACTLY* what is being done with 
GPLv3. With GPLv2 and prior there was a simple guarantee that 
every "Licensee" had exactly the same rights. With GPLv3 you are forcing your 
ethics and morals on people - and isn't this exactly what the Roman Catholic 
church did during the Spanish Inquisition?

> > If this "run modified copies on the same hardware you received the
> > original on" *IS* the "spirit" of the license, then why isn't it
> > stated anywhere before GPLv3?
>
> For the same reasons that the pro-DRM laws weren't mentioned before,
> and the patent retaliation clauses weren't mention before: these
> specific cases hadn't been studied, only the general idea of
> respecting users' freedoms was.

Bzzt! Wrong! The reason is that it wasn't necessary - at all. It still isn't, 
but a group that feels modification == replacement wants it to be, so it has 
suddenly become necessary. (Note that anti-DRM stuff *IS* good - DRM is part 
of an attempt by failing business models to stop the failure)

> > I'll grant you that. But, at this point, where can I find a copy of
> > the GPLv1 without having to dig around the net ?
>
> In the program you received under GPLv1.
>
> Hey, you said there was code under GPLv1.1 in the Linux tree.  Then,
> there should be a copy of GPLv1.1 in there, otherwise AFAICT the
> distribution of that code is copyright infringement.  IANAL.

Ah, but I never said I had a GPLv1 program. If GPLv1 is still valid and 
available I should be able to find a copy of it *RIGHT* *NOW* to license a 
new project if I want to use GPLv1 as its license. So your logic is again 
flawed.

> >> In contrast, your TiVO may get a software upgrade without your
> >> permission that will take your rights away from that point on, and
> >> there's very little you can do about it, other than unplugging it from
> >> the network to avoid the upgrade if it's not too late already.
> >
> > And because its a device that connects to their network - and TiVO
> > isn't a telecommunications company - they have the right to upgrade
> > and configure the software inside however they want. (In the US at
> > least)
>
> But do they have the right to not pass this right on, under the GPL?

Yes, they do. It isn't a right they have as "copyright holders" - in fact, it 
isn't a part of their rights under the copyright at all. It's a part of their 
rights as the owners of the network. 

> >> > A lot of them would probably have private modifications that would
> >> > never be distributed - and under the GPLv2 it is clear that you can
> >> > keep modifications private as long as you don't distribute them.
> >>
> >> Likewise with GPLv3.
> >
> > I can see this, but will a company see this?
>
> In what sense does the GPLv3 make this particular point any less
> obscure?

Never claimed it was less obscure, just that you've usually got a board-room 
filled with middle-aged men that might have problems agreeing that it is a 
clear-cut case.

> > True. But that doesn't save them from lawsuits trying to force them
> > to obey the terms of the new revision even though they received the
> > software under an earlier version.
>
> Nothing saves anyone from silly lawsuits.  This one would likely be
> laughed out of court in no time.  Anyone worried about this should
> also be concerned about their neighbor suing them for copyright
> infringment every time they set their stereo loud enough for the
> neighbors to listen and be annoyed.  (Hint: only the copyright holder
> would stand a chance of winning such a lawsuit)

Yes, but the fact that it would cost money to get the suit dropped is a 
problem. 

> >> > (and don't try to argue that even though those modifications are
> >> > truly private (to the company) they should be released anyway to
> >> > comply with the "spirit" of the license. It is made clear that it
> >> > isn't by the text of the license itself)
> >>
> >> How could you possibly come to the conclusion that forcing anyone to
> >> release private modifications would be in compliance with the spirit
> >> of the license?  can != must
> >
> > I was trying to be sarcastic and inject a little humor here. Guess I
> > should have used the old <sarcasm> tag :)
>
> Aah.  I'm not sure I'd have understood it either.
>
> >> >> > Why should I repeat Linus' explanation of the ways that GPLv3
> >> >> > violates the spirit of GPLv2?
> >> >>
> >> >> Don't worry about parrotting here, he hasn't provided that
> >> >> explanation yet ;-)  Please give it a try.
> >> >
> >> > But he has. Whether you have accepted that his explanations are
> >> > valid or not doesn't change the fact.
> >>
> >> His explanation is based on a reading of the license that doesn't
> >> match what its authors meant.  I guess the authors know better what
> >> they meant the spirit of the license to be than someone else who
> >> studied it a lot but that until very recently couldn't even tell the
> >> spirit from the legal terms.
> >
> > And his interpretation is no less valid than that of anyone else. In
> > fact, after a recent conversation with a couple of lawyers that I
> > know, I can state that his interpretation isn't that far off from
> > theirs.
>
> Interpretation as applied to the legal terms, yes.  As for the spirit
> of the license, the authors ought to know better than anyone else what
> they meant.  Sure, other interpretations might lead to different
> understandings as to what the readers *think* it means, but that
> doesn't change what it was *intended* to mean.

Doesn't matter what the author intended it to mean - at all. What matters is 
how its interpreted when/if it shows up in court. This can be seen *ALL* the 
time with laws across the globe. In the US there is this thing called 
the "RICO" Laws- "Racketeering Influenced Criminal Organization" - that give 
the government the ability to seize anything that is deemed a "profit of drug 
sales" or of a "Criminal Organization". It has been used for that purpose, 
but its interpretation has caused it to be used to seize money that has had 
no source in either.

> > Then you're lucky. I've had a lot of people say something similar to the
> > following: "Oh, I've heard about that. So which version of the GNU-Linux
> > kernel are you running?"
>
> Oh my.  That's indeed unfortunate and unfair.
>
> > As I've stated before - I can find nothing in the history of the GPL or
> > the FSF that makes the "on the same hardware" requirement clear and part
> > of the "spirit" of "Free Software".
>
> Put the considerations above, about a single computer or a
> uniformly-limited computing platform, and you'll see that this "on the
> same hardware" argument is just a means to deny people freedom.  If I
> could stop you from running modified versions on one piece of
> hardware, then I could on two, and 3, and then soon it's all of them,
> and we're back to square zero in terms of freedom.
>
> >> > What I won't do is release whatever tools and such that are needed to
> >> > make the hardware run a different version of the kernel. Why? Because:
> >> > the hardware was designed so that a specific version of the kernel
> >> > runs without problems, there is hardware that is very picky and
> >> > running a customized kernel could cause that hardware to fail, etc...
> >>
> >> Why do you care?  It's no longer your hardware, it's theirs.
> >
> > Legal requirements in some countries that require manufacturers to
> > provide support for their product for a period of time after it has been
> > purchased.
>
> If you replace a component in the hardware, are you still required to
> provide support or offer warranty?  Why should this be different just
> because it's a software component?

Artificial distinctions in the law

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> With GPLv2 and prior there was a simple guarantee that every
> "Licensee" had exactly the same rights. With GPLv3 you are forcing
> your ethics and morals on people - and isn't this exactly what the
> Roman Catholic church did during the Spanish Inquisition?

I fail to see the distinction you're making between GPLv2 and GPLv3.
AFAICT, with GPLv3, there still is a simple guarantee that every
licensee has exactly the same rights.

Sure, GPLv3 follows the spirit of the GPLs more strictly than GPLv2
possibly could.  How is that "forcing ethics and morals" any more than
GPLv2 was?

> Ah, but I never said I had a GPLv1 program.

I thought you had a copy of Linux and, per what you'd said before,
there was GPLv1 code in it.  I was just trying to make it easy for
you.

> If GPLv1 is still valid and available I should be able to find a
> copy of it *RIGHT* *NOW* to license a new project if I want to use
> GPLv1 as its license.

http://www.gnu.org/copyleft/copying-1.0.html

>> > And because its a device that connects to their network - and TiVO
>> > isn't a telecommunications company - they have the right to upgrade
>> > and configure the software inside however they want. (In the US at
>> > least)
>> 
>> But do they have the right to not pass this right on, under the GPL?

> Yes, they do. It isn't a right they have as "copyright holders" - in fact, it 
> isn't a part of their rights under the copyright at all. It's a part of their 
> rights as the owners of the network. 

How about the "no further restrictions" bit?

> Never claimed it was less obscure, just that you've usually got a board-room 
> filled with middle-aged men that might have problems agreeing that it is a 
> clear-cut case.

> Yes, but the fact that it would cost money to get the suit dropped is a 
> problem. 

Again, how are these arguments against GPLv3?  They apply equally to
any other license, including GPLv2.

>> Interpretation as applied to the legal terms, yes.  As for the spirit
>> of the license, the authors ought to know better than anyone else what
>> they meant.  Sure, other interpretations might lead to different
>> understandings as to what the readers *think* it means, but that
>> doesn't change what it was *intended* to mean.

> Doesn't matter what the author intended it to mean - at all. What matters is 
> how its interpreted when/if it shows up in court.

You're talking about the legal terms.  The spirit of the license is a
very different matter.  It can guide the interpretation of the legal
terms, but the author is at a better position than anyone else to know
what he meant.

>> If you replace a component in the hardware, are you still required to
>> provide support or offer warranty?  Why should this be different just
>> because it's a software component?

> Artificial distinctions in the law

Well, then, lock down the software.  Make it irreplaceable, even by
yourself.  Problem solved.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Thursday 14 June 2007 17:19:51 Alexandre Oliva wrote:
> On Jun 14, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> > With GPLv2 and prior there was a simple guarantee that every
> > "Licensee" had exactly the same rights. With GPLv3 you are forcing
> > your ethics and morals on people - and isn't this exactly what the
> > Roman Catholic church did during the Spanish Inquisition?
>
> I fail to see the distinction you're making between GPLv2 and GPLv3.
> AFAICT, with GPLv3, there still is a simple guarantee that every
> licensee has exactly the same rights.
>
> Sure, GPLv3 follows the spirit of the GPLs more strictly than GPLv2
> possibly could.  How is that "forcing ethics and morals" any more than
> GPLv2 was?

Because GPLv2 doesn't enforce limitations on the hardware a GPL'd work can be 
put on. It doesn't make artificial distinctions between "Commercial", 
"Industrial" and "User". What it does is *ATTEMPT* to ensure that nobody 
receiving a copy of a GPL'd work has the same rights as any other person that 
gets a copy. GPLv3 gives people *additional* rights beyond those. In 
the "TiVO" case it forces somebody releasing a HW platform to grant 
*additional* rights if they are going to use software covered by the GPLv3. 
The reason for forcing the giving of those additional rights is "the FSF and 
GPLv3 committees think that what TiVO did is 'morally and ethically' wrong, 
so were are enforcing our morals and ethics".

Note that these are the rights that TiVO got from the GPLv2:
1) The ability to make copies of Linux
2) The ability to modify Linux
3) The ability to distribute Linux
*NOTE* that those are the rights *GUARANTEED* by the GPL - despite what anyone 
*WISHES* it to say, or what the "Intent" or "Spirit" of the license may be, 
those are the only guaranteed rights.

In shipping their devices with an "object code" version of Linux on them they 
exercised their right to perform such a distribution, granted under section 3 
of the GPLv2. They made modifications to the Linux so it functioned properly 
on their devices, as allowed by the GPL. They have made numerous copies of 
Linux, as allowed by the GPL. And, as required by the GPLv2, they made the 
source code form of their changes available. In fact, they went beyond the 
requirements of the GPL (which only requires you make the source available to 
people you have given an "object code" version to) in making it fully 
available to the public *AND* in contributing those changes back to Linux.

What rights did they give to "downstream" recipients of the "object code" 
version? *EXACTLY* those they received from the GPLv2.

What rights do they have as creators of a *PROPRIETARY* hardware platform:
1) The right to restrict what programs it will run
2) The right to update it as they choose, even if it makes it incompatible 
with earlier versions

Does the GPL *require* them to give up those rights? Version 3 does, but not 
any earlier version. Why does version 3 do this? Because one or more of the 
people behind its design and language feel that executing a piece of software 
on any given hardware platform automatically entitles them to all rights to 
the hardware that the creator of the hardware has. 

> > Ah, but I never said I had a GPLv1 program.
>
> I thought you had a copy of Linux and, per what you'd said before,
> there was GPLv1 code in it.  I was just trying to make it easy for
> you.
>
> > If GPLv1 is still valid and available I should be able to find a
> > copy of it *RIGHT* *NOW* to license a new project if I want to use
> > GPLv1 as its license.
>
> http://www.gnu.org/copyleft/copying-1.0.html

Ah, see, I didn't even know it was still there. I hadn't seen it in a complete 
form anywhere.

> >> > And because its a device that connects to their network - and TiVO
> >> > isn't a telecommunications company - they have the right to upgrade
> >> > and configure the software inside however they want. (In the US at
> >> > least)
> >>
> >> But do they have the right to not pass this right on, under the GPL?
> >
> > Yes, they do. It isn't a right they have as "copyright holders" - in
> > fact, it isn't a part of their rights under the copyright at all. It's a
> > part of their rights as the owners of the network.
>
> How about the "no further restrictions" bit?

As applies to the software. The rights that the GPL has (historically) granted 
are what I stated above. TiVO, and companies like them, are *NOT* imposing 
any restrictions on rights granted under GPLv2 and prior. Remember, because 
I'm getting tired of repeating myself: replace != modify

> > Never claimed it was less obscure, just that you've usually got a
> > board-room filled with middle-aged men that might have problems agreeing
> > that it is a clear-cut case.
> >
> > Yes, but the fact that it would cost money to get the suit dropped is a
> > problem.
>
> Again, how are these arguments against GPLv3?  They apply equally to
> any other license, including GPLv2.

Granted. But GPLv2 has been examined and re-examined by lawyer and lay-person 
alike for more than 10 years. If you recall, however, my original statement 
was about the "revise or change the license at any time" clause. If that is 
possible, then it is also possible for the mentioned "frivolous lawsuit" by 
someone applying the "revised" license to something licensed with the 
non-revised version.

Your argument is flawed, however. Most companies have done the work for 
defense with the "other licenses". And if it isn't possible for a "Licensor" 
to say "My work is released under version X of this license only" (with the 
GPL) - as you seem to have argued at several points - then it wouldn't matter 
if they've done the legwork to protect them from all but the most frivolous 
of lawsuits. All that needs to happen is for someone to release the same work 
under a later license (with modifications) and the company is now open to a 
new, unplanned for class of "frivolous lawsuits".

> >> Interpretation as applied to the legal terms, yes.  As for the spirit
> >> of the license, the authors ought to know better than anyone else what
> >> they meant.  Sure, other interpretations might lead to different
> >> understandings as to what the readers *think* it means, but that
> >> doesn't change what it was *intended* to mean.
> >
> > Doesn't matter what the author intended it to mean - at all. What matters
> > is how its interpreted when/if it shows up in court.
>
> You're talking about the legal terms.  The spirit of the license is a
> very different matter.  It can guide the interpretation of the legal
> terms, but the author is at a better position than anyone else to know
> what he meant.

I do agree that the intent can guide interpretation, but in the US it is very 
rare. Consider the "Second Amendment", which states "The formation of 
militias being necessary for the defense of the state, the right of the 
people to keep and bear arms shall not be abridged." The *INTENT* is 
explained by the authors in several places, but the interpretation is still 
argued about. The meaning is *VERY* clear to me - and makes all "Gun Control" 
laws and such in direct violation of the US Constitution. But that doesn't 
matter, because nobody pays attention to the intent.

> >> If you replace a component in the hardware, are you still required to
> >> provide support or offer warranty?  Why should this be different just
> >> because it's a software component?
> >
> > Artificial distinctions in the law
>
> Well, then, lock down the software.  Make it irreplaceable, even by
> yourself.  Problem solved.

Yes, problem solved. But at the cost of taking hardware back 20 years.

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Michael Poole]

Daniel Hazelton writes:

> What rights did they give to "downstream" recipients of the "object code" 
> version? *EXACTLY* those they received from the GPLv2.

Doing the e-mail equivalent of yelling about this will not change the
fact that people who think Tivo did something wrong -- legally and/or
morally -- consider DRM locks on a piece of software to be part of the
"work based on the Program" that is governed by the GPL.

The fundamental reason for this is that neither the executable code
nor the digital signature serves the desired function alone.  The user
received a copy of the executable for a particular purpose: to run the
program on a particular platform.  With DRM signatures, only the
combination of program and signature will perform that function, and
separating the two based on strictly read legal definitions is risky.

The question of whether DRM signatures are covered by the license must
be resolved before one can determine whether Tivo gave "*EXACTLY*" the
same rights to object-code recipients as Tivo received.  GPLv2 is
worded such that the answer to this does not depend on whether one is
in file A and the other in file B, or whether one is on hard drive C
and the other is in flash device D, as long as they are delivered as
part of one unit; it *might* matter if, say, one is received on
physical media and the other is downloaded on demand.

(Linus likes to say that FSF counsel thinks that Tivo did not violate
GPLv2.  I suspect the actual situation is that FSF counsel believes
that there is no case law on point, and that it could go either way,
making it improper to publicly claim that Tivo violated any copyright.
Until a court rules on a close-enough case, the question of whether
GPLv2 covers DRM signatures remains open.  In the mean time, it makes
more sense for the FSF to issue a new license that squarely addresses
this -- such as the GPLv3 -- and persuade as many developers as
possible that using it is the best way to protect free software.)

Michael Poole

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Thursday 14 June 2007 22:13:13 Michael Poole wrote:
> Daniel Hazelton writes:
> > What rights did they give to "downstream" recipients of the "object code"
> > version? *EXACTLY* those they received from the GPLv2.
>
> Doing the e-mail equivalent of yelling about this will not change the

Sorry, I wasn't trying to "yell" - just provide a note that at that point I 
would be providing verbal stress.

> fact that people who think Tivo did something wrong -- legally and/or
> morally -- consider DRM locks on a piece of software to be part of the
> "work based on the Program" that is governed by the GPL.

All I've done is get a little annoyed that, despite evidence that it isn't 
legally wrong - at least under the laws I am most familiar with - people 
continue repeat that it is.

I can't argue that it isn't "morally" wrong. While it may not be against my 
morals, it could be against those of another person. It has never been my 
intent to try and convince people that their morals are wrong.

> The fundamental reason for this is that neither the executable code
> nor the digital signature serves the desired function alone.  The user
> received a copy of the executable for a particular purpose: to run the
> program on a particular platform.  With DRM signatures, only the
> combination of program and signature will perform that function, and
> separating the two based on strictly read legal definitions is risky.

I agree.

> The question of whether DRM signatures are covered by the license must
> be resolved before one can determine whether Tivo gave "*EXACTLY*" the
> same rights to object-code recipients as Tivo received.  GPLv2 is
> worded such that the answer to this does not depend on whether one is
> in file A and the other in file B, or whether one is on hard drive C
> and the other is in flash device D, as long as they are delivered as
> part of one unit; it *might* matter if, say, one is received on
> physical media and the other is downloaded on demand.

I have read the GPLv2 at least three times since it was pointed out that I had 
forgotten part of it. At no point can I find a point where Tivo broke the 
GPLv2 requirement that they give the recipients of the object code the same 
rights they received when they acquired a copy of the object or source code.

> (Linus likes to say that FSF counsel thinks that Tivo did not violate
> GPLv2.  I suspect the actual situation is that FSF counsel believes
> that there is no case law on point, and that it could go either way,
> making it improper to publicly claim that Tivo violated any copyright.
> Until a court rules on a close-enough case, the question of whether
> GPLv2 covers DRM signatures remains open.  In the mean time, it makes
> more sense for the FSF to issue a new license that squarely addresses
> this -- such as the GPLv3 -- and persuade as many developers as
> possible that using it is the best way to protect free software.)

In examining the GPLv2 and the situation from a strictly factual basis I can 
believe Linus' statement fully. The facts are as I stated them in a previous 
mail.

DRH

> Michael Poole



-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Michael Poole]

Daniel Hazelton writes:

> On Thursday 14 June 2007 22:13:13 Michael Poole wrote:
>
>> The fundamental reason for this is that neither the executable code
>> nor the digital signature serves the desired function alone.  The user
>> received a copy of the executable for a particular purpose: to run the
>> program on a particular platform.  With DRM signatures, only the
>> combination of program and signature will perform that function, and
>> separating the two based on strictly read legal definitions is risky.
>
> I agree.
>
>> The question of whether DRM signatures are covered by the license must
>> be resolved before one can determine whether Tivo gave "*EXACTLY*" the
>> same rights to object-code recipients as Tivo received.  GPLv2 is
>> worded such that the answer to this does not depend on whether one is
>> in file A and the other in file B, or whether one is on hard drive C
>> and the other is in flash device D, as long as they are delivered as
>> part of one unit; it *might* matter if, say, one is received on
>> physical media and the other is downloaded on demand.
>
> I have read the GPLv2 at least three times since it was pointed out that I had 
> forgotten part of it. At no point can I find a point where Tivo broke the 
> GPLv2 requirement that they give the recipients of the object code the same 
> rights they received when they acquired a copy of the object or source code.

I am trying to reconcile your responses to those two paragraphs.

If the DRM signature and program executable are coupled such that they
are not useful when separated, the implication to me is that they form
one work that is based on the original Program.  This is beyond the
GPL's permission for "mere aggregation".

If they are one work, and the original Program was licensed under the
GPLv2, the combined work must also be licensed under the terms of the
GPLv2.

The input files required to generate a DRM-valid digital signature are
part the preferred form for modifying that work.

If those bits are not distributed along with the rest of the GPL'ed
work, the distributor is either not giving the same rights to the end
user, not distributing the source code for the work, or both.  Which
is it?

Michael Poole

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Thu, 14 Jun 2007, Michael Poole wrote:
>
> If the DRM signature and program executable are coupled such that they
> are not useful when separated, the implication to me is that they form
> one work that is based on the original Program.  This is beyond the
> GPL's permission for "mere aggregation".

So you want to make things like a 160-bit SHA1 hash of a binary be a 
"derived work" of that software?

Trust me, you *really* don't want to go there. It's an insane legal 
standpoint, and if you were right, we'd be in a *world* of trouble.

Think about something as simple as security software that creates 
filesystem checksums for verifying the integrity of the filesystem, and 
protects against tampering.

Do you *really* want to claim that the SHA1 checksum of your "oracle" 
binary is owned by Oracle, and you need to have a special license to copy 
that checksum around and verify it?

Do you *really* want to claim that the RIAA owns the CDDB checksums (well, 
I guess "feedb", these days) of the CD's that get uploaded for music 
databases? 

Do you realize that in your INSANE world, there is no notion of "fair 
use", and you just tried to extend the notion of copyright so far that you 
turned your utopia into a total distopia.

In other words, anybody who claims that copyright in a program extends to 
the cryptographic hash of the binary, and at the same time makes a "free 
software" kind of argument is so damn clueless that it's not even funny. 
You're arguing for "freedom" by using logic that is the very *antithesis* 
of freedom.

That's just incredibly stupid and incredibly short-sighted. 

If that were to seriously be an FSF argument, then I would officially lump 
the FSF as a *much*worse* danger to the free world than the RIAA and the 
MPAA combined!

I seriously doubt you really thought your idea through! Because it goes 
beyond stupid.

			Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Michael Poole]

Linus Torvalds writes:

> On Thu, 14 Jun 2007, Michael Poole wrote:
>>
>> If the DRM signature and program executable are coupled such that they
>> are not useful when separated, the implication to me is that they form
>> one work that is based on the original Program.  This is beyond the
>> GPL's permission for "mere aggregation".
>
> So you want to make things like a 160-bit SHA1 hash of a binary be a 
> "derived work" of that software?

No.  That is why I specified "not useful when separated".  I also
intentionally avoided the phrase "derived work": the legal definition
of derived work is based on entirely different factors.

If the signature is one that serves to indicate origin, to detect
tampering, or the other things you mentioned, the program's binary is
useful when separated from the signature.  My objection arises when a
functionally equivalent binary -- including advertised functions such
as "runs on platform XYZ" -- cannot be produced from the distributed
source code.

Michael Poole

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Thu, 14 Jun 2007, Michael Poole wrote:
> 
> If the signature is one that serves to indicate origin, to detect
> tampering, or the other things you mentioned, the program's binary is
> useful when separated from the signature.  My objection arises when a
> functionally equivalent binary -- including advertised functions such
> as "runs on platform XYZ" -- cannot be produced from the distributed
> source code.

Ahh.

Ok, that's a totally different issue, and is one where I heartily agree 
with you. I would actually *love* for the GPL (any version) to have a 
"guarantee of authenticity", where if you distribute a binary, there has 
to be some documented way to get *exactly* that binary out of the source 
code that got distributed.

Of course, SHA1's can be used to verify that, although, quite frankly, I'd 
expect that a simple "cmp" would be the more straightforward approach.

So the "verification" can be used both to lock down a particular binary 
_and_ to authenticate that the binary really came from the source code it 
was claimed to come from.

Of course, in practice, it's actually really nasty to do that 
verification. Many compilers actually do things like insert date-stamps in 
the object files etc. So it's probably not all that practical.

			Linus

Re: Dual-Licensing Linux And Medical Devices

[Tim Post]

On Thu, 2007-06-14 at 21:23 -0700, Linus Torvalds wrote:
> 
> On Thu, 14 Jun 2007, Michael Poole wrote:
> > 
> > If the signature is one that serves to indicate origin, to detect
> > tampering, or the other things you mentioned, the program's binary is
> > useful when separated from the signature.  My objection arises when a
> > functionally equivalent binary -- including advertised functions such
> > as "runs on platform XYZ" -- cannot be produced from the distributed
> > source code.
> 
> Ahh.
> 
> Ok, that's a totally different issue, and is one where I heartily agree 
> with you. I would actually *love* for the GPL (any version) to have a 
> "guarantee of authenticity", where if you distribute a binary, there has 
> to be some documented way to get *exactly* that binary out of the source 
> code that got distributed.

I would hope that this is *required*, somehow, when dealing with medical
equipment. I don't think those appliances even have the capacity to
build every upgrade from source. None that I've tinkered with do. These
things almost need a license of their own.

As long as the signing mechanism can't be used to force clinics to pay
for the privilege of upgrading free software, that is. It would truly
suck if an ultrasound loaded with free software sat in a corner useless
because a free clinic could not afford to pay for what they already paid
for.

If you guys can find a way to make that practical given my above
concerns, that would be entirely useful. I hate the fact that this kind
of trust is needed because it is so very easily mis-used, but people
dying due to hacked IV regulators really wouldn't much care about those
politics.

I think, also privacy implications for patients. A rootkit in a MRI
would be very bad.

Regardless, like it or not, kernel code is in or headed for medical
devices, so I hope some more brain power is burned on this. 

Best,
--Tim

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 15, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:

> On Thu, 14 Jun 2007, Michael Poole wrote:
>> 
>> If the DRM signature and program executable are coupled such that they
>> are not useful when separated, the implication to me is that they form
>> one work that is based on the original Program.  This is beyond the
>> GPL's permission for "mere aggregation".

> So you want to make things like a 160-bit SHA1 hash of a binary be a 
> "derived work" of that software?

How about the combination of the software binary with the hash?

Considering that the hash is a functional part of the software, as in,
if you take that out, it no longer works?

> If that were to seriously be an FSF argument,

Remember: I don't speak for the FSF, and I don't speak for FSFLA, just
like I don't speak for Red Hat.

Just like you shouldn't redirect your qualms with the FSF to me, you
shouldn't direct your qualms with me to the FSF.  That would be very
wrong.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Ingo Molnar]

* Linus Torvalds <torvalds@linux-foundation.org> wrote:

> On Thu, 14 Jun 2007, Michael Poole wrote:
> >
> > If the DRM signature and program executable are coupled such that 
> > they are not useful when separated, the implication to me is that 
> > they form one work that is based on the original Program.  This is 
> > beyond the GPL's permission for "mere aggregation".
> 
> So you want to make things like a 160-bit SHA1 hash of a binary be a 
> "derived work" of that software?
> 
> Trust me, you *really* don't want to go there. It's an insane legal 
> standpoint, and if you were right, we'd be in a *world* of trouble.
> 
> Think about something as simple as security software that creates 
> filesystem checksums for verifying the integrity of the filesystem, 
> and protects against tampering.
> 
> Do you *really* want to claim that the SHA1 checksum of your "oracle" 
> binary is owned by Oracle, and you need to have a special license to 
> copy that checksum around and verify it?

not only that, but it would instantly turn everyone who owns a hard 
drive or a CD-ROM into a copyright violator: the disks checksums the 
content of the disk _in a reversible way_. Same for RAID5 and RAID6 
techniques. By installing Quake3 on a Windows box one sure does not have 
permission to create a derived work of Windows and Quake3, right? =B-)

a checksum, a one-way hash, or even reversible parity bit(s) that 
'mixes' the copies of multiple works together clearly cannot be new work 
that falls under copyright protection.

Firstly, it is not a new work, because a 'work' has to be created by a 
human - and here the new content was created by a machine. Copyright 
protection only applies to sufficiently original works created by 
humans.

Secondly, it is _at most_ a new, partial copy of existing works and 
hence you need the permission to copy all the works in question. (but 
you needed that permission to create the harddisk anyway)

Thirdly, it could be argued that the sha1 is not even a copy, because it 
is irreversible and hence not even a single bit of the original work can 
be reconstructed from it hence it cannot even be a 'partial copy of the 
original'.

that's at least 3 robust levels of argument against the insane and 
absurd notion that the SHA1 key is somehow a derived work of the copy it 
checksums.

	Ingo

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 15, 2007, Ingo Molnar <mingo@elte.hu> wrote:

> Secondly, it is _at most_ a new, partial copy of existing works and 
> hence you need the permission to copy all the works in question.

Wouldn't you consider the signing key as one of these existing works?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Thursday 14 June 2007 23:04:37 Michael Poole wrote:
> Daniel Hazelton writes:
> > On Thursday 14 June 2007 22:13:13 Michael Poole wrote:
> >> The fundamental reason for this is that neither the executable code
> >> nor the digital signature serves the desired function alone.  The user
> >> received a copy of the executable for a particular purpose: to run the
> >> program on a particular platform.  With DRM signatures, only the
> >> combination of program and signature will perform that function, and
> >> separating the two based on strictly read legal definitions is risky.
> >
> > I agree.
> >
> >> The question of whether DRM signatures are covered by the license must
> >> be resolved before one can determine whether Tivo gave "*EXACTLY*" the
> >> same rights to object-code recipients as Tivo received.  GPLv2 is
> >> worded such that the answer to this does not depend on whether one is
> >> in file A and the other in file B, or whether one is on hard drive C
> >> and the other is in flash device D, as long as they are delivered as
> >> part of one unit; it *might* matter if, say, one is received on
> >> physical media and the other is downloaded on demand.
> >
> > I have read the GPLv2 at least three times since it was pointed out that
> > I had forgotten part of it. At no point can I find a point where Tivo
> > broke the GPLv2 requirement that they give the recipients of the object
> > code the same rights they received when they acquired a copy of the
> > object or source code.
>
> I am trying to reconcile your responses to those two paragraphs.
>
> If the DRM signature and program executable are coupled such that they
> are not useful when separated, the implication to me is that they form
> one work that is based on the original Program.  This is beyond the
> GPL's permission for "mere aggregation".
>
> If they are one work, and the original Program was licensed under the
> GPLv2, the combined work must also be licensed under the terms of the
> GPLv2.
>
> The input files required to generate a DRM-valid digital signature are
> part the preferred form for modifying that work.
>
> If those bits are not distributed along with the rest of the GPL'ed
> work, the distributor is either not giving the same rights to the end
> user, not distributing the source code for the work, or both.  Which
> is it?

Following your logic it would be a "failure to distribute the source code for 
a work".

However, since the signing is an automated process it cannot generate a "new" 
work - at least, not under the laws of the US - so the signature itself 
cannot have a copyright at all.

DRH
PS: This is the exact same reason that the GPL cannot apply to a Bison 
generated parser in the US. The "input" file that causes Bison to generate 
the output can have a copyright, but not the output - no matter what RMS or 
anyone else wants, and no matter what the GPL says about it.

>
> Michael Poole



-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Michael Poole]

Daniel Hazelton writes:

> Following your logic it would be a "failure to distribute the source code for 
> a work".
>
> However, since the signing is an automated process it cannot generate a "new" 
> work - at least, not under the laws of the US - so the signature itself 
> cannot have a copyright at all.
>
> DRH
> PS: This is the exact same reason that the GPL cannot apply to a Bison 
> generated parser in the US. The "input" file that causes Bison to generate 
> the output can have a copyright, but not the output - no matter what RMS or 
> anyone else wants, and no matter what the GPL says about it.

I do not suggest that copyright subsists in the signature or in the
signing key.  Whether it does is irrelevant to the signing key being
part of the source code (when the signature is needed for the binary
to work properly).

Similarly, copyright might not subsist in a simple linker script --
its content being determined by the operating system and perhaps the
rest of the program's source code -- but under the GPL, the linker
script would be part of the source code for a compiled version.

Michael Poole

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Ingo Molnar]

* Michael Poole <mdpoole@troilus.org> wrote:

> > However, since the signing is an automated process it cannot 
> > generate a "new" work - at least, not under the laws of the US - so 
> > the signature itself cannot have a copyright at all.
[...]
> 
> I do not suggest that copyright subsists in the signature or in the 
> signing key.  Whether it does is irrelevant to the signing key being 
> part of the source code (when the signature is needed for the binary 
> to work properly).

it is very much relevant. By admitting that the key is not part of the 
"work", you have lost all moral basis to claim control over it. Cutely 
"defining it" into the source code just hides what this really is: the 
key is a "payment" in exchange for the license, which payment goes 
outside the scope of the software itself. It has no relevance to the 
software work being "free", it reaches for paymeant beyond the work to 
advance the FSF's agenda.

yes, a copyright license can be used to control other works, it can be 
used to control the movement of non-copyrightable items as well (such as 
money), but the GPL always tried to stay out of that kind of business.

Where does this "reach out for more resources in exchange for the 
license" process stop? As the value of free software increases, will the 
FSF iterate the GPL to ask for more and more consideration for the 
privilege to license that software? (All in the name of achieving more 
freedom of course.)

> Similarly, copyright might not subsist in a simple linker script -- 
> its content being determined by the operating system and perhaps the 
> rest of the program's source code -- but under the GPL, the linker 
> script would be part of the source code for a compiled version.

the linker script is still part of the whole work though - even if that 
particular element might not be copyrightable in isolation. Likewise, 
the kernel contains code that is in the public domain - to which 
copyright protection does not extend either. But you cannot argue that 
the Tivo 'key' is part of the whole work. It is part of the _hardware_. 
The Tivo box is a compilation (at most a collection) of multiple works, 
and allowing the GPL to jump over derivation/modification lines is 
wrong. The GPLv2 certain doesnt do that land-grab.

	Ingo

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Michael Poole]

Ingo Molnar writes:

> * Michael Poole <mdpoole@troilus.org> wrote:
>
>> > However, since the signing is an automated process it cannot 
>> > generate a "new" work - at least, not under the laws of the US - so 
>> > the signature itself cannot have a copyright at all.
> [...]
>> 
>> I do not suggest that copyright subsists in the signature or in the 
>> signing key.  Whether it does is irrelevant to the signing key being 
>> part of the source code (when the signature is needed for the binary 
>> to work properly).
>
> it is very much relevant. By admitting that the key is not part of the 
> "work", you have lost all moral basis to claim control over it.

I have not admitted any such thing.  I have said the key and signature
do not have separate copyright protection.  Variables named "i" in a
file are not protected by copyright, but they are very much part of
the source code in that file.

>> Similarly, copyright might not subsist in a simple linker script -- 
>> its content being determined by the operating system and perhaps the 
>> rest of the program's source code -- but under the GPL, the linker 
>> script would be part of the source code for a compiled version.
>
> the linker script is still part of the whole work though - even if that 
> particular element might not be copyrightable in isolation. Likewise, 
> the kernel contains code that is in the public domain - to which 
> copyright protection does not extend either. But you cannot argue that 
> the Tivo 'key' is part of the whole work. It is part of the _hardware_. 
> The Tivo box is a compilation (at most a collection) of multiple works, 
> and allowing the GPL to jump over derivation/modification lines is 
> wrong. The GPLv2 certain doesnt do that land-grab.

Where in the Tivo hardware is the signing key?  There is a related key
in the hardware, but that one is not used to generate an integral part
of the kernel binary.

Michael Poole

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Ingo Molnar]

* Michael Poole <mdpoole@troilus.org> wrote:

> >> I do not suggest that copyright subsists in the signature or in the 
> >> signing key.  Whether it does is irrelevant to the signing key 
> >> being part of the source code (when the signature is needed for the 
> >> binary to work properly).
> >
> > it is very much relevant. By admitting that the key is not part of 
> > the "work", you have lost all moral basis to claim control over it.
> 
> I have not admitted any such thing.  I have said the key and signature 
> do not have separate copyright protection.  Variables named "i" in a 
> file are not protected by copyright, but they are very much part of 
> the source code in that file.

the problem with your argument is that the definition of what 
constitutes "work" is up to copyright law, _not_ the license writer. 
I.e. you cannot just cleverly define "source code" to include something 
unrelated and then pretend that it's all in one work. And that's exactly 
what the GPLv3 does: it creatively defines the hardware's key into the 
'source code' of the software and then asks for that to be provided 
_not_ because somehow the key derives from the software (it clearly does 
not), but as a "compensation" for the right to redistribute! I.e. it's 
trying to extend its scope to some item that is not part of the 
software. See?

	Ingo

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Michael Poole]

Ingo Molnar writes:

> * Michael Poole <mdpoole@troilus.org> wrote:
>
>> >> I do not suggest that copyright subsists in the signature or in the 
>> >> signing key.  Whether it does is irrelevant to the signing key 
>> >> being part of the source code (when the signature is needed for the 
>> >> binary to work properly).
>> >
>> > it is very much relevant. By admitting that the key is not part of 
>> > the "work", you have lost all moral basis to claim control over it.
>> 
>> I have not admitted any such thing.  I have said the key and signature 
>> do not have separate copyright protection.  Variables named "i" in a 
>> file are not protected by copyright, but they are very much part of 
>> the source code in that file.
>
> the problem with your argument is that the definition of what 
> constitutes "work" is up to copyright law, _not_ the license writer. 

Linux is unquestionably a work protected under copyright law.  When I
compile Linux, copyright law still protects the executable form.  This
is not a problem.

> I.e. you cannot just cleverly define "source code" to include something 
> unrelated and then pretend that it's all in one work. And that's exactly 
> what the GPLv3 does: it creatively defines the hardware's key into the 
> 'source code' of the software and then asks for that to be provided 
> _not_ because somehow the key derives from the software (it clearly does 
> not), but as a "compensation" for the right to redistribute! I.e. it's 
> trying to extend its scope to some item that is not part of the 
> software. See?

No.  The GPL does not care about the hardware's key, as I pointed out
in the part of my email that you cut out.  The GPL cares about the key
used to generate an integral part of the executable form of the GPLed
work.  The executable does not function properly if it lacks that
part.  This is exactly the same way in which the GPL cares about the
programming instructions in other parts of the source code: if you
remove them, the resulting work does something quite different.  See?

Michael Poole

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Ingo Molnar]

* Michael Poole <mdpoole@troilus.org> wrote:

> > I.e. you cannot just cleverly define "source code" to include 
> > something unrelated and then pretend that it's all in one work. And 
> > that's exactly what the GPLv3 does: it creatively defines the 
> > hardware's key into the 'source code' of the software and then asks 
> > for that to be provided _not_ because somehow the key derives from 
> > the software (it clearly does not), but as a "compensation" for the 
> > right to redistribute! I.e. it's trying to extend its scope to some 
> > item that is not part of the software. See?
> 
> No.  The GPL does not care about the hardware's key, as I pointed out 
> in the part of my email that you cut out.  The GPL cares about the key 
> used to generate an integral part of the executable form of the GPLed 
> work.  The executable does not function properly if it lacks that 
> part. [...]

it is a false statement on your part that the executable "does not 
function properly" if it lacks that part. Try it: take out the harddisk 
from the Tivo (it's a bog standard IDE harddisk), put into a nice Linux 
PC, mount it, modify a bit in the kernel image header and it will likely 
still boot just fine on that PC.

now if you put the harddisk back into the Tivo, the Tivo's bootloader 
will refuse to run that modified kernel. So will it (and any Linux 
bootloader) refuse to load the kernel if you corrupt the compressed 
format and the gunzip function finds a CRC error. You cannot run 
arbitrary binaries on hardware without knowing the properties of that 
hardware. One such property of the hardware might be: "i only run 
applications that use at most 500 MB of RAM" - because ... the hardware 
might only have 512 MB of RAM. Another property of the hardware might 
be: "i will only trust and run applications that match a given 
signature". Dont buy that hardware if you dont like its inherent 
limitations!

The modification the GPL talks about is about modification of the SOURCE 
CODE. But if you have a new binary, you have no expectation of being 
able to run that on a piece of hardware. It might or might not run. (for 
example if you modified the software to include a 1 GB static array then 
the software might not work on a system that has only 512 MB of RAM.)

go download the Tivo Linux kernel from:

  http://dynamic.tivo.com/linux/811/linux-2.4.tar.gz

modify and build it. Boot it on your general purpose PC. It will quite 
likely work just fine!

	Ingo

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Michael Poole]

Ingo Molnar writes:

> * Michael Poole <mdpoole@troilus.org> wrote:
>
>> > I.e. you cannot just cleverly define "source code" to include 
>> > something unrelated and then pretend that it's all in one work. And 
>> > that's exactly what the GPLv3 does: it creatively defines the 
>> > hardware's key into the 'source code' of the software and then asks 
>> > for that to be provided _not_ because somehow the key derives from 
>> > the software (it clearly does not), but as a "compensation" for the 
>> > right to redistribute! I.e. it's trying to extend its scope to some 
>> > item that is not part of the software. See?
>> 
>> No.  The GPL does not care about the hardware's key, as I pointed out 
>> in the part of my email that you cut out.  The GPL cares about the key 
>> used to generate an integral part of the executable form of the GPLed 
>> work.  The executable does not function properly if it lacks that 
>> part. [...]
>
> it is a false statement on your part that the executable "does not 
> function properly" if it lacks that part. Try it: take out the harddisk 
> from the Tivo (it's a bog standard IDE harddisk), put into a nice Linux 
> PC, mount it, modify a bit in the kernel image header and it will likely 
> still boot just fine on that PC.

Tivo did not program or sell the hard drive to be used in an arbitrary
Linux PC.  They sold the hard drive to be used in their hardware, with
a Linux kernel specifically modified for that.  Without the right
digital signature, it does not do the same thing: it is *incomplete*.
That is eminently a software issue.  Hardware limitations -- whether
they be RAM size or requirement for a certain digital signature -- are
beside the point.

The requirement that I "modify a bit in the kernel image header" is
also one of the most pathetic cop-outs I have seen.  What makes that
binary format the preferred form for modification of Linux?

Michael Poole

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 15, 2007, Ingo Molnar <mingo@elte.hu> wrote:

> it is a false statement on your part that the executable "does not 
> function properly" if it lacks that part. Try it: take out the harddisk 
> from the Tivo (it's a bog standard IDE harddisk), put into a nice Linux 
> PC, mount it, modify a bit in the kernel image header and it will likely 
> still boot just fine on that PC.

Ok, try this: take the disk out, remove/replace/modify the signature,
put the disk back in, and tell me what it is that fail to run.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Ingo Molnar]

* Alexandre Oliva <aoliva@redhat.com> wrote:

> On Jun 15, 2007, Ingo Molnar <mingo@elte.hu> wrote:
> 
> > it is a false statement on your part that the executable "does not 
> > function properly" if it lacks that part. Try it: take out the harddisk 
> > from the Tivo (it's a bog standard IDE harddisk), put into a nice Linux 
> > PC, mount it, modify a bit in the kernel image header and it will likely 
> > still boot just fine on that PC.
> 
> Ok, try this: take the disk out, remove/replace/modify the signature, 
> put the disk back in, and tell me what it is that fail to run.

you mean back into the Tivo? That is not support for what you claimed. 
You claimed the "executable does not function properly" if it lacks that 
part (and you did not qualify your statement with anything). That was a 
false statement, because it still works fine in just about any 
bog-standard PC. A true statement would be: "the modified executable 
does not function properly _in the Tivo_". It still works fine on a 
general purpose PC.

In fact, you couldnt even modify the binary on the Tivo, because the 
Tivo is not a general purpose PC, it is a PVR. You'd have to put the 
disk into a PC to modify the binary. And then you'd have to put it back 
into the Tivo. So even in this silly example of yours you _already_ have 
to have a general purpose programmable system where the free software 
runs fine, and even under your strained and invalid interpretation of 
the GPLv2, your "rights" to modify the software are very well present on 
that general purpose system.

But you didnt really want to make use of Tivo's free software 
enhancements, right? Lets face the sad truth: the overwhelming majority 
of Tivo 'modders' wanted to hack the PVR not to enhance the Tivo, they 
more likely wanted to watch pay-per-view content without the pay bit and 
they perhaps wanted to get around service restrictions that the Tivo 
implements (and through which it funds lower-than-production-cost for 
the PVR). So the 'rights' you are trying to protect are invented 
'rights' of mostly _freeloaders_ in fact. The 'Tivo community' was 
conjured up after the fact. So even in this supposedly golden and 
hand-picked DRM example of RMS, the whole story stinks from beginning to 
end and has all the classic earmarks of detached-from-the-real-world 
religious extremism in the works ...

and the whole effort is totally pointless anyway. Consumers are already 
voting with their feet against DRM restrictions. So the only DRM victims 
of the GPLv3 attack measures will be the _good_ uses of DRM. People will 
be able to use tamper-proof, vendor-upgradable hardware based on 
FreeBSD, but not based on any GPLv3 kernel. Nobody will care about 
content DRM at that point anymore, and all that remains is a license 
that is crippled for the _good_ uses. Smart move to advance free 
software, eh? Really, RMS should have kept hacking code a bit longer so 
that he doesnt become that totally detached from the rest of us.

	Ingo

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Michael Poole]

Ingo Molnar writes:

> * Alexandre Oliva <aoliva@redhat.com> wrote:
>
>> On Jun 15, 2007, Ingo Molnar <mingo@elte.hu> wrote:
>> 
>> > it is a false statement on your part that the executable "does not 
>> > function properly" if it lacks that part. Try it: take out the harddisk 
>> > from the Tivo (it's a bog standard IDE harddisk), put into a nice Linux 
>> > PC, mount it, modify a bit in the kernel image header and it will likely 
>> > still boot just fine on that PC.
>> 
>> Ok, try this: take the disk out, remove/replace/modify the signature, 
>> put the disk back in, and tell me what it is that fail to run.
>
> you mean back into the Tivo? That is not support for what you claimed. 
> You claimed the "executable does not function properly" if it lacks that 
> part (and you did not qualify your statement with anything). That was a 
> false statement, because it still works fine in just about any 
> bog-standard PC. A true statement would be: "the modified executable 
> does not function properly _in the Tivo_". It still works fine on a 
> general purpose PC.

I claimed that.  Unless I missed something, Alexandre did not.

Ability to run on a standard PC is irrelevant.  Tivo distributes the
executable for the specific purpose of running on their hardware.
Having the signature accepted by the hardware is a critical aspect of
the executable.  That purpose and function are what make the signature
part of the work based on Linux.

Courts consider purpose and intent when analyzing actions; except when
one has bought the best available legal system, they would not follow
your logic.  (The role the signature plays in controlling access to a
copyrighted work, per DMCA, might also separately identify it as part
of the work based on Linux.)

If I wished to distribute a kernel with extended functionality from a
C file but not the C source files, under your logic I need not give
them out -- a user could modify the binary and run it on a general
purpose PC.  Right?  At most it would take clever linker tricks to
make the change small enough.

As to the suggestion that vendors would use another kernel: I would
not mind.  A huge fraction of the interesting and useful work in open
source kernels happens in Linux (first or only).  Using any third
party software is a trade-off of what you get versus what you give up.

Michael Poole

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 17, 2007, Ingo Molnar <mingo@elte.hu> wrote:

> * Alexandre Oliva <aoliva@redhat.com> wrote:

>> On Jun 15, 2007, Ingo Molnar <mingo@elte.hu> wrote:
>> 
>> > it is a false statement on your part that the executable "does not 
>> > function properly" if it lacks that part. Try it: take out the harddisk 
>> > from the Tivo (it's a bog standard IDE harddisk), put into a nice Linux 
>> > PC, mount it, modify a bit in the kernel image header and it will likely 
>> > still boot just fine on that PC.
>> 
>> Ok, try this: take the disk out, remove/replace/modify the signature, 
>> put the disk back in, and tell me what it is that fail to run.

> you mean back into the Tivo? That is not support for what you claimed. 
> You claimed the "executable does not function properly" if it lacks that 
> part (and you did not qualify your statement with anything). That was a 
> false statement, because it still works fine in just about any 
> bog-standard PC. A true statement would be: "the modified executable 
> does not function properly _in the Tivo_". It still works fine on a 
> general purpose PC.

I stand """corrected""".  It doesn't matter, because the TiVo is where
the combination of the executable with the signature shipped, and,
see, I didn't talk about modifying the executable, what I wrote about
above was modifying the signature alone.  See?

> But you didnt really want to make use of Tivo's free software 
> enhancements, right? Lets face the sad truth: the overwhelming majority 
> of Tivo 'modders' wanted to hack the PVR not to enhance the Tivo, they 
> more likely wanted to watch pay-per-view content without the pay bit and 
> they perhaps wanted to get around service restrictions that the Tivo 
> implements (and through which it funds lower-than-production-cost for 
> the PVR).  So the 'rights' you are trying to protect are invented 
> 'rights' of mostly _freeloaders_ in fact.

Sony Betamax anyone?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Florin Malita]

On 06/15/2007 10:56 AM, Michael Poole wrote:
> The GPL cares about the key
> used to generate an integral part of the executable form of the GPLed
> work.  

GLPv2 doesn't: why do you think the digital signature is an integral 
part of the executable? It can be a totally separate blob, distributed 
via a separate channel and even stored at a different location than the 
executable. Does it still look like an integral part of the executable 
to you then?

(unless of course you're trying to argue that the hash itself is a 
derivative work, but that has already been refuted many times before.)

> The executable does not function properly if it lacks that
> part.

It works just fine given the right environment. The right environment 
may be some other hardware (without DRM restrictions) or the DRMed 
device + an authorized digital signature. The digital signature is not 
part of your executable.

Do you honestly believe GPLv2 requires the distributor to provide you 
with the right environment for your modified copy to "function 
properly"? I would say it doesn't, but feel free to point me to specific 
sections which *state* otherwise. AFAICT, GPLv2 is specifically limited 
to "copying, distribution and modification". How you use (or don't use, 
or can't use) your modified copy is totally outside its scope.

---
fm

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Michael Poole]

Florin Malita writes:

> On 06/15/2007 10:56 AM, Michael Poole wrote:
>> The GPL cares about the key
>> used to generate an integral part of the executable form of the GPLed
>> work.  
>
> GLPv2 doesn't: why do you think the digital signature is an integral
> part of the executable? It can be a totally separate blob, distributed
> via a separate channel and even stored at a different location than
> the executable. Does it still look like an integral part of the
> executable to you then?

Yes.  If I cut a book in half and store the halves separately, does
the second half become an independent work?  The integral-ness is a
function of how the thing is created and how it functions, not how it
is stored.  If you need part B for part A to execute as intended, then
part A is not a complete work in itself.  On top of this, in the Tivo
case the two are distributed together, and even part of the same file.

[snip]
> Do you honestly believe GPLv2 requires the distributor to provide you
> with the right environment for your modified copy to "function
> properly"? I would say it doesn't, but feel free to point me to
> specific sections which *state* otherwise. AFAICT, GPLv2 is
> specifically limited to "copying, distribution and modification". How
> you use (or don't use, or can't use) your modified copy is totally
> outside its scope.

The GPL does not require a distributor to provide me with any kind of
environment.  If I get a Tivo kernel image but do not have a Tivo, the
GPL does not require anyone to give me hardware.  Fortunately, that is
not at all my argument.

Michael Poole

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Florin Malita]

On 06/15/2007 12:18 PM, Michael Poole wrote:
> Florin Malita writes:
>
>   
>> On 06/15/2007 10:56 AM, Michael Poole wrote:
>>     
>>> The GPL cares about the key
>>> used to generate an integral part of the executable form of the GPLed
>>> work.  
>>>       
>> GLPv2 doesn't: why do you think the digital signature is an integral
>> part of the executable? It can be a totally separate blob, distributed
>> via a separate channel and even stored at a different location than
>> the executable. Does it still look like an integral part of the
>> executable to you then?
>>     
>
> Yes.  If I cut a book in half and store the halves separately, does
> the second half become an independent work?  

Except in this case you're not touching the book at all. If you write a 
review for a book (much better analogy methinks), then your review is 
obviously not an integral part of the book even though it's based on its 
content.

> The integral-ness is a
> function of how the thing is created and how it functions, not how it
> is stored.  If you need part B for part A to execute as intended, then
> part A is not a complete work in itself.

Being an integral part (as in combined or derived work) has nothing to 
do with usability. There are many other bits and pieces your executable 
needs in order to function properly (or at all) but that doesn't make 
your CPU microcode & electricity provider an integral part of the 
program, does it?

Luckily, it doesn't really matter what you or I think that 
"integral-ness" means, all it matters is how copyright law defines a 
"derivative work" and whether a cryptographic hash is such a thing. Now 
are you seriously arguing that a hash is a derivative work?

>   On top of this, in the Tivo
> case the two are distributed together, and even part of the same file.
>   

It's mere aggregation, but it's totally irrelevant because they could 
just as easily change their approach.

---
fm

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Michael Poole]

Florin Malita writes:

> On 06/15/2007 12:18 PM, Michael Poole wrote:
>> Florin Malita writes:
>>
>>   
>>> On 06/15/2007 10:56 AM, Michael Poole wrote:
>>>     
>>>> The GPL cares about the key
>>>> used to generate an integral part of the executable form of the GPLed
>>>> work.        
>>> GLPv2 doesn't: why do you think the digital signature is an integral
>>> part of the executable? It can be a totally separate blob, distributed
>>> via a separate channel and even stored at a different location than
>>> the executable. Does it still look like an integral part of the
>>> executable to you then?
>>>     
>>
>> Yes.  If I cut a book in half and store the halves separately, does
>> the second half become an independent work?  
>
> Except in this case you're not touching the book at all. If you write
> a review for a book (much better analogy methinks), then your review
> is obviously not an integral part of the book even though it's based
> on its content.

Extremely poor analogy.  I do not distribute my review with the book.
Someone buying the book is able to use the book just fine (for the
purpose for which it was sold) without my review.  They need neither
my review nor other modifications before the book becomes readable.
As Ingo said, you need either the digital signature or other changes
before a Tivo kernel image will load.

>> The integral-ness is a
>> function of how the thing is created and how it functions, not how it
>> is stored.  If you need part B for part A to execute as intended, then
>> part A is not a complete work in itself.
>
> Being an integral part (as in combined or derived work) has nothing to
> do with usability. There are many other bits and pieces your
> executable needs in order to function properly (or at all) but that
> doesn't make your CPU microcode & electricity provider an integral
> part of the program, does it?

No.  Those are independent works.  They are not distributed to make a
certain piece of software function in a particular way or place.  The
presence of software in a box with CPU microcode is -- at least in
every case I have seen -- what GPL calls "mere aggregation".

> Luckily, it doesn't really matter what you or I think that
> "integral-ness" means, all it matters is how copyright law defines a
> "derivative work" and whether a cryptographic hash is such a
> thing. Now are you seriously arguing that a hash is a derivative work?

No.  I explained this before.  Try reading the thread and the GPL.  I
am not sure where people get the (wrong) idea that the GPL only
concerns itself with "derivative work[s]".

>>   On top of this, in the Tivo
>> case the two are distributed together, and even part of the same file.
>>   
>
> It's mere aggregation, but it's totally irrelevant because they could
> just as easily change their approach.

If and when they do, I'll consider the rules that might apply.  Until
then, it is fairly stupid to try to defend Tivo by saying they *might*
do something they currently don't, and if they did, they *might* have
a defense that they currently don't.

Michael Poole

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Florin Malita]

On 06/15/2007 02:30 PM, Michael Poole wrote:
> Florin Malita writes:
>
>   
>> On 06/15/2007 12:18 PM, Michael Poole wrote:
>>     
>>> Yes.  If I cut a book in half and store the halves separately, does
>>> the second half become an independent work?  
>>>       
>> Except in this case you're not touching the book at all. If you write
>> a review for a book (much better analogy methinks), then your review
>> is obviously not an integral part of the book even though it's based
>> on its content.
>>     
>
> Extremely poor analogy.  I do not distribute my review with the book.
>   

But you do (because I say so ; ), and guess what? It makes no 
difference: your review is not a derivative work anymore than it was before.

> Someone buying the book is able to use the book just fine (for the
> purpose for which it was sold) without my review.  They need neither
> my review nor other modifications before the book becomes readable.
>   

Exactly. So what's your difficulty in downloading the Tivo code, reading 
it and re-using it in your own projects, on your other devices? How is 
the missing signing key preventing you from doing any of that?

Someone buying the book may be free to read it anywhere but if they 
insist on reading it at your table you may sensibly require they bring a 
copy of your review with them (to prove their genuine interest ; ). 
Failure to comply only means they have to read the book someplace else. 
Can they read the book? Sure. Can they read it at your table? Only if 
you choose to allow them.

> As Ingo said, you need either the digital signature or other changes
> before a Tivo kernel image will load.
>   

GPLv2 guarantees that the book remains readable. It does not grant you 
(doesn't even try) the right to execute a modified copy on any 
particular piece of hardware. Your kernel is perfectly functional on any 
platform that supports it - it just so happens that the Tivo device does 
not support it.

>> Being an integral part (as in combined or derived work) has nothing to
>> do with usability. There are many other bits and pieces your
>> executable needs in order to function properly (or at all) but that
>> doesn't make your CPU microcode & electricity provider an integral
>> part of the program, does it?
>>     
>
> No.  Those are independent works.

So is a digital signature. Again, are you arguing the digital signature 
is a derivative work?


>   
>> Luckily, it doesn't really matter what you or I think that
>> "integral-ness" means, all it matters is how copyright law defines a
>> "derivative work" and whether a cryptographic hash is such a
>> thing. Now are you seriously arguing that a hash is a derivative work?
>>     
>
> No.  I explained this before.  Try reading the thread and the GPL.  I
> am not sure where people get the (wrong) idea that the GPL only
> concerns itself with "derivative work[s]".
>   

I guess you'll have to explain again because copyright law and its 
definition of derivative works are the things that make the GPL work:

"0.This License applies to any program or other work which contains a 
notice placed by the copyright holder saying it may be distributed under 
the terms of this General Public License. The "Program", below, refers 
to any such program or work, and a "work based on the Program" means 
either the Program or any derivative work under copyright law".

What I can't find though is any reference to "integral parts" or your 
taken-for-granted right to run a modified copy of the program on the 
same device used for distribution (or any mention of functionality at 
all for that matter). Actually: "Activities other than copying, 
distribution and modification are not covered by this License; they are 
outside its scope".

>>>   On top of this, in the Tivo
>>> case the two are distributed together, and even part of the same file.
>>>   
>>>       
>> It's mere aggregation, but it's totally irrelevant because they could
>> just as easily change their approach.
>>     
>
> If and when they do, I'll consider the rules that might apply.  Until
> then, it is fairly stupid to try to defend Tivo by saying they *might*
> do something they currently don't, and if they did, they *might* have
> a defense that they currently don't.
>   

But you're missing the whole point: the rules are the same, nothing 
changes! You are drawing an artificial distinction between the two cases 
and focusing on aggregation, which is totally irrelevant: either the 
digital signature is a derivative work or it isn't, and in either case 
its distribution method makes no difference in the world.

The reason I brought up the separate-signature example is to illustrate 
just how ridiculous is to think of the signature keys as source files: 
you can implement an equivalent DRM system without ever modifying the 
kernel blob. Only difference is the channel used for signature 
distribution, and I hope you won't argue that mere aggregation changes 
its nature.

---
fm

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 15, 2007, Ingo Molnar <mingo@elte.hu> wrote:

> that's exactly what the GPLv3 does: it creatively defines the
> hardware's key into the 'source code' of the software and then asks
> for that to be provided _not_ because somehow the key derives from
> the software

Dude, you're 3 drafts behind.  And some people already read GPLv2 like
that.

> (it clearly does not), but as a "compensation" for the right to
> redistribute! I.e. it's trying to extend its scope to some item that
> is not part of the software. See?

The "compensation" is and has always been "respecting others'
freedoms".  If the key is used to disrespect others freedoms, as it is
by TiVO, then TiVO is failing to keep its part in the deal.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 15, 2007, Ingo Molnar <mingo@elte.hu> wrote:

> it is very much relevant. By admitting that the key is not part of the 
> "work", you have lost all moral basis to claim control over it.

legal basis, maybe.  legality and morality are quite different
concepts.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Michael Gerdau]

[-- Attachment #1: Type: text/plain, Size: 2969 bytes --]

> Because GPLv2 doesn't enforce limitations on the hardware a GPL'd work can be 
> put on. It doesn't make artificial distinctions between "Commercial", 
> "Industrial" and "User". What it does is *ATTEMPT* to ensure that nobody 
> receiving a copy of a GPL'd work has the same rights as any other person that 
> gets a copy. GPLv3 gives people *additional* rights beyond those.

IMO this statement expressedly exposes the different viewpoints as used
in various factions in this discussion.

Without adopting all the details I think I can agree to the above stmt.
However I don't agree with the implied msg as I perceive it.

In the following I'll try to explain what I mean by the above.

I don't know whether what TiVo did actually was allowed by the legal
phrases of the GPLv2. I can image it was legally valid but I don't know.
But then I'm convinced it was one of the things the inventors of the
GPL wanted to make illegal by it -- they may have failed to do so when
wording the legal part.

I like to remind you of the story with the broken closed source printer
driver RMS tried to fix at MIT (if I recall correctly) and the frustration
that he couldn't do so that finally made him start the FSF.

No customer can fix his TiVo box without the cooperation of the HW
vendor. If they refuse there is nothing that can be done. For me this
is very much like printer story above.

Assuming you (the reader) agree so far:
I find it obvious that the GPL was meant to prevent such to be possible.
This is what I mean by the "the spirit of the GPL".

Living in germany I'm also used to the courts valueing the intention over
the exact wording of a contract (a licence after all is a contract). So
I _think_ in germany TiVo would have lost a lawsuit if they had tried it.


Now for a different PoV:
Do I think Tivoisation is bad for the community ?
Of course I think it is but your mileage may vary.

Anyway, if one considers Tivoisation acceptable then there is no reason
to stop using GPLv2.

If one wishes to prevent it there are two related questions:
- does GPLv2 prevent it ?
- if GPLv2 does not prevent it then how can we change it to achieve that ?

To me it seems as if the FSF tends to answer the first question with 'no'
and consequently answers the second question with 'GPLv3'.

Whether or not the GPLv3 is truely an acceptable answer to prevent
Tivoisation is a completely different issue that I can't really judge.

Last not least:
Nothing of the above has to do with ethics, moral or any such cathegories.
This is by intention.

Thank you for reading thus far -- I hope I made myself clear.

Best wishes,
Michael
-- 
 Technosis GmbH, Geschäftsführer: Michael Gerdau, Tobias Dittmar
 Sitz Hamburg; HRB 89145 Amtsgericht Hamburg
 Vote against SPAM - see http://www.politik-digital.de/spam/
 Michael Gerdau       email: mgd@technosis.de
 GPG-keys available on request or at public keyserver

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Friday 15 June 2007 02:24:37 Michael Gerdau wrote:
> > Because GPLv2 doesn't enforce limitations on the hardware a GPL'd work
> > can be put on. It doesn't make artificial distinctions between
> > "Commercial", "Industrial" and "User". What it does is *ATTEMPT* to
> > ensure that nobody receiving a copy of a GPL'd work has the same rights
> > as any other person that gets a copy. GPLv3 gives people *additional*
> > rights beyond those.
>
> IMO this statement expressedly exposes the different viewpoints as used
> in various factions in this discussion.

Agreed. And thanks for ignoring the obvious typo.

> Without adopting all the details I think I can agree to the above stmt.
> However I don't agree with the implied msg as I perceive it.
>
> In the following I'll try to explain what I mean by the above.
>
> I don't know whether what TiVo did actually was allowed by the legal
> phrases of the GPLv2. I can image it was legally valid but I don't know.
> But then I'm convinced it was one of the things the inventors of the
> GPL wanted to make illegal by it -- they may have failed to do so when
> wording the legal part.

No doubt. However, GPLv2 actually states in clear and concise english that it 
doesn't cover *anything* but the rights to "copy, distribute and modify" 
covered works. It actually states that other rights are beyond the scope of 
the license. That statement, IMHO (and IANAL), obviates any other "intent" 
the "inventor" of the license may have had by making the scope of the license 
clear. 

<snip>

> No customer can fix his TiVo box without the cooperation of the HW
> vendor. If they refuse there is nothing that can be done. For me this
> is very much like printer story above.

I own an XBox 360. If it breaks I can't fix it without the cooperation of MS. 
The fact that a TiVO runs an OS that is licensed under the GPL doesn't change 
the fact that the situation is *exactly* the same. TiVO breaks? Manufacturer 
(or someone certified and licensed for the task by the manu) fixes it. XBox 
breaks? Manufacturer fixes it. My laptop breaks? As long as its under 
warranty, the manufacturer fixes it *FOR* *FREE* - if it's out of warranty, I 
pay for the "service" but they still fix it. QED: The "manufacturer must 
cooperate in or perform the repair" is not some new idea - its actually 
common practice.

> Assuming you (the reader) agree so far:
> I find it obvious that the GPL was meant to prevent such to be possible.
> This is what I mean by the "the spirit of the GPL".
>
> Living in germany I'm also used to the courts valueing the intention over
> the exact wording of a contract (a licence after all is a contract). So
> I _think_ in germany TiVo would have lost a lawsuit if they had tried it.

It might be that you are correct. However, thanks to someone having actually 
identified the exact scope of the GPLv2 *in* the legally active text of the 
license the "intent" shouldn't have much weight or bearing. Look at the first 
sentence of the second paragraph of section 0. "Activities other than 
copying, distribution and modification are not covered by this License; they 
are outside its scope."

>
> Now for a different PoV:
> Do I think Tivoisation is bad for the community ?
> Of course I think it is but your mileage may vary.

And I happen to agree with you. What I disagree with is taking steps to 
make "bad == illegal". I also have a problem with doing things that force my 
viewpoint on other people.

> Anyway, if one considers Tivoisation acceptable then there is no reason
> to stop using GPLv2.
>
> If one wishes to prevent it there are two related questions:
> - does GPLv2 prevent it ?
> - if GPLv2 does not prevent it then how can we change it to achieve that ?
>
> To me it seems as if the FSF tends to answer the first question with 'no'
> and consequently answers the second question with 'GPLv3'.
>
> Whether or not the GPLv3 is truely an acceptable answer to prevent
> Tivoisation is a completely different issue that I can't really judge.

And neither am I. My whole point in arguing over it has been that, despite 
what some people want to believe, it isn't violating the GPLv2 in any way, 
shape or form.

>
> Last not least:
> Nothing of the above has to do with ethics, moral or any such cathegories.
> This is by intention.

Thank you for that.

> Thank you for reading thus far -- I hope I made myself clear.

Yes, you have.

DRH

> Best wishes,
> Michael



-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Greaves]

Daniel Hazelton wrote:
>> Now for a different PoV:
>> Do I think Tivoisation is bad for the community ?
>> Of course I think it is but your mileage may vary.
> 
> And I happen to agree with you. What I disagree with is taking steps to 
> make "bad == illegal". I also have a problem with doing things that force my 
> viewpoint on other people.

Surely it's more:
   bad == go away and don't use future improvements to our software anymore please.
??

*If* you think it's bad (Linus doesn't as far as the kernel goes) then isn't it 
reasonable to exclude 'bad for the community' from the community?

This isn't retroactive - they can continue to use any V2 software they had, they 
wouldn't be able to use V3 developments.

That seems to me to be a very, very reasonable thing to do (and very much *not* 
bad == illegal IMHO)

David

PS well, I was just seeing if anyone had fixed my libata/md bug yet but this 
seemed more interesting.
PPS and Tejun has, I'm off...

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Friday 15 June 2007 06:18:59 David Greaves wrote:
> Daniel Hazelton wrote:
> >> Now for a different PoV:
> >> Do I think Tivoisation is bad for the community ?
> >> Of course I think it is but your mileage may vary.
> >
> > And I happen to agree with you. What I disagree with is taking steps to
> > make "bad == illegal". I also have a problem with doing things that force
> > my viewpoint on other people.
>
> Surely it's more:
>    bad == go away and don't use future improvements to our software anymore
> please. ??

I agree. I stated it in the terms I did because Alexandre originally 
brought "ethics and morality" into the discussion. Not that my word choice is 
wrong at all - violating a copyright license is, after all, illegal.

> *If* you think it's bad (Linus doesn't as far as the kernel goes) then
> isn't it reasonable to exclude 'bad for the community' from the community?

I agree that it is "bad for the community". The impression I've gotten from 
reading the GPLv3, reading transcripts of interviews with RMS, reading 
transcripts of interview with Eben Moglen *and* from the FSFLA members 
participating in this discussion the reason for the anti-tivoization language 
in GPLv3 isn't "its bad for the community" but "we find it ethically and 
morally wrong". That being the reason, what they are saying with the language 
in the GPLv3 is "this is bad/evil and things that are bad/evil should be 
illegal".

> This isn't retroactive - they can continue to use any V2 software they had,
> they wouldn't be able to use V3 developments.
>
> That seems to me to be a very, very reasonable thing to do (and very much
> *not* bad == illegal IMHO)

:) Same here

> David
>
> PS well, I was just seeing if anyone had fixed my libata/md bug yet but
> this seemed more interesting.

It is. Consumed a solid six-hour stretch of my day (5PM EDT to 11PM EDT, June 
14, 2007) (FYI: EDT is UTC - 4)

> PPS and Tejun has, I'm off...

Good for you!

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Fri, 15 Jun 2007, David Greaves wrote:
> 
> Surely it's more:
>   bad == go away and don't use future improvements to our software anymore
> please.
> ??

Well, with the understanding that I don't think that what Tivo did was bad 
in the first place, let me tackle that question *anyway*.

The answer is: Not necessarily.

Some people can be "bad" for the community. They may simply be disruptive 
and not productive at all. They may troll the mailing lists without 
actually ever doing something good, or they may do other "bad" things.

In fact, let's make it *very* specific: let's say that the bad person is a 
cracker, and specializes in finding security holes, and writing exploits 
for them, and selling those exploits to spammers.

Most of us might agree that that is a "bad" person for the community, no?

Now, by your own logic, let's look at what that means for the license. 
Should we write into our copyright license that you cannot try to find 
security holes? Would that be a good addition to the GPLv2?

Now, I stated that in a way where the answer is obvious: that would be a 
*horrible* addition to the GPLv2. I think everybody can agree on that. It  
would be really stupid to say "you cannot look for security holes" just 
because *some* people who do it are bad.

Now, think about that for a moment, and then go back to your question 
about whether Tivo is bad for the community, and whether being bad for the 
community should mean that the license should be written to say "go away 
and don't use future improvements to our software".

See where I'm trying to take you?

I think that even people who *do* think that what Tivo did was "bad", 
should think very deeply about the issue whether you should try to lock 
out "bad uses" in your license. Yes, the answer may be "yes, you should". 
But I'm arguing that the answer _may_ also be: "No, you shouldn't, becasue 
it turns out that you might lock out _good_ people too".

So in my cracker/spammer example, by trying to lock out the bad people, 
the obvious (and _stupid_ - don't get me wrong, I'm not at *all* 
suggesting anything like that should ever be done) license addition of 
"don't expose security problems" actually just causes more problems than 
it solves (if it solves anything at all - really bad people don't actually 
tend to even care about the license!).

It makes it harder for *valid* uses of security problem discovery. It 
makes it potentially illegal to try to do security research. And don't 
tell me stupid licenses and laws like that don't happen: people really 
*do* make these kinds of shortsighted decisions, to "protect" themselves 
from bad people.

I personally think that the same is true of the GPLv3 anti-tivoization 
clauses. Even if you don't like Tivo, you may well recognize that there 
are lots of *other* reasons for lock-down. Maybe you hate Tivo, and the 
RIAA and the MPAA. Fine. What about the FCC? Or what about secure 
terminals? What about any number of *other* reasons to use validated 
kernel images?

Now, I cannot speak for Alan Cox (he can speak damn well for himself), but 
I think Alan is an example of a person who actually really detests what 
Tivo did. But I also am pretty sure that he's also quite smart enough to 
see that the GPLv3 anti-tivoization clauses may stop *other* uses that he 
doesn't dislike and he doesn't think are "wrong", and even though he 
dislikes what Tivo does, as far as I know, I think his stance on the GPLv3 
is that it's actually wrong for Linux.

See? You don't actually have to like Tivo to see downsides to trying to 
stop them. Because these kinds of things have consequences *outside* of 
just stopping Tivo.

			Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[david]

On Fri, 15 Jun 2007, Linus Torvalds wrote:

>
> Now, by your own logic, let's look at what that means for the license.
> Should we write into our copyright license that you cannot try to find
> security holes? Would that be a good addition to the GPLv2?
>
> Now, I stated that in a way where the answer is obvious: that would be a
> *horrible* addition to the GPLv2. I think everybody can agree on that. It
> would be really stupid to say "you cannot look for security holes" just
> because *some* people who do it are bad.
>
> Now, think about that for a moment, and then go back to your question
> about whether Tivo is bad for the community, and whether being bad for the
> community should mean that the license should be written to say "go away
> and don't use future improvements to our software".
>
> See where I'm trying to take you?
>
> I think that even people who *do* think that what Tivo did was "bad",
> should think very deeply about the issue whether you should try to lock
> out "bad uses" in your license. Yes, the answer may be "yes, you should".
> But I'm arguing that the answer _may_ also be: "No, you shouldn't, becasue
> it turns out that you might lock out _good_ people too".
>
> So in my cracker/spammer example, by trying to lock out the bad people,
> the obvious (and _stupid_ - don't get me wrong, I'm not at *all*
> suggesting anything like that should ever be done) license addition of
> "don't expose security problems" actually just causes more problems than
> it solves (if it solves anything at all - really bad people don't actually
> tend to even care about the license!).
>
> It makes it harder for *valid* uses of security problem discovery. It
> makes it potentially illegal to try to do security research. And don't
> tell me stupid licenses and laws like that don't happen: people really
> *do* make these kinds of shortsighted decisions, to "protect" themselves
> from bad people.

in fact there was news in the last week or two about a law in Germany that 
does exactly this. it outlaws all programs that can be used for hacking 
systems.

David Lang

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Greaves]

Linus Torvalds wrote:
> 
> On Fri, 15 Jun 2007, David Greaves wrote:
>> Surely it's more:
>>   bad == go away and don't use future improvements to our software anymore
>> please.
>> ??
> 
> Well, with the understanding that I don't think that what Tivo did was bad 
> in the first place, let me tackle that question *anyway*.
> 
> The answer is: Not necessarily.
I do agree with what you say here. Maybe a summary:
Babies, bathwater...
When you have a hammer (license) everything looks like a nail...

> See? You don't actually have to like Tivo to see downsides to trying to 
> stop them. Because these kinds of things have consequences *outside* of 
> just stopping Tivo.

My concern is around embedded type systems and maybe even the 'trusted' 
frameworks etc.

I _think_ I can see a completely opensource system that the end user cannot 
modify _in any way_. Which kinda defeats the point (to me) of opensource.

This 5 minute design undoubtedly has flaws but it shows a direction:
A basically standard 'De11' PC with some flash.
A Tivoised boot system so only signed kernels boot.
A modified kernel that only runs (FOSS) executables whose signed hash lives in 
the flash.

Do we (you) _want_ to prevent this?

Do we trust in 'the market' to prevent this?

Do we use license tools?

David

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Krzysztof Halasa]

David Greaves <david@dgreaves.com> writes:

> This 5 minute design undoubtedly has flaws but it shows a direction:
> A basically standard 'De11' PC with some flash.
> A Tivoised boot system so only signed kernels boot.
> A modified kernel that only runs (FOSS) executables whose signed hash
> lives in the flash.

How hard would it be to reprogramm the flash?
-- 
Krzysztof Halasa

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Greaves]

Krzysztof Halasa wrote:
> David Greaves <david@dgreaves.com> writes:
> 
>> This 5 minute design undoubtedly has flaws but it shows a direction:
>> A basically standard 'De11' PC with some flash.
>> A Tivoised boot system so only signed kernels boot.
>> A modified kernel that only runs (FOSS) executables whose signed hash
>> lives in the flash.
> 
> How hard would it be to reprogramm the flash?

The flash contains hashes signed by the companies private key.

The kernel contains the public key. It can decrypt the hashes but the private 
key isn't available to encrypt them. So although you can put a new application 
onto the system, you can't create a signed hash to write to the flash.

The kernel only runs the executable if the hash is valid.
You can re-write the kernel to avoid this check - but the hardware is Tivoised - 
so you can't run it.

I am not suggesting the kernel should go down the GPLV2 route - I am wondering 
if this is a viable scenario or one of Schneiers'  "movie-plot" threats :)

David

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Krzysztof Halasa]

David Greaves <david@dgreaves.com> writes:

>> How hard would it be to reprogramm the flash?
>
> The flash contains hashes signed by the companies private key.
>
> The kernel contains the public key. It can decrypt the hashes but the
> private key isn't available to encrypt them. So although you can put a
> new application onto the system, you can't create a signed hash to
> write to the flash.
>
> The kernel only runs the executable if the hash is valid.
> You can re-write the kernel to avoid this check - but the hardware is
> Tivoised -
> so you can't run it.

I read it: the flash contains everything from the bootloader to the
kernel and file system.
The bootloader contains the public key and checks if the kernel/fs
are ok. That includes calculating hashes and checking signatures.
No encryption/decryption there at all.

Right?

Then how hard would it be to reprogram the flash, to get rid of all
this crap? Or to just put your public key there.

Do they at least use BGA type of flash chips so you can't attach
a clip and have to use something more demanding?
-- 
Krzysztof Halasa

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Greaves]

Krzysztof Halasa wrote:
> David Greaves <david@dgreaves.com> writes:
> 
>>> How hard would it be to reprogramm the flash?
>> The flash contains hashes signed by the companies private key.
>>
>> The kernel contains the public key. It can decrypt the hashes but the
>> private key isn't available to encrypt them. So although you can put a
>> new application onto the system, you can't create a signed hash to
>> write to the flash.
 >
> Then how hard would it be to reprogram the flash, to get rid of all
> this crap? Or to just put your public key there.
> 
> Do they at least use BGA type of flash chips so you can't attach
> a clip and have to use something more demanding?

Stop trying to technically crack my 5-minute fag-packet design - that's easy and 
boring :)

Tivo have solved this problem - use their solution - but do it on something more 
general purpose.

Help fix it - the point is more "is this feasible". And if it is, "does it matter?"


David

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Fri, 15 Jun 2007, Michael Gerdau wrote:
> 
> I find it obvious that the GPL was meant to prevent such to be possible.
> This is what I mean by the "the spirit of the GPL".

Umm. It may well have been meant by *rms*. But your argument fatally falls 
down on the fact that rms has had *nothing* to do with the Linux kernel.

> Living in germany I'm also used to the courts valueing the intention over
> the exact wording of a contract (a licence after all is a contract). So
> I _think_ in germany TiVo would have lost a lawsuit if they had tried it.

Ehh. The intent that matters is not the intent of the person who authored 
the license, but the intent of the person who *chose* the license.

In other words, rms has *no* input on the kernel.

What matters is *my* intent in *choosing* the GPLv2, not *his* intent in 
writing it. 

But to make it even less relevant: intent really only legally matters when 
the legal issues are unclear.

And they really aren't that unclear here.

> Anyway, if one considers Tivoisation acceptable then there is no reason
> to stop using GPLv2.

Indeed.

> If one wishes to prevent it there are two related questions:
> - does GPLv2 prevent it ?
> - if GPLv2 does not prevent it then how can we change it to achieve that ?

Well, I think it's fairly unquestionable that the GPLv3 does prevent it. 

So your second question isn't even really interesting. We know the answer. 
So the only question that is even remotely interesting is the first one.

> To me it seems as if the FSF tends to answer the first question with 'no'
> and consequently answers the second question with 'GPLv3'.

Yes, I do agree with that reasoning, but there are *other*, and more 
direct, reasons than just the FSF's answer to say that the answer to your 
first question is "no".

The fact is, plain reading of the license (which *always* takes precedence 
over "intent", even in Europe) simply doesn't make what Tivo did illegal.

You literally have to read the GPLv2 in ways that are obviously not true 
to get to any other situation.

For example, Alexandre made the same two mistakes over and over in his 
reading when he tried to argue that the GPLv2 disallows what Tivo did:

 (a) The right to modify means "modify in place"

     This was a point that Alexandre (and others) have tried to make, but 
     it really is *not* supported by any reality.

     First off, the GPLv2 simply never *ever* says "in place". That
     wording (or anything equivalent) simply does not exist! So you really 
     have to add it by "reading" it some special way, and quite frankly, 
     no such reading is sensible.

     I can logically *prove* that such a reading is not sensible by the 
     two examples I already made clear to Alexandre multiple times:

	- Red Hat sends out DVD's with GPL'd software, and thus 
	  distributes copies that CANNOT be "modified in place". So 
	  thinking that "modified in place" is made illegal by the GPLv2 
	  is simply untenable, unless you think that what Red Hat does is 
	  against the GPLv2.

	  Do you really believe that what Red Hat does is against the GPL? 
	  Does Alexandre? Judging by his email address, I don't think he 
	  does.

	- I myself distribute Linux by making it available for download 
	  publicly, and no, I do *not* allow people to modify it in place, 
	  even though I do distribute Linux.

	  Alexandre tried to argue against this point by saying that I'm 
	  not really "distributing" Linux, and that the actual copying is 
	  being done by people who download it, and that they *can* modify 
	  their downloads "in place", but that only shows a lack of 
	  knowledge about what "distribution" means in copyright law.

	  The US courts have made it very clear that you "distribute" 
	  things not by downloading them, but by making them *available* 
	  for public download. So from a purely *legal* standpoint, I do 
	  actually *distribute* Linux by making it available! And no, I 
	  don't allow people to modify it in place!

	  Again, do you *really* think that I am in violation of the GPL 
	  by making the kernel available for download (== distributing 
	  it), but not letting others modify it in place?

So clearly, the whole "modify in place" argument is simply *wrong*.

It cannot *possibly* be a valid reading of the GPLv2! When the GPLv2 talks 
about "legal permissions to copy, distribute and/or modify" the software, 
it does *not* mean that you have to have the ability to modify it in 
place!

The other argument Alexandre made was different, but equally invalid from 
a legal standpoint:

 (b) The language in the preamble: "must give the recipients all the 
     rights that you have" means really *all* the rights and abilities!

     This too is a totally flawed reading of the license, for a few 
     reasons:

	- It's not actually in the *conditions* section. It's part of the 
	  explanation of the conditions, and while it's part of the 
	  license, it actually *is* different in the sense that it's not 
	  even meant to specify the "precise terms and conditions". Those 
	  are explicitly laid out _elsewhere_ (look for that term: 
	  "precise terms and conditions").

	- However, I don't dispute that the language exists, I just 
	  dispute Alexandres _reading_ of it, and I offer an alternate 
	  meaning for it. I do not guarantee that *my* reading is the 
	  correct one (although practically speaking, I pretty much can 
	  guarantee that too), but I *do* guarantee that Alexandre's 
	  reading is wrong, because it is nonsensical and impossible.

	  It is nonsensical and impossible in the exact same way I
	  showed that he was trying to claim something nonsensical 
	  and impossible in adding the "in place" reading to "right to 
	  modify".

	- So the reason I want to mention the fact that it is in the 
	  "preamble" is exactly because the preamble is *not* meant to be 
	  the "exact terms and conditions", it's meant to be a much softer 
	  thing - something that _explains_ the exact terms and 
	  conditions.

	  In particular, in this case, it explains "Section 6" (and 
	  you could also say that it explains section 7 too, but the 
	  preamble actually talks about patents separately, so you should 
	  probably see "section 6" as just the most obvious example of 
	  what the preamble is talking about).

	- In other words, the preamble is just a preliminary and inexact 
	  explanation for the _real_ requirements, which are spelled out 
	  elsewhere, in places like section 6 of the license. And *that* 
	  section actually makes it clear that "all rights" only concerns 
	  the rights SPELLED OUT IN THE LICENSE ITSELF.

	  Section 6 explicitly says:

		"You may not impose any further restrictions on the 
		 recipients' exercise of the rights granted herein."

	  and that is what I claim is the *correct* and exact reading of 
	  the explanatory language in the preamble. Yes, it is "all 
	  rights", but it is "all rights" ONLY AS FAR AS THE GPLv2 itself 
	  is concerned! It's not about any _other_ additional rights you 
	  may have outside the GPLv2!

	- To finish that point, let me just point out that Alexandre's 
	  reading *cannot* possibly work!

	  In particular, reading the preamble to mean that you give the 
	  recipient *all* rights you have (even outside the actual 
	  requirements of the license itself) that Alexandre tried to 
	  argue, is actually legally simply impossible. The copyright 
	  owner *fundamentally* has more rights than any licensee, and no 
	  license can *ever* transfer all those legal rights!

	  In other words, Alexandre's reading of the preamble is simply 
	  not legally tenable, because it assumes that the license says 
	  something that simply is not legally possible. And obviously the 
	  license doesn't say anything like that at all: it's a legal 
	  document, it can only state legally valid things!

	  I'll use the *exact* same examples as previously to show how 
	  Alexandre's reading of "all the rights that you have" can not 
	  possibly be correct:

	  (a) Again, Red Hat makes DVD's that contain GPLv2'd programs on 
	      them. Red Hat is bound by the GPL, so each work they put 
	      on the DVD is always under the GPL, and Red Hat *must* give 
	      you the rights that the GPL specifies.

	      But does Red Hat actually give you *all* the rights they 
	      hold on the DVD? No, they definitely do not. They hold a 
	      compilation copyright on RHEL, and they very  much do *not* 
	      give you the right to copy the whole distribution and sell 
	      it as RHEL. You only get the rights to the individual 
	      pieces, not to the whole thing!

	  (b) Again, I make the Linux kernel available to you on a web 
	      site, and thus distribute it to you. Do I actually give you 
	      *all* the rights I have in it? Hell no. I cannot (and do 
	      not even want to). As an author, I have special rights in my 
	      code that you do not get. You get the rights spelled out in 
	      the GPLv2, and *nothing* more.

     In other words, Alexandre's reading of the text in the preamble is 
     *impossible*. It absolutely *cannot* be the way the license works. 
     It's not how Red Hat itself reads it, and it's not how it can even 
     legally be made to work even if somebody *wanted* to read it that 
     way.

See? Both of Alexandre's arguments about why Tivo did something "against 
the license" were actually totally bogus. Neither of them was relevant, 
they both hinged on Alexandre reading the GPLv2 *wrong*.

> Whether or not the GPLv3 is truely an acceptable answer to prevent
> Tivoisation is a completely different issue that I can't really judge.

Absolutely. I do think it prevents Tivoisation, but I personally think 
it's unacceptable in even *trying* to prevent it, and as I've tried to 
make clear, the GPLv2 definitely did *not* prevent Tivo from doing what 
they did.

			Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 15, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:

> On Fri, 15 Jun 2007, Michael Gerdau wrote:
>> 
>> I find it obvious that the GPL was meant to prevent such to be possible.
>> This is what I mean by the "the spirit of the GPL".

> Umm. It may well have been meant by *rms*. But your argument fatally falls 
> down on the fact that rms has had *nothing* to do with the Linux kernel.

You're mixing two separate issues:

1- does GPLv3 change the spirit of the GPL?

2- is GPLv3 better than GPLv2 for Linux?

The answers may be different, and the reason I got into this debate
was to set the record straight on 1.  As the discussion evolved (if
developing into a flamewar can be characterized as evolving ;-), I
realized the motivations for preferring v2 over v3 were not clear to
me (and they still appear contradictory to me), so I started
investigating that, which is indeed 2., but is not about 3:

3- is Linux going to switch to v3?

>> Living in germany I'm also used to the courts valueing the
>> intention over the exact wording of a contract (a licence after all
>> is a contract). So I _think_ in germany TiVo would have lost a
>> lawsuit if they had tried it.

> Ehh. The intent that matters is not the intent of the person who
> authored the license, but the intent of the person who *chose* the
> license.

+1

> So clearly, the whole "modify in place" argument is simply *wrong*.

When you leave an essential portion of the reasoning out, which you
repeatedly did, this conclusion is obvious.  But it's also obviously
wrong to try to apply this conclusion to the argument that I phrased.

> It cannot *possibly* be a valid reading of the GPLv2! When the GPLv2
> talks about "legal permissions to copy, distribute and/or modify"
> the software, it does *not* mean that you have to have the ability
> to modify it in place!

Isn't a restriction on in-place modification a further restriction on
the permission to modify granted by the license?  A further
restriction that is not permitted by the license?

Again, this is not about ROM, CD-ROMs and other unmodifiable media.
In this case, the distributor is not imposing this restriction, it's
not selecting the media with the strict purpose of forbidding
modification.  It doesn't retain the ability to modify without failing
to pass it on.  This is the key distiction that you repeatedly
dropped.  And then, presented it as if it were a separate argument.

> 	  rights", but it is "all rights" ONLY AS FAR AS THE GPLv2 itself 
> 	  is concerned! It's not about any _other_ additional rights you 
> 	  may have outside the GPLv2!

I agree, and I don't think I've ever claimed otherwise.  It's rights
as far as the software is concerned, and even this might be pushing it
a bit too far.  That's why the spirit gives the intuition, but the
legal terms are precise in turning that into "no further
restrictions", as I'd already explained long before you did.

But then, again, the license grants the right to modify, and prohibits
further restrictions to it, so I claim that saying "you can modify,
just not in place, because I won't let you do it" (rather than because
it's impossible), that's a further restriction of a freedom granted by
the license, which turns into a license violation.


Now we can turn into the debate on whether replacing is modifying, and
the conclusion is quite possibly that, in legal terms, it isn't.  I
don't care.  I'm not here to debate the legal terms.  I'm in this
debate to set the record straight on whether GPLv3 changes the spirit
of the GPL.

> See? Both of Alexandre's arguments about why Tivo did something "against 
> the license" were actually totally bogus.

Actually...  What you name as two separate arguments were two parts of
*one* of the 3 arguments I've raised so far.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Michael Gerdau]

[-- Attachment #1: Type: text/plain, Size: 8443 bytes --]

> > I find it obvious that the GPL was meant to prevent such to be possible.
> > This is what I mean by the "the spirit of the GPL".
> 
> Umm. It may well have been meant by *rms*. But your argument fatally falls 
> down on the fact that rms has had *nothing* to do with the Linux kernel.

While I raised this argument on the Lunix kernel ML it was not meant
to be valid specifically for it.

My observation in this thread is that almost everybody discusses different
aspects of the same thing and everybody is somehow right. I was trying
to "go back to start" and have the look at the overall picture which in
this case for me is the question what the GPL's spirit is.

Whether and which of it _you_ intended to adopt for the kernel I had
no intention to sencond guess.

> > Living in germany I'm also used to the courts valueing the intention over
> > the exact wording of a contract (a licence after all is a contract). So
> > I _think_ in germany TiVo would have lost a lawsuit if they had tried it.
> 
> Ehh. The intent that matters is not the intent of the person who authored 
> the license, but the intent of the person who *chose* the license.

That seems to imply that we have to deal with myriads of intended meanings,
namely those of all who contributed to the kernel.

I'm pretty sure I don't wish to walk that road. If you want to we'll have
to agree to disagree.

> What matters is *my* intent in *choosing* the GPLv2, not *his* intent in 
> writing it. 

I beg to differ. By adopting _his_ license you adopted his view. If you
don't like that then choose a different license (which obviously you are
free to do).

It's just not feaseable to have something like "my GPL means a different
thing than your GPL".

> But to make it even less relevant: intent really only legally matters when 
> the legal issues are unclear.
> 
> And they really aren't that unclear here.

We have to agree to disagree then.

> > If one wishes to prevent it there are two related questions:
> > - does GPLv2 prevent it ?
> > - if GPLv2 does not prevent it then how can we change it to achieve that ?
> 
> Well, I think it's fairly unquestionable that the GPLv3 does prevent it. 
> 
> So your second question isn't even really interesting. We know the answer. 
> So the only question that is even remotely interesting is the first one.

My second question leaves out whether or not GPLv3 is an acceptable answer
to my second question. While the FSF says it is it is by no means clear
that I will agree -- all I wanted to do is present the situation as I see it.

> Yes, I do agree with that reasoning, but there are *other*, and more 
> direct, reasons than just the FSF's answer to say that the answer to your 
> first question is "no".

Note I only said the FSF seems to say "no". I'm not yet sure I do.

> The fact is, plain reading of the license (which *always* takes precedence 
> over "intent", even in Europe) simply doesn't make what Tivo did illegal.

I disagree and I don't see that plain reading of the license is that
obvious w/r to the SHA1 key because from a certain perspective said key
is required to create a working modification which I'm entitled to under
the GPLv2. I also agree that your perspective has merrit too. I'm simply
not sure which of the above is "correct" (as in agreeable from a judge's
PoV).

Based on that I disagree with your above stmt, at least I don't think
your implication every other reading being outright wrong is false.

This thread IMO clearly shows that apparently it isn't that clear -- far
too many intelligent people do disagree on this IMO not so obvious reading.

> You literally have to read the GPLv2 in ways that are obviously not true 
> to get to any other situation.

I object against the word obvious as an obsolute measure. You have all right
to consider it obvious from your PoV. My PoV may differ and I strongly claim
it being equally valid.

> For example, Alexandre made the same two mistakes over and over in his 
> reading when he tried to argue that the GPLv2 disallows what Tivo did:

I do not agree with everything Alexandre wrote but I do agree with some
parts.

>  (a) The right to modify means "modify in place"
[snip]
> So clearly, the whole "modify in place" argument is simply *wrong*.

I tend to agree and I didn't like it when it was brought up. But IMO that
does not invalidate his position as such.

>  (b) The language in the preamble: "must give the recipients all the 
>      rights that you have" means really *all* the rights and abilities!

I always did imply a "within reason". To me that means "if it is simple
for them to do it and can be simply extended to me as well then they have
to extend it". Handing out a SHA1 key definitely is simple and thus IMO
something I can expect them to do.

I'm aware we have to disagree on this. I'm also aware that this is
subject to the (IMO implied right) to run the modified code on the
original HW, mostly because they can do that very easily (which would
be covered by "within reason").

BTW:
I can easily take the position that from a metaphysical PoV the whole
concept of a copy is an illusion because every copy is in fact a modified
new version (that happens to be rather similar from a certain PoV... but
I disgress and don't really wish to walk that bridge :)

> 	  (a) Again, Red Hat makes DVD's that contain GPLv2'd programs on 
> 	      them. Red Hat is bound by the GPL, so each work they put 
> 	      on the DVD is always under the GPL, and Red Hat *must* give 
> 	      you the rights that the GPL specifies.
[snipped the paragraph on RH "withholding rights]

Here the "within reason" kicks in. And I readily admit that you are an
able intellectual acrobat who can twist arguments such that they become
silly.

My experience with german courts has shown me that the judges I had to
deal with always and foremost did apply a reality check and did not try
to bisect the consequences like an algorithm evaluated by a machine, i.e.
the tried to decide what is right and wrong and not whether the letter
of the contract could be twisted this or that way.

> 	  (b) Again, I make the Linux kernel available to you on a web 
> 	      site, and thus distribute it to you. Do I actually give you 
> 	      *all* the rights I have in it? Hell no. I cannot (and do 
> 	      not even want to). As an author, I have special rights in my 
> 	      code that you do not get. You get the rights spelled out in 
> 	      the GPLv2, and *nothing* more.

I probably haven't read all Alexandre wrote but from what I read I'd be
surprised he'd disagree (I don't).

>      In other words, Alexandre's reading of the text in the preamble is 
>      *impossible*. It absolutely *cannot* be the way the license works. 
>      It's not how Red Hat itself reads it, and it's not how it can even 
>      legally be made to work even if somebody *wanted* to read it that 
>      way.

I don't know whether Alexandre does read to preamble to mean what you
imply he does. But whether he does or not is irrelevant for me to decide
that IMO there is a strong argument to claim that what TiVo did is illegal
even under the GPLv2. Note I'm merely saying "strong argument" as opposed
to "clear beyond doubt".

> > Whether or not the GPLv3 is truely an acceptable answer to prevent
> > Tivoisation is a completely different issue that I can't really judge.
> 
> Absolutely. I do think it prevents Tivoisation, but I personally think 
> it's unacceptable in even *trying* to prevent it, and as I've tried to 
> make clear, the GPLv2 definitely did *not* prevent Tivo from doing what 
> they did.

Well, I think trying to prevent it is totally acceptable, after all I'm
free to impose whatever restrictions to my code I see fit (I think it was
long ago agreed that the GPLv2 does impose restrictions as well; they are
just different).

We have to agree to disagree on both whether trying to prevent Tivoisation
is acceptable and whether GPLv2 already prevents it.

Best wishes,
Michael
-- 
 Technosis GmbH, Geschäftsführer: Michael Gerdau, Tobias Dittmar
 Sitz Hamburg; HRB 89145 Amtsgericht Hamburg
 Vote against SPAM - see http://www.politik-digital.de/spam/
 Michael Gerdau       email: mgd@technosis.de
 GPG-keys available on request or at public keyserver

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Ingo Molnar]

* Michael Gerdau <mgd@technosis.de> wrote:

> > What matters is *my* intent in *choosing* the GPLv2, not *his* 
> > intent in writing it.
> 
> I beg to differ. By adopting _his_ license you adopted his view. [...]

ianal, but fortunately that's not what the law is. The license says what 
it says, and that is what controls. The intent of the author (of Linus 
and other copyright holders) is a secondary source of information /if 
and only if/ any ambiguity of meaning arises (as determined by a judge, 
not by you or me). But the opinion and intent of RMS (unless adopted by 
Linus) is quite immaterial.

( there is a legalistic special-case. If any dispute arises over
  what license the COPYING file in Linux itself (and only that file) is 
  under, then the intent of RMS matters too, but only for that limited
  matter for that single file. Btw., the COPYING file itself is not 
  licensed under the GPL. )

	Ingo

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Michael Gerdau]

[-- Attachment #1: Type: text/plain, Size: 1239 bytes --]

> > > What matters is *my* intent in *choosing* the GPLv2, not *his* 
> > > intent in writing it.
> > 
> > I beg to differ. By adopting _his_ license you adopted his view. [...]
> 
> ianal, but fortunately that's not what the law is. The license says what 
> it says, and that is what controls. The intent of the author (of Linus 
> and other copyright holders) is a secondary source of information /if 
> and only if/ any ambiguity of meaning arises (as determined by a judge, 
> not by you or me). But the opinion and intent of RMS (unless adopted by 
> Linus) is quite immaterial.

I agree with the "/if and only if/ any ambiguity of meaning arises" part.
I'm sorry I didn't make that clear before.

However if that situation arises (i.e. the judge decides there is an
ambiguity) then as far as my experience tells me it is the intention of
the author (RMS et al in this case) that counts. But I erred before...

Best wishes,
Michael
-- 
 Technosis GmbH, Geschäftsführer: Michael Gerdau, Tobias Dittmar
 Sitz Hamburg; HRB 89145 Amtsgericht Hamburg
 Vote against SPAM - see http://www.politik-digital.de/spam/
 Michael Gerdau       email: mgd@technosis.de
 GPG-keys available on request or at public keyserver

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Friday 15 June 2007 19:39:57 Michael Gerdau wrote:
> > > > What matters is *my* intent in *choosing* the GPLv2, not *his*
> > > > intent in writing it.
> > >
> > > I beg to differ. By adopting _his_ license you adopted his view. [...]
> >
> > ianal, but fortunately that's not what the law is. The license says what
> > it says, and that is what controls. The intent of the author (of Linus
> > and other copyright holders) is a secondary source of information /if
> > and only if/ any ambiguity of meaning arises (as determined by a judge,
> > not by you or me). But the opinion and intent of RMS (unless adopted by
> > Linus) is quite immaterial.
>
> I agree with the "/if and only if/ any ambiguity of meaning arises" part.
> I'm sorry I didn't make that clear before.
>
> However if that situation arises (i.e. the judge decides there is an
> ambiguity) then as far as my experience tells me it is the intention of
> the author (RMS et al in this case) that counts. But I erred before...

I doubt this. In a situation like that the intent of the licensor is what 
matter, not the intent of the original author of the license.

DRH

> Best wishes,
> Michael



-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Friday 15 June 2007 18:06:11 Michael Gerdau wrote:
> > > I find it obvious that the GPL was meant to prevent such to be
> > > possible. This is what I mean by the "the spirit of the GPL".
> >
> > Umm. It may well have been meant by *rms*. But your argument fatally
> > falls down on the fact that rms has had *nothing* to do with the Linux
> > kernel.
>
> While I raised this argument on the Lunix kernel ML it was not meant
> to be valid specifically for it.
>
> My observation in this thread is that almost everybody discusses different
> aspects of the same thing and everybody is somehow right. I was trying
> to "go back to start" and have the look at the overall picture which in
> this case for me is the question what the GPL's spirit is.
>
> Whether and which of it _you_ intended to adopt for the kernel I had
> no intention to sencond guess.
>
> > > Living in germany I'm also used to the courts valueing the intention
> > > over the exact wording of a contract (a licence after all is a
> > > contract). So I _think_ in germany TiVo would have lost a lawsuit if
> > > they had tried it.
> >
> > Ehh. The intent that matters is not the intent of the person who authored
> > the license, but the intent of the person who *chose* the license.
>
> That seems to imply that we have to deal with myriads of intended meanings,
> namely those of all who contributed to the kernel.
>
> I'm pretty sure I don't wish to walk that road. If you want to we'll have
> to agree to disagree.

Each person that contributed code to the linux kernel *CAN* have their own 
interpretations of the GPLv2 as it *APPLIES* to their code. The 
interpretation that matters when talking about the kernel, a a whole, is 
Linus'.

> > What matters is *my* intent in *choosing* the GPLv2, not *his* intent in
> > writing it.
>
> I beg to differ. By adopting _his_ license you adopted his view. If you
> don't like that then choose a different license (which obviously you are
> free to do).
>
> It's just not feaseable to have something like "my GPL means a different
> thing than your GPL".

Wrong. If I adopt the GPL it will be because of the *interpretation* I give it 
when reading it. And because *I* am the one then granting the license, it is 
*MY* interpretation that matters.

> > But to make it even less relevant: intent really only legally matters
> > when the legal issues are unclear.
> >
> > And they really aren't that unclear here.
>
> We have to agree to disagree then.
>
> > > If one wishes to prevent it there are two related questions:
> > > - does GPLv2 prevent it ?
> > > - if GPLv2 does not prevent it then how can we change it to achieve
> > > that ?
> >
> > Well, I think it's fairly unquestionable that the GPLv3 does prevent it.
> >
> > So your second question isn't even really interesting. We know the
> > answer. So the only question that is even remotely interesting is the
> > first one.
>
> My second question leaves out whether or not GPLv3 is an acceptable answer
> to my second question. While the FSF says it is it is by no means clear
> that I will agree -- all I wanted to do is present the situation as I see
> it.
>
> > Yes, I do agree with that reasoning, but there are *other*, and more
> > direct, reasons than just the FSF's answer to say that the answer to your
> > first question is "no".
>
> Note I only said the FSF seems to say "no". I'm not yet sure I do.
>
> > The fact is, plain reading of the license (which *always* takes
> > precedence over "intent", even in Europe) simply doesn't make what Tivo
> > did illegal.
>
> I disagree and I don't see that plain reading of the license is that
> obvious w/r to the SHA1 key because from a certain perspective said key
> is required to create a working modification which I'm entitled to under
> the GPLv2. I also agree that your perspective has merrit too. I'm simply
> not sure which of the above is "correct" (as in agreeable from a judge's
> PoV).

A plain reading of the license doesn't entitle you to create a "working 
modification". See the disclaimers of warranty and guarantee - sections 11 
and 12 of the GPLv2. They mean that the person *GRANTING* the license doesn't 
have to make sure that the program will be useful for your purposes, is 
modifiable to fit your purposes or will even *NOT* damage your hardware when 
it runs. "THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE PROGRAM IS 
WITH YOU." 

In other words there is no guarantee in the GPLv2 that you will be able to 
create a working modification. The *ONLY* guarantee that exists, in regards 
to modification, is that you will be able to make a modification. 

> Based on that I disagree with your above stmt, at least I don't think
> your implication every other reading being outright wrong is false.
>
> This thread IMO clearly shows that apparently it isn't that clear -- far
> too many intelligent people do disagree on this IMO not so obvious reading.

I agree that "too many intelligent people disagree". But this is just human 
nature. That people believe it says one thing when it doesn't is analogous to 
the *erroneous* belief that the world is flat.

> > You literally have to read the GPLv2 in ways that are obviously not true
> > to get to any other situation.
>
> I object against the word obvious as an obsolute measure. You have all
> right to consider it obvious from your PoV. My PoV may differ and I
> strongly claim it being equally valid.
>
> > For example, Alexandre made the same two mistakes over and over in his
> > reading when he tried to argue that the GPLv2 disallows what Tivo did:
>
> I do not agree with everything Alexandre wrote but I do agree with some
> parts.
>
> >  (a) The right to modify means "modify in place"
>
> [snip]
>
> > So clearly, the whole "modify in place" argument is simply *wrong*.
>
> I tend to agree and I didn't like it when it was brought up. But IMO that
> does not invalidate his position as such.

But the "modify in place" argument is part of the reasoning behind the claim 
that the "anti-tivoization" language in GPL3 doesn't break spirit with GPL2.

> >  (b) The language in the preamble: "must give the recipients all the
> >      rights that you have" means really *all* the rights and abilities!
>
> I always did imply a "within reason". To me that means "if it is simple
> for them to do it and can be simply extended to me as well then they have
> to extend it". Handing out a SHA1 key definitely is simple and thus IMO
> something I can expect them to do.

But the "within reason" isn't there. That some people have inferred that term 
applies doesn't matter. 

> I'm aware we have to disagree on this. I'm also aware that this is
> subject to the (IMO implied right) to run the modified code on the
> original HW, mostly because they can do that very easily (which would
> be covered by "within reason").
>
> BTW:
> I can easily take the position that from a metaphysical PoV the whole
> concept of a copy is an illusion because every copy is in fact a modified
> new version (that happens to be rather similar from a certain PoV... but
> I disgress and don't really wish to walk that bridge :)

You know, I actually could agree to that "metaphysical" view.

> > 	  (a) Again, Red Hat makes DVD's that contain GPLv2'd programs on
> > 	      them. Red Hat is bound by the GPL, so each work they put
> > 	      on the DVD is always under the GPL, and Red Hat *must* give
> > 	      you the rights that the GPL specifies.
>
> [snipped the paragraph on RH "withholding rights]
>
> Here the "within reason" kicks in. And I readily admit that you are an
> able intellectual acrobat who can twist arguments such that they become
> silly.

And here is where it becomes obvious that an inference is needed. *BUT* the 
argument was that the preamble, as written, states the intent of the license. 
Having to add an unwritten phrase invalidates the argument.

> My experience with german courts has shown me that the judges I had to
> deal with always and foremost did apply a reality check and did not try
> to bisect the consequences like an algorithm evaluated by a machine, i.e.
> the tried to decide what is right and wrong and not whether the letter
> of the contract could be twisted this or that way.

This is the way it should be. However, the letter of the contract, in this 
case, is very clear and that hasn't stopped Herr Welte at all.

> > 	  (b) Again, I make the Linux kernel available to you on a web
> > 	      site, and thus distribute it to you. Do I actually give you
> > 	      *all* the rights I have in it? Hell no. I cannot (and do
> > 	      not even want to). As an author, I have special rights in my
> > 	      code that you do not get. You get the rights spelled out in
> > 	      the GPLv2, and *nothing* more.
>
> I probably haven't read all Alexandre wrote but from what I read I'd be
> surprised he'd disagree (I don't).

He doesn't. But the fact is that he has claimed that the "give you all the 
rights I have" is to be read literally. Hence Linus' above statement.

> >      In other words, Alexandre's reading of the text in the preamble is
> >      *impossible*. It absolutely *cannot* be the way the license works.
> >      It's not how Red Hat itself reads it, and it's not how it can even
> >      legally be made to work even if somebody *wanted* to read it that
> >      way.
>
> I don't know whether Alexandre does read to preamble to mean what you
> imply he does. But whether he does or not is irrelevant for me to decide
> that IMO there is a strong argument to claim that what TiVo did is illegal
> even under the GPLv2. Note I'm merely saying "strong argument" as opposed
> to "clear beyond doubt".
>
> > > Whether or not the GPLv3 is truely an acceptable answer to prevent
> > > Tivoisation is a completely different issue that I can't really judge.
> >
> > Absolutely. I do think it prevents Tivoisation, but I personally think
> > it's unacceptable in even *trying* to prevent it, and as I've tried to
> > make clear, the GPLv2 definitely did *not* prevent Tivo from doing what
> > they did.
>
> Well, I think trying to prevent it is totally acceptable, after all I'm
> free to impose whatever restrictions to my code I see fit (I think it was
> long ago agreed that the GPLv2 does impose restrictions as well; they are
> just different).
>
> We have to agree to disagree on both whether trying to prevent Tivoisation
> is acceptable and whether GPLv2 already prevents it.

Agreed.

DRH

> Best wishes,
> Michael



-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Ingo Molnar]

* Daniel Hazelton <dhazelton@enter.net> wrote:

> > My experience with german courts has shown me that the judges I had 
> > to deal with always and foremost did apply a reality check and did 
> > not try to bisect the consequences like an algorithm evaluated by a 
> > machine, i.e. the tried to decide what is right and wrong and not 
> > whether the letter of the contract could be twisted this or that 
> > way.
> 
> This is the way it should be. However, the letter of the contract, in 
> this case, is very clear and that hasn't stopped Herr Welte at all.

btw., still ianal, but the GPLv2 is not a "contract" but a "pure 
copyright license". A contract, almost by definition is a restriction of 
rights in exchange for consideration - while if you accept the license 
of a GPLv2-ed work this act only gives rights that you did not have 
before. Furthermore when you get source code of free software then there 
is no "meeting of minds" needed for you to accept the GPL's conditions, 
and only the letter of the license (and, in case of any ambiguities, the 
intent of the author of the code) matters to the interpretation of the 
license, not the intent of the recipient. (while in contract cases both 
the meeting of minds is needed and the intent and understanding of both 
parties matters to the interpretation of the contract.)

	Ingo

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Sat, 16 Jun 2007, Ingo Molnar wrote:
>
> btw., still ianal, but the GPLv2 is not a "contract" but a "pure 
> copyright license".

I've been told by several independent sources that it really doesn't 
matter.

The "pure license" argument was born largely for silly reasons: people 
claimed (a _loong_ time ago) that the GPL wasn't enforceable in the US 
because in order to be enforceable, something of value has to change hands 
(in the US, for example, it would be common to "sell" something for a 
nominal sum of $1 USD rather than to give it outright, to "seal the deal" 
and make it irrevocable).

That's generally considered a specious argument, apparently. In most 
jurisdictions in the US, a license and a contract are judged to be legally 
exactly the same thing, and if you don't follow the GPL and have no other 
contract to show for it, you're in violation of federal copyright law, so 
whether it is a license or a contract really doesn't matter.

So it's true: the GPL just gives you rights, and without it you have no 
rights (other than fair use ones etc), and blah blah. But the distinction 
between "license" vs "contract" really isn't a very important one in any 
case.

> Furthermore when you get source code of free software then there is no 
> "meeting of minds" needed for you to accept the GPL's conditions, and 
> only the letter of the license (and, in case of any ambiguities, the 
> intent of the author of the code) matters to the interpretation of the 
> license, not the intent of the recipient. (while in contract cases both 
> the meeting of minds is needed and the intent and understanding of both 
> parties matters to the interpretation of the contract.)

I do agree that you can probably use this to say that the intent of the 
copyright has a stronger position, and that his "intent" thus matters 
more.

But I suspect that the "intent" angle is fairly weak legally to begin 
with, and if you cannot show that the intent was mutual, it's probably 
weaker still. So yeah, the intent of the copyright owner arguably might 
matter more, but quite frankly, I suspect everbody is better off not 
worrying so much about "intent", and worrying more about the "terms and 
conditions" part.

(I've said several times that intent _matters_, I just don't want people 
to think that it matters a whole lot).

What is pretty clear, though, is that the intent of a third party in the 
license/cotnract matters not at all. In the case of the kernel, the FSF 
being such a third party.

			Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Rob Landley]

On Friday 15 June 2007 18:59:14 Linus Torvalds wrote:
> So it's true: the GPL just gives you rights, and without it you have no
> rights (other than fair use ones etc), and blah blah. But the distinction
> between "license" vs "contract" really isn't a very important one in any
> case.

Er, copyright law is federal, contract law is generally state level?  So not 
only does contract law vary a lot more by jurisdiction, but it's enforced by 
different courts than suits over copyright?  (You'll notice the GPL doesn't 
say which state law holds sway.  If it was a contract this would be kind of 
important.)

Also, in addition to the "exchange of value" bit there's "privity of contract" 
and "informed consent" when dealing with contract, which are cans of worms 
which can be avoided by Not Going There (tm)...

(These were largeish issues in the SCO vs Novell case, involving lots of 
motions in Utah detailed blow-by-blow on Groklaw...)

Rob
-- 
"One of my most productive days was throwing away 1000 lines of code."
  - Ken Thompson.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Michael Gerdau]

[-- Attachment #1: Type: text/plain, Size: 1112 bytes --]

> On Friday 15 June 2007 18:59:14 Linus Torvalds wrote:
> > So it's true: the GPL just gives you rights, and without it you have no
> > rights (other than fair use ones etc), and blah blah. But the distinction
> > between "license" vs "contract" really isn't a very important one in any
> > case.
> 
> Er, copyright law is federal, contract law is generally state level?  So not 
> only does contract law vary a lot more by jurisdiction, but it's enforced by 
> different courts than suits over copyright?  (You'll notice the GPL doesn't 
> say which state law holds sway.  If it was a contract this would be kind of 
> important.)

That seems to be a special property of the US legal system. At least I'm
not aware of this or a similar distinction in e.g. germany (or most parts
of europe AFAIK).

Best,
Michael
-- 
 Technosis GmbH, Geschäftsführer: Michael Gerdau, Tobias Dittmar
 Sitz Hamburg; HRB 89145 Amtsgericht Hamburg
 Vote against SPAM - see http://www.politik-digital.de/spam/
 Michael Gerdau       email: mgd@technosis.de
 GPG-keys available on request or at public keyserver

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Ingo Molnar]

* Linus Torvalds <torvalds@linux-foundation.org> wrote:

> > Furthermore when you get source code of free software then there is 
> > no "meeting of minds" needed for you to accept the GPL's conditions, 
> > and only the letter of the license (and, in case of any ambiguities, 
> > the intent of the author of the code) matters to the interpretation 
> > of the license, not the intent of the recipient. (while in contract 
> > cases both the meeting of minds is needed and the intent and 
> > understanding of both parties matters to the interpretation of the 
> > contract.)
> 
> I do agree that you can probably use this to say that the intent of 
> the copyright has a stronger position, and that his "intent" thus 
> matters more.
> 
> But I suspect that the "intent" angle is fairly weak legally to begin 
> with, and if you cannot show that the intent was mutual, it's probably 
> weaker still. So yeah, the intent of the copyright owner arguably 
> might matter more, but quite frankly, I suspect everbody is better off 
> not worrying so much about "intent", and worrying more about the 
> "terms and conditions" part.

yeah - and from everything i know about this subject the distinction 
between contract and license is small and more of a technicality - but 
still, it's a nice touch that the "pure license" argument that the FSF 
has advanced for a long time (and which it is now more silent about, 
given the GPLv3's not so pure structure) neatly defeats the common 
argument:

  "but, but, when i received the Tivo with GPL-ed software on it the 
   GPLv2 was not intended to be like that, there is a right to run 
   modified software on the hardware!"

> What is pretty clear, though, is that the intent of a third party in 
> the license/cotnract matters not at all. In the case of the kernel, 
> the FSF being such a third party.

yeah. But the argument goes a bit further: people who chose to _license_ 
the kernel (by receiving a Tivo for example and downloading its kernel 
source) claim that _their_ interpretation of the GPL is that of the 
FSF's and that Tivo ought to follow it. The whole "Tivo is cheating the 
GPL deal with the end users" line of PR. As far as license 
interpretation goes there is _no end user deal_ and the 'end user' does 
not even play in terms of intent - only if she choses to be an active 
member of the community. That's why i think it's better to talk about a 
license than a contract. (even though legally, at least in the US, the 
two are quite close to each other.)

so a 'pure copyright license' stresses the point even more that you only 
really count in the ecosystem if you contribute in one way or another. 
The system should be and _is_ assymetric towards the actual black letter 
text of the license and, as a second layer, towards the intent of the 
people who actually produced this 1+ billion lines of code, 
documentation, bugzilla entries and other nice works. And that is a 
thing the FSF is missing sometimes i believe - the "listen to _all_ the 
people who enabled this cool stuff" part.

	Ingo

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

> * Daniel Hazelton <dhazelton@enter.net> wrote:

>> > My experience with german courts has shown me that the judges I had 
>> > to deal with always and foremost did apply a reality check and did 
>> > not try to bisect the consequences like an algorithm evaluated by a 
>> > machine, i.e. the tried to decide what is right and wrong and not 
>> > whether the letter of the contract could be twisted this or that 
>> > way.

>> This is the way it should be. However, the letter of the contract, in 
>> this case, is very clear and that hasn't stopped Herr Welte at all.

And this is the beauty of a multi-author project.  Even if some
authors think that the license permits something, if any of them
understands it doesn't, he can try to enforce that WRT his own
contributions.  So those exploiting the gray areas of the license can
still get caught.

On Jun 15, 2007, Ingo Molnar <mingo@elte.hu> wrote:

> btw., still ianal, but the GPLv2 is not a "contract" but a "pure 
> copyright license". A contract, almost by definition is a restriction of 
> rights in exchange for consideration - while if you accept the license 
> of a GPLv2-ed work this act only gives rights that you did not have 
> before.

In Brazil, this is kind of contract/license is called a beneficial
contract.

> Furthermore when you get source code of free software then there 
> is no "meeting of minds" needed for you to accept the GPL's conditions, 
> and only the letter of the license (and, in case of any ambiguities, the 
> intent of the author of the code) matters to the interpretation of the 
> license, not the intent of the recipient.

That's correct, but with a catch: since the contract or license is
chosen by the licensor, in case of ambiguity in the terms, many courts
will interpret it in a way that privileges the licensee, regardless of
the fact that copyright licenses are to be interpreted restrictively
(at least in Brazilian law).  And IANAL ;-)

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Scott Preece]

On 6/15/07, Alexandre Oliva <aoliva@redhat.com> wrote:
> > * Daniel Hazelton <dhazelton@enter.net> wrote:
>
> That's correct, but with a catch: since the contract or license is
> chosen by the licensor, in case of ambiguity in the terms, many courts
> will interpret it in a way that privileges the licensee, regardless of
> the fact that copyright licenses are to be interpreted restrictively
> (at least in Brazilian law).  And IANAL ;-)
---

Hmm. In such a suit, however, the user would not be "the licensee" and
would not be a party to the suit - some author would be the plaintiff
and would be suing someone for doing something in violation of the
license that author granted - that is, the *defendant* would be the
licensee who would get the benefit of the doubt...

scott

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 15, 2007, "Scott Preece" <sepreece@gmail.com> wrote:

> On 6/15/07, Alexandre Oliva <aoliva@redhat.com> wrote:
>> > * Daniel Hazelton <dhazelton@enter.net> wrote:
>> 
>> That's correct, but with a catch: since the contract or license is
>> chosen by the licensor, in case of ambiguity in the terms, many courts
>> will interpret it in a way that privileges the licensee, regardless of
>> the fact that copyright licenses are to be interpreted restrictively
>> (at least in Brazilian law).  And IANAL ;-)
> ---

> Hmm. In such a suit, however, the user would not be "the licensee" and
> would not be a party to the suit - some author would be the plaintiff
> and would be suing someone for doing something in violation of the
> license that author granted - that is, the *defendant* would be the
> licensee who would get the benefit of the doubt...

Yes.  And so justice is made.  Licensor gets to pick the license,
licensee gets the benefit of the doubt.  What's the 'however' about?
Was this not obvious?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Scott Preece]

On 6/15/07, Alexandre Oliva <aoliva@redhat.com> wrote:
> On Jun 15, 2007, "Scott Preece" <sepreece@gmail.com> wrote:
>
> > On 6/15/07, Alexandre Oliva <aoliva@redhat.com> wrote:
> >> > * Daniel Hazelton <dhazelton@enter.net> wrote:
> >>
> >> That's correct, but with a catch: since the contract or license is
> >> chosen by the licensor, in case of ambiguity in the terms, many courts
> >> will interpret it in a way that privileges the licensee, regardless of
> >> the fact that copyright licenses are to be interpreted restrictively
> >> (at least in Brazilian law).  And IANAL ;-)
> > ---
>
> > Hmm. In such a suit, however, the user would not be "the licensee" and
> > would not be a party to the suit - some author would be the plaintiff
> > and would be suing someone for doing something in violation of the
> > license that author granted - that is, the *defendant* would be the
> > licensee who would get the benefit of the doubt...
>
> Yes.  And so justice is made.  Licensor gets to pick the license,
> licensee gets the benefit of the doubt.  What's the 'however' about?
> Was this not obvious?
---

Sorry - I thought you were saying ambiguity would be resolved in favor
of the user. If you meant in favor of the licensee (regardless of that
limiting the user's rights), then I agree.

scott

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Sat, 16 Jun 2007, Michael Gerdau wrote:
> 
> I beg to differ. By adopting _his_ license you adopted his view.

I'm sorry, but that's simply bullshit. 

The GPLv2 does not state that you have to become a slave of rms and follow 
him in all things, and agree with him. Really. You must have read some 
other (perhaps unreleased early draft?) version.

The GPLv2 says what it says. Not what you (or rms) *wished* it says.

You don't enter into contracts and licenses based on wishes and intents. 
That's just not how it works. 

> >  (b) The language in the preamble: "must give the recipients all the 
> >      rights that you have" means really *all* the rights and abilities!
> 
> I always did imply a "within reason". 

Your view is not relevant. The fact that the "preamble" is not the 
"conditions" is what's relevant. 

The preamble is explicitly stated to be *different* from the exact 
conditions. It's not the real "terms of copying". It's there to explain, 
it's not there to *be* the license.

It's explanatory, but the wording that actually *matters* is the "terms 
and conditions".

And the fact that *you* can mentally add words to it when you read the 
license (adding a "within reason") has absolutely no relevance 
what-so-ever.

			Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Rob Landley]

On Friday 15 June 2007 13:03:53 Linus Torvalds wrote:

> 	      But does Red Hat actually give you *all* the rights they
> 	      hold on the DVD? No, they definitely do not. They hold a
> 	      compilation copyright on RHEL, and they very  much do *not*
> 	      give you the right to copy the whole distribution and sell
> 	      it as RHEL. You only get the rights to the individual
> 	      pieces, not to the whole thing!

Technically what they're holding back is _trademark_ rights, which are a 
different area of IP law and not addressed by the GPL.  (I know you know 
this, but just for the record...)

The five main areas of IP law as I understand them are copyright, patent, 
trademark, contract, and trade secret.  Each of which is a different animal 
with a different legal foundation and different enabling legislation.  The 
GPL is a copyright license with some language about patents.  It is not based 
on contract law (although that's a common misperception that Lawrence Lessig 
and Eben Moglen have spent some effort debunking), and doesn't even mention 
trademarks.

So Red Hat isn't saying "you can use some of our copyrights but not others", 
last I heard all of their copyrights are licensed GPLv2 as a matter of 
corporate policy.  What you can't use is their trademarked name or logo, 
because they are explicitly refusing to license the trademarks to third 
parties.

And under GPLv2, this is allowed.

Rob
-- 
"One of my most productive days was throwing away 1000 lines of code."
  - Ken Thompson.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Fri, 15 Jun 2007, Rob Landley wrote:
> 
> Technically what they're holding back is _trademark_ rights, which are a 
> different area of IP law and not addressed by the GPL.  (I know you know 
> this, but just for the record...)

No, technically Red Hat really *does* have copyrights of their own.

Red Hat owns the "compilation copyright" on their distribution. That 
means, for example, that even if they have _only_ open source programs on 
their DVD image, you still are not necessarily able (without their 
permission) to set up a "cheap-cd's" kind of operation, and sell their 
CD-ROM/DVD images for a lower price.

So yes, they do own the Red Hat trademark too, but they fundamentally do 
own copyrights over and beyond those of the individual programs they 
distribute!

Now, I think it so happens that the RHEL DVD's contains other programs 
than just open source, and that you couldn't legally copy them *anyway*, 
but that's a different issue.

Also, happily, a lot of vendors do not *want* to exercise their 
copyright in the compilation, so you can go to cheapbytes.com, and you'll 
find Fedora CD's, OpenSuSE CD's, Ubuntu CD's, etc, and as far as I know, 
they're all perfectly legal. Exactly because open-source vendors usually 
don't want to look nasty by limiting the compilation, when they can't 
really limit the individual parts anyway.

> The five main areas of IP law as I understand them are copyright, patent, 
> trademark, contract, and trade secret.

I'd not put contract there, but fair enough. But what I was really trying 
to point out is that there are many different "levels" of copyright.

So you can own a "copyright in the compilation" - which just means that 
you own the details of how you set it all together - _without_ actually 
necessarily owning the copyrights in any of the individual packages 
(although you obviously have to have a license to _make_ a compilation of 
them - but the GPLv2 is one such license).

			Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Rob Landley]

On Friday 15 June 2007 21:29:22 Linus Torvalds wrote:
> On Fri, 15 Jun 2007, Rob Landley wrote:
> > Technically what they're holding back is _trademark_ rights, which are a
> > different area of IP law and not addressed by the GPL.  (I know you know
> > this, but just for the record...)
>
> No, technically Red Hat really *does* have copyrights of their own.

They do have copyrights.  They license them under the GPL, and afterwards they 
still have them.

> Red Hat owns the "compilation copyright" on their distribution. That
> means, for example, that even if they have _only_ open source programs on
> their DVD image, you still are not necessarily able (without their
> permission) to set up a "cheap-cd's" kind of operation, and sell their
> CD-ROM/DVD images for a lower price.

I agree that they have this right, but that wasn't the rationale they gave in 
the cease and desist letters they sent out in 2001.  Those said it was ok to 
redistribute, but you can't use their trademarks to promote it when you did 
so:
http://www.newsforge.com/article.pl?sid=01/12/10/2014239&mode=thread

[Rummages around for their current policy statement...]

The restriction is embodied in their "trademark guidelines and policies":
http://www.redhat.com/about/companyprofile/trademark/

If you open the PDF:
http://www.redhat.com/f/pdf/corp/RH-3573_284204_TM_Gd.pdf

Page one, the right hand side talks about copyright, second paragraph:

> At the same time, the combined body of work that consitutes Red Hat (R)
> Enterprise Linux (R) is a collective work which has been organized by Red
> hat, and Red hat holds the copyright in that collective work.  Red hat then
> permits others to copy, modify and redistribute the collective work.  To
> grant this permission Red hat usually uses the GNU General Public License
> ("GPL") version 2 and Red Hat's own End User License Agreement.  Although
> software licensed under the GPL is "open source software," Red Hat retains
> ownership of the copyright in its collective work.  If someone violates the
> GPL regarding that collective work, only Red Hat, as the copyright owner and
> licensor of that collective work, has legal authority to enforce the GPL
> against the violator.  Although Red hat "owns" the collective work, in
> licenseing it under the GPL, Red Hat grants broad rights in the collective
> work to others.  Neither the GPL nor Red Hat's End User License Agreement
> grant any right to use Red hat's trademarks in the redistribution of the
> collective work.

That seems to say that their compilation copyright is also licensed under the 
GPL, and what they're not giving permission to use is the trademark.

> So yes, they do own the Red Hat trademark too, but they fundamentally do
> own copyrights over and beyond those of the individual programs they
> distribute!

I agree they claim compilation copyrights.  But they seem to have licensed 
their compilation copyrights under GPLv2.

If they're including GPLed works in the compilation, this may actually be a 
requirement.  (Lawyers would happily fight over this issue for months: 
asserting a copyright over the aggregation takes the "mere" out of it, don't 
you think?  Is a compilation a derived work of the components that were 
compiled, at least for the purposes of GPLv2?  Can it be "mere aggregation" 
if you're enforcing a copyright on that aggregation?  Does it not then become 
a larger work with GPL components?  I dunno.)

You'll notice that back under Bob Young, Red Hat carefully didn't go there, by 
licensing the compilation GPLv2 and segregating incompatibly licensed content 
to a separate CD.  After the IPO, he retired and different management took 
over, and introduced Red Hat Enterprise to eat Sun's market[1], and who knows 
what they're thinking now?  I suspect their lawyers still want the 
GPL-incompatible stuff on a separate CD so they can sleep at night.

(Either that, or they just don't assert a compilation copyright, but why give 
that up if they don't have to?)

> Now, I think it so happens that the RHEL DVD's contains other programs
> than just open source, and that you couldn't legally copy them *anyway*,
> but that's a different issue.

The existence of CentOS seems to argue against this.  They say:

> The vast majority of changes made will be made to comply with the upstream
> vendor's re-distribution policies concerning trademarked names or logos.
> Any other changes made will be spelled out in the Release Notes for the
> individual CentOS product.

I haven't noticed any specific non-GPL packages removed from Centos.  Buried 
down in their FAQ they say they're building the same set of packages as are 
in Red Hat Enterprise AS.

I suspect the reason for a lack of obvious proprietary stuff in the RHEL base 
distro is that Red Hat's lawyers don't want to give up the option to enforce 
the compilation copyright on said base distro, but also don't want to raise 
the spectre of enforcing a copyright on a derived work of GPL code (the 
compilation, which includes GPL code and if copyrighted probably isn't "mere 
aggregation") that contains code which can't be licensed under the GPL.

I don't know the answer to this one.  This is the sort of thing that gives 
lawyers ulcers.  (And yachts.)  But I can believe their lawyers want to keep 
their options open if they're sufficiently unsure how the hypothetical would 
play out in court...

> Also, happily, a lot of vendors do not *want* to exercise their
> copyright in the compilation, so you can go to cheapbytes.com, and you'll
> find Fedora CD's, OpenSuSE CD's, Ubuntu CD's, etc, and as far as I know,
> they're all perfectly legal.

Red Hat created the "Fedora" trademark to have a separate and more liberally 
licensed trademark that people like cheapbytes.com could use without 
reflecting on Red Hat Enterprise.  Unfortunately, trying to find reference 
for this is non-obvious, because, the Fedora Trademarks page is: 
http://fedoraproject.org/wiki/Legal/TrademarkGuidelines

Which contains no actual content, but instead links to a fedora.redhat.com 
page which is 404. :P

But again, it's trademark not copyright they're restricting.

> > The five main areas of IP law as I understand them are copyright, patent,
> > trademark, contract, and trade secret.
>
> I'd not put contract there, but fair enough. But what I was really trying
> to point out is that there are many different "levels" of copyright.
>
> So you can own a "copyright in the compilation" - which just means that
> you own the details of how you set it all together - _without_ actually
> necessarily owning the copyrights in any of the individual packages
> (although you obviously have to have a license to _make_ a compilation of
> them - but the GPLv2 is one such license).

Sure, it's spelled out in some detail by the copyright office right here:
http://www.copyright.gov/circs/circ14.html#compilations

And in less penetrable legalese here [2]:
http://www.copyright.gov/title17/92chap1.html#103

However, mostly I just deal with straightforward derivative works when 
worrying about GPL enforcement.  (When I brought up compilation copyright to 
the SFLC guys last year in relation to enforcing the busybox copyrights, they 
didn't think they needed to go there since between Erik and myself they 
represent copyrights to a significant portion of the entire codebase, 
including the central plumbing you can't run the thing without).

Intellectual property law has been a hobby of mine ever since I visited the 
Patent and Trademark office in washington DC with my grandfather back in the 
90's, and picked up the big green brochures explaining this stuff.  I think I 
still have the badge in a box somewhere.  There's all SORTS of weird corner 
cases to this stuff.  (Did you know the federal government can't create a 
copyright?  Everything created by federal employees is automatically in the 
public domain.  So the way they get around that is they hire contractors, 
have them create copyrighted materials, and transfer the copyrights to the 
government...)

> 			Linus

Rob

[1] Large purchasing contracts, especially for government projects, often cap 
a supplier's profits at a percentage of costs.  This means the way for a 
supplier to get a larger profit off the contract is to user more expensive 
components.  So they'd rather grab $5000/seat OS than a $20/seat OS because 
the same 10% profit is $500 vs $2.  This is a big reason you saw lots of 
Solaris and AIX and such in Navy bids 10 years ago: it was popular _because_ 
it was expensive.  Then Red Hat figured out "wait, you mean if I can come up 
with an excuse to charge a couple orders of magnitude more per seat, MORE 
people will buy it?"  And they invented "Red Hat Enterprise" to soak that 
kind of purchaser for all they were worth, and it was so lucrative they 
retreated up into that space to attack the soft underbelly of Sun...

[2]  The section relating specifically to computer programs is interesting 
reading too:  http://www.copyright.gov/title17/92chap1.html#117

The stuff about "archival copies" is explicitly statutory.  When I first heard 
about it I thought it was just case law...
-- 
"One of my most productive days was throwing away 1000 lines of code."
  - Ken Thompson.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alan Cox]

> Red Hat created the "Fedora" trademark to have a separate and more liberally 
> licensed trademark that people like cheapbytes.com could use without 
> reflecting on Red Hat Enterprise.  Unfortunately, trying to find reference 
> for this is non-obvious, because, the Fedora Trademarks page is: 
> http://fedoraproject.org/wiki/Legal/TrademarkGuidelines
> 
> Which contains no actual content, but instead links to a fedora.redhat.com 
> page which is 404. :P

That would be a bug 8)

> But again, it's trademark not copyright they're restricting.

Correct - all the vendors face the problem that there are people out
there who want to try and pass crap on using someone elses good name
whether its toothpaste, perfume or software (or in the case of RHEL
software/services/support bundles)

The Fedora mark is thus used to make sure that if you get a Fedora CD, it
actually has -Fedora- on it etc.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Paul Mundt]

On Thu, Jun 14, 2007 at 02:51:13AM -0300, Alexandre Oliva wrote:
> On Jun 14, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> 
> > I've never had a reason to want to change the way any device like a TiVO 
> > works. So I can't comment on this.
> 
> Have you never wanted to improve any aspect of the software in your
> cell phone?  In your TV, VCR, DVD player, anything?  In the microwave
> oven, maybe?
> 
This is perhaps the part that's the most interesting. For the very small
number of people that _do_ want to change these things (usually at the
expense of a voided warranty, in the consumer device case), there's
always a way to make these changes, even if you must resort to hardware
hacking. Trying to mandate this sort of functionality in the license
might make it easier for a few people to get their code loaded, but the
vast majority of users have zero interest in anything like this.

I don't see how you can claim that the vendor is infringing on your
freedom, _you_ made the decision to go out and buy the product knowing
that the vendor wasn't going to go out of their way to help you hack
the device. In many cases the vendor doesn't even have the option
(802.11b channels and certification come to mind, GSM, etc.) of opening
things up to the end user, and making changes to the license isn't going
to magically change any of this.

If you don't like what the vendor has done with the product, you have the
freedom to not support the vendor, and to try and encourage people to
follow suit. As an example, I simply opted not to buy a tivo since I
wasn't able to do what I wanted with it out of the box, rather than
opting to rant about it (or coin an idiotic buzzword) much to the dismay
of every other person on a mailing list. This was neither something I
lost a great deal of sleep over, nor did I at any time feel like my
freedom was being eroded. True story.

If the vendor's bottom line is measurably impacted, they may even
reevaluate their position on supporting device hacking, but it's
certainly not going to be through draconian licensing that vendors
suddenly decide to play nice.

There were certainly enough vendors that followed the letter of the
GPLv2 without following the spirit of the license, with varied benefit
(especially with consumer device vendors). Imposing additional
constraints under the guise of the FSF's current version of "freedom"
isn't going to get these sorts of vendors working any better with the
community. You could of course argue that these vendors have nothing to
offer the community, and so they shouldn't be tolerated at all, but that
assumes that being able to load arbitrary code on their hardware and
the usefulness of whatever contributions they have to make to the software
are directly coupled some how. This has never been the case, and it only
seems to be a mindset that has started circulating when the GPLv3 came
about.

One idly wonders who exactly the FSF feels they're speaking for at this
point..

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bernd Paysan]

[-- Attachment #1: Type: text/plain, Size: 2609 bytes --]

On Thursday 14 June 2007 09:32, Paul Mundt wrote:
> This is perhaps the part that's the most interesting. For the very small
> number of people that _do_ want to change these things (usually at the
> expense of a voided warranty, in the consumer device case), there's
> always a way to make these changes, even if you must resort to hardware
> hacking. Trying to mandate this sort of functionality in the license
> might make it easier for a few people to get their code loaded, but the
> vast majority of users have zero interest in anything like this.

I don't feel this is a very conclusive argument.

How many computer users do want to change their OS? I mean not only want to 
change the OS in the sense of "apply patches released by Microsoft", but on 
their own? Many typical computer users ask for help to "fix their computer" 
when turning it off and on again already "fixes it". They would never ever 
change the source code of their OS even if they technically could do it - 
they are not programmers.

However, if there is the technical possibility to change the firmware of an 
appliance, somebody does it, and often mere users upload these changes to 
their own device (like the OpenWRT stuff).

Let me give one example: My parents own a DVB-T DVR. It was a cheap one, and 
it was cheap because the software is lackluster. Unfortunately it isn't 
free. Many users of this device complain to the manufacturer about the 
stability and quality of the software, but with no avail - there haven't 
been any updates in the last two years. I suppose I would be able to fix 
the problem, most other users probably wouldn't (and my parents neither). 
But if I did fix the problem, and provided them with an updated firmware, 
they would install it on their device.

That's the "help your neighbour" right in the GNU manifesto. It's as 
important as the "help yourself" right, maybe even more. It was the 
original motivation of RMS to make free software - the frustration of not 
being able to help his neighbours. He had an NDA to help himself.

What people want is software that works. If the firmware of your microwave 
or DVR works, you don't care so much if it is free or not. You only care if 
it doesn't work, and you feel the urge to fix it (and turning it off and on 
again doesn't fix it). That's why people complain loud about ATI drivers 
not being open, and don't care that much about the Nvidia driver, which is 
just as closed, but works.

-- 
Bernd Paysan
"If you want it done right, you have to do it yourself"
http://www.jwdt.com/~paysan/

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Paul Mundt <lethal@linux-sh.org> wrote:

> I don't see how you can claim that the vendor is infringing on your
> freedom, _you_ made the decision to go out and buy the product knowing
> that the vendor wasn't going to go out of their way to help you hack
> the device.

But I also made this decision fully aware that the software included
in the package was published under a license that said I was entitled
to modify it.  More than once I purchased a device that claimed to
have GNU/Linux software on it, only to find out that I couldn't use
the freedoms, because the distributor was infringing the license in
various ways.

> If you don't like what the vendor has done with the product, you have the
> freedom to not support the vendor, and to try and encourage people to
> follow suit.

Sure.  But wouldn't it be nice if the copyright holder could also help
in this effort?  It doesn't mean the copyright holder has to: s/he can
always grant an additional permission, or simply refrain from
enforcing this provision of the license.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> <somewhat sarcastic>
> And the companies that produce devices that come with Linux and/or
> other GPL'd software installed and place limits such that only
> people that have purchased that hardware have access to the
> "modified" source running on the device are following the letter,
> and the spirit, of the GPL.

WAIT, WAIT, THAT'S... :-)

> Before you start yelling I'm wrong, think about it this way: they
> make the source available to the people that they've given binary
> versions to, and there is nothing stopping one of those people from
> making the source available to the rest of the world.

The *only* in your sentence betrayed you.

If they place the limits such that nobody else can access the sources,
they're in violation of the license.

If they merely refrain from distributing the sources to others, but
still enable the recipients to do so, this is not a violation of the
license.

But then IANAL.

> *AND* the GPL has never been about making the source available to
> everyone - just to those that get the binaries.

Exactly.  Not even to the upstream distributor.  That's where Linus'
theory of tit-for-tat falls apart.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Chris Friesen]

Alexandre Oliva wrote:
> On Jun 14, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

>>*AND* the GPL has never been about making the source available to
>>everyone - just to those that get the binaries.

> Exactly.  Not even to the upstream distributor.  That's where Linus'
> theory of tit-for-tat falls apart.

Nope.

case 1:  Upstream provides source, tivo modifies and distributes it (to 
their customers).

case 2: tivo provides source, end user modifies and distributes it 
(possibly to their customers, maybe to friends, possibly even to upstream).

See?  Tit for tat.

Chris

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, "Chris Friesen" <cfriesen@nortel.com> wrote:

> Alexandre Oliva wrote:
>> On Jun 14, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

>>> *AND* the GPL has never been about making the source available to
>>> everyone - just to those that get the binaries.

>> Exactly.  Not even to the upstream distributor.  That's where Linus'
>> theory of tit-for-tat falls apart.

> Nope.

> case 1:  Upstream provides source, tivo modifies and distributes it
> (to their customers).

> case 2: tivo provides source, end user modifies and distributes it
> (possibly to their customers, maybe to friends, possibly even to
> upstream).

> See?  Tit for tat.

case 2': tivo provides source, end user tries to improve it, realizes
the hardware won't let him and gives up

Where's the payback, or the payforward?

And then, tit-for-tat is about equivalent retaliation, an eye for an
eye.  Where's the retaliation here?

If GPLv2 were tit-for-tat, if someone invents artifices to prevent the
user from making the changes the user wants on the software, wouldn't
it be "equivalent retaliation" to prevent the perpetrator from making
the changes it wants on the software?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Thursday 14 June 2007 18:45:07 Alexandre Oliva wrote:
> On Jun 14, 2007, "Chris Friesen" <cfriesen@nortel.com> wrote:
> > Alexandre Oliva wrote:
> >> On Jun 14, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> >>> *AND* the GPL has never been about making the source available to
> >>> everyone - just to those that get the binaries.
> >>
> >> Exactly.  Not even to the upstream distributor.  That's where Linus'
> >> theory of tit-for-tat falls apart.
> >
> > Nope.
> >
> > case 1:  Upstream provides source, tivo modifies and distributes it
> > (to their customers).
> >
> > case 2: tivo provides source, end user modifies and distributes it
> > (possibly to their customers, maybe to friends, possibly even to
> > upstream).
> >
> > See?  Tit for tat.
>
> case 2': tivo provides source, end user tries to improve it, realizes
> the hardware won't let him and gives up

Faulty logic. The hardware doesn't *restrict* you from *MODIFYING* any fscking 
thing. 

DRH

>
> Where's the payback, or the payforward?
>
> And then, tit-for-tat is about equivalent retaliation, an eye for an
> eye.  Where's the retaliation here?
>
> If GPLv2 were tit-for-tat, if someone invents artifices to prevent the
> user from making the changes the user wants on the software, wouldn't
> it be "equivalent retaliation" to prevent the perpetrator from making
> the changes it wants on the software?



-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> Faulty logic. The hardware doesn't *restrict* you from *MODIFYING*
> any fscking thing.

Ok, lemme try again:

case 2'': tivo provides source, end user tries to improve it, realizes
the hardware won't let him use the result of his efforts, and gives up

> On Thursday 14 June 2007 18:45:07 Alexandre Oliva wrote:
>> Where's the payback, or the payforward?
>> 
>> And then, tit-for-tat is about equivalent retaliation, an eye for an
>> eye.  Where's the retaliation here?
>> 
>> If GPLv2 were tit-for-tat, if someone invents artifices to prevent the
>> user from making the changes the user wants on the software, wouldn't
>> it be "equivalent retaliation" to prevent the perpetrator from making
>> the changes it wants on the software?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Fri, 15 Jun 2007, Alexandre Oliva wrote:
> 
> case 2'': tivo provides source, end user tries to improve it, realizes
> the hardware won't let him use the result of his efforts, and gives up

So you're blaming Tivo for the fact that your end user was a lazy bum and 
wanted to take advantage of somebody elses hard work without permission?

Quite frankly, I know who the bad guy in that scenario is, and it ain't 
Tivo. It's your lazy bum, that thought he would just take what Tivo did, 
sign the contract, and then not follow it. And just because the box 
_contained_ some piece of free software, that lazy bum suddenly has all 
those rights? Never mind all the *other* effort that went into bringing 
that box to market?

You do realize that Tivo makes all their money on the service, don't you? 
The actual hardware they basically give away at cost, exactly to get the 
service contracts. Not exactly a very unusual strategy in the high-tech 
world, is it?

You know what? I respect the pro-FSF opinions less and less, the more you 
guys argue for it. Michael Poole seems to argue that things like fair use 
shouldn't exist, and even the cryptographic _signatures_ of the programs 
should be under total control of the copyright owner.

And you seem to argue that it's perfectly fine to ignore the people who 
design hardware and the services around them, and once you have that piece 
of hardware in your grubby hands you can do anythign you want to it, and 
_their_ rights and the contracts you signed don't matter at all.

Guys, you should be ashamed of calling yourself "free software" people. 

You sound more like the RIAA/MPAA ("we own all the rights! We _own_ your 
sorry asses for even listening to our music") and a bunch of whiners that 
think that just because you have touched a piece of hardware you 
automatically can do anythign you want to it, and nobody elses rights 
matter in the least!

Guys, in fighting for "your rights", you should look a bit at *other* 
peoples rights too. Including the rights of hw manufacturers, and the 
service providers. Because this is all an eco-system, where in order to 
actually succeed, you need to make _everybody_ succeed.

			Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 15, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:

> On Fri, 15 Jun 2007, Alexandre Oliva wrote:
>> 
>> case 2'': tivo provides source, end user tries to improve it, realizes
>> the hardware won't let him use the result of his efforts, and gives up

> So you're blaming Tivo for the fact that your end user was a lazy bum and 
> wanted to take advantage of somebody elses hard work without permission?

-ENONSEQUITUR

> Quite frankly, I know who the bad guy in that scenario is, and it ain't 
> Tivo. It's your lazy bum, that thought he would just take what Tivo did, 
> sign the contract, and then not follow it. And just because the box 
> _contained_ some piece of free software, that lazy bum suddenly has all 
> those rights?

Yes, because the software license that TiVo signed up for says that
TiVo must pass on certain rights and not impose any further
restrictions.

And all that because TiVo wanted to use kernel and userland that were
readily available, and at no cost other than respecting others'
freedoms, while at that?

Who's the lazy bum, again?

> And you seem to argue that it's perfectly fine to ignore the people who 
> design hardware and the services around them,

Just like they seem to think it's perfectly fine to ignore a number of
people who design and maintain the software they decided to use in
their hardware.

> Guys, you should be ashamed of calling yourself "free software" people. 

> You sound more like the RIAA/MPAA ("we own all the rights! We _own_ your 
> sorry asses for even listening to our music") and a bunch of whiners that 
> think that just because you have touched a piece of hardware you 
> automatically can do anythign you want to it, and nobody elses rights 
> matter in the least!

> Guys, in fighting for "your rights", you should look a bit at *other* 
> peoples rights too. Including the rights of hw manufacturers, and the 
> service providers. Because this is all an eco-system, where in order to 
> actually succeed, you need to make _everybody_ succeed.

Good.  How about thinking of the users, the customers of your dear
friends too?  The ones who might be contributing much more to your
project.

Then look how what you said in the paragraph before about RIAA/MPAA
applies to what TiVO is doing to the software, and realize that you're
accusing us of doing what the party you support does.

I'm not trying to impose anything.  I'm not pushing anything.  I'm
defending the GPLv3 from accusations that it's departing from the GPL
spirit, and I'm trying to find out in what way Tivoization promotes
the goals you perceive as good for Linux, that make GPLv2
advantageous.  So far, you haven't given any single reason about this.
You talked about tit-for-tat, you said anti-Tivoization in GPLv3 was
bad, but you don't connect the dots.  Forgive if I get the impression
that you're just fooling yourself, and misguiding a *lot* of people
out there in the process.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Al Viro]

On Fri, Jun 15, 2007 at 01:14:49AM -0300, Alexandre Oliva wrote:
> I'm not trying to impose anything.  I'm not pushing anything.  I'm
> defending the GPLv3 from accusations that it's departing from the GPL
> spirit, and I'm trying to find out in what way Tivoization promotes
> the goals you perceive as good for Linux, that make GPLv2
> advantageous.  So far, you haven't given any single reason about this.
> You talked about tit-for-tat, you said anti-Tivoization in GPLv3 was
> bad, but you don't connect the dots.  Forgive if I get the impression
> that you're just fooling yourself, and misguiding a *lot* of people
> out there in the process.

Give.  Me.  A.  Break.

Section 6 is inherently broken.  It tries to gerrymander the "bad"
cases and ends up with a huge mess.  Definition of user device is
arbitrary and reeks with discrimination against the field of use.
Trying to be more explicit about installation instructions walks
straight into a minefield:
	* is it enough to give some installation methods?  If so,
should they be as cheap as the rest?  As efficient as the rest in
some sense?  Representative in some sense?  The same as what
manufacturer ever uses?
	* if all installation methods should be given, where does
one stop?  Should one describe unsupported ones?  All of them?
Is that a violation of license to omit some?  How does one prove
that omission hadn't been malicious violation in face of complaint?

Trying to be explicit enough to get a rope on the TiVo neck ends
up with not just clumsy rules; it opens a can of worms worse than
what we have in matching part of v2.

It looks like trying to be tight enough to be sure to catch the cases FSF
doesn't like and trying to avoid getting the stuff that really shouldn't
be caught.  And looks like these requirements conflict.

So in the end you get an ugliness that satisfies neither those who think
that TiVo case is not a problem nor those who agree that it is a problem
and consider v2 sufficient in that area.

And BTW, you've been told just that about an hour before you've sent that
mail.  I don't mind repeating that on l-k, but please don't pretend to be
unaware of the problems in that area.  I've no idea which problems Linus
has with that turd, but there's certainly enough in there.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 15, 2007, Al Viro <viro@ftp.linux.org.uk> wrote:

> On Fri, Jun 15, 2007 at 01:14:49AM -0300, Alexandre Oliva wrote:
>> I'm not trying to impose anything.  I'm not pushing anything.  I'm
>> defending the GPLv3 from accusations that it's departing from the GPL
>> spirit, and I'm trying to find out in what way Tivoization promotes
>> the goals you perceive as good for Linux, that make GPLv2
>> advantageous.  So far, you haven't given any single reason about this.
>> You talked about tit-for-tat, you said anti-Tivoization in GPLv3 was
>> bad, but you don't connect the dots.  Forgive if I get the impression
>> that you're just fooling yourself, and misguiding a *lot* of people
>> out there in the process.

> Give.  Me.  A.  Break.

> Section 6 is inherently broken.

You mean the bits against Tivoization in it, right.

You point out reasons you dislike the particular wording (good, can
you relay them to gplv3.fsf.org, please?), but nowhere do you show
where such provisions hamper the tit-for-tat goal that Linus likes
about GPLv2 and claims to be the reason he chose v2.  In fact, I've
shown evidence that anti-Tivoization increases this tit-for-tat.
Still, v2 is preferred over v3 under these grounds.  How can it be?

> And BTW, you've been told just that about an hour before you've sent that
> mail.

Yup.  I got your opinion.  Now I got it twice, and others got it
once.  How about others' opinions?

And, more importantly, how does that provision conflict with your
personal goals WRT Linux?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Al Viro]

On Fri, Jun 15, 2007 at 05:08:52PM -0300, Alexandre Oliva wrote:
> You point out reasons you dislike the particular wording (good, can
> you relay them to gplv3.fsf.org, please?), but nowhere do you show
> where such provisions hamper the tit-for-tat goal that Linus likes
> about GPLv2 and claims to be the reason he chose v2.  In fact, I've
> shown evidence that anti-Tivoization increases this tit-for-tat.
> Still, v2 is preferred over v3 under these grounds.  How can it be?

It's not about particular wording.  It's about disgust at that kind of
games in general.  Now bugger off and stop deliberately misparsing what
I said.

Oh, and could you please lose the "Free Software Evangelist" bit in your
sig when posting on l-k?  Unless you want to get the treatment normally
reserved for evangelist pests, that is...  Your personal kinks are your
personal kinks; kindly keep your religion to consenting partners.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Friday 15 June 2007 00:14:49 Alexandre Oliva wrote:
> On Jun 15, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:
> > On Fri, 15 Jun 2007, Alexandre Oliva wrote:
> >> case 2'': tivo provides source, end user tries to improve it, realizes
> >> the hardware won't let him use the result of his efforts, and gives up
> >
> > So you're blaming Tivo for the fact that your end user was a lazy bum and
> > wanted to take advantage of somebody elses hard work without permission?
>
> -ENONSEQUITUR
>
> > Quite frankly, I know who the bad guy in that scenario is, and it ain't
> > Tivo. It's your lazy bum, that thought he would just take what Tivo did,
> > sign the contract, and then not follow it. And just because the box
> > _contained_ some piece of free software, that lazy bum suddenly has all
> > those rights?
>
> Yes, because the software license that TiVo signed up for says that
> TiVo must pass on certain rights and not impose any further
> restrictions.

They don't add any restrictions that don't already exist. As stated in section 
0 of the GPLv2 the scope of the license is "copying, distributing and 
modification".

> And all that because TiVo wanted to use kernel and userland that were
> readily available, and at no cost other than respecting others'
> freedoms, while at that?
>
> Who's the lazy bum, again?

Still the same one that Linus pointed out. No amount of faulty logic can 
change that.

> > And you seem to argue that it's perfectly fine to ignore the people who
> > design hardware and the services around them,
>
> Just like they seem to think it's perfectly fine to ignore a number of
> people who design and maintain the software they decided to use in
> their hardware.

Huh? They have complied with the terms - and, IMHO, the spirit - of the 
license under which that software was released.

> > Guys, you should be ashamed of calling yourself "free software" people.
> >
> > You sound more like the RIAA/MPAA ("we own all the rights! We _own_ your
> > sorry asses for even listening to our music") and a bunch of whiners that
> > think that just because you have touched a piece of hardware you
> > automatically can do anythign you want to it, and nobody elses rights
> > matter in the least!
> >
> > Guys, in fighting for "your rights", you should look a bit at *other*
> > peoples rights too. Including the rights of hw manufacturers, and the
> > service providers. Because this is all an eco-system, where in order to
> > actually succeed, you need to make _everybody_ succeed.
>
> Good.  How about thinking of the users, the customers of your dear
> friends too?  The ones who might be contributing much more to your
> project.

If they are *CONTRIBUTING* then they are not just simple users anymore.

> Then look how what you said in the paragraph before about RIAA/MPAA
> applies to what TiVO is doing to the software, and realize that you're
> accusing us of doing what the party you support does.
>
> I'm not trying to impose anything.  I'm not pushing anything.  I'm
> defending the GPLv3 from accusations that it's departing from the GPL
> spirit, and I'm trying to find out in what way Tivoization promotes
> the goals you perceive as good for Linux, that make GPLv2
> advantageous.  So far, you haven't given any single reason about this.
> You talked about tit-for-tat, you said anti-Tivoization in GPLv3 was
> bad, but you don't connect the dots.  Forgive if I get the impression
> that you're just fooling yourself, and misguiding a *lot* of people
> out there in the process.

And failing. You say that the intent and spirit of the GPLv2 is clear if you 
read it. I read it and I feel its pretty clear that the only things that the 
GPLv2 sought to protect are *EXACTLY* "copying, distribution and 
modification". 

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 15, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> On Friday 15 June 2007 00:14:49 Alexandre Oliva wrote:
>> On Jun 15, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:
>> > Guys, in fighting for "your rights", you should look a bit at *other*
>> > peoples rights too. Including the rights of hw manufacturers, and the
>> > service providers. Because this is all an eco-system, where in order to
>> > actually succeed, you need to make _everybody_ succeed.

>> Good.  How about thinking of the users, the customers of your dear
>> friends too?  The ones who might be contributing much more to your
>> project.

> If they are *CONTRIBUTING* then they are not just simple users anymore.

Yeah.  Isn't that great?  Isn't this even more contributions "in
kind"?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Helge Hafting]

Linus Torvalds wrote:
> On Fri, 15 Jun 2007, Alexandre Oliva wrote:
>   
>> case 2'': tivo provides source, end user tries to improve it, realizes
>> the hardware won't let him use the result of his efforts, and gives up
>>     
>
> So you're blaming Tivo for the fact that your end user was a lazy bum and 
> wanted to take advantage of somebody elses hard work without permission?
>
> Quite frankly, I know who the bad guy in that scenario is, and it ain't 
> Tivo. It's your lazy bum, that thought he would just take what Tivo did, 
> sign the contract, and then not follow it. And just because the box 
> _contained_ some piece of free software, that lazy bum suddenly has all 
> those rights? Never mind all the *other* effort that went into bringing 
> that box to market?
>
> You do realize that Tivo makes all their money on the service, don't you? 
> The actual hardware they basically give away at cost, exactly to get the 
> service contracts. Not exactly a very unusual strategy in the high-tech 
> world, is it?
>   
Not unusual - but so what.
This strategy may backfire if users find a way
to use the cheap box for something without Tivo's service contract.

That don't make me feel sorry for Tivo - they can then sell their
boxes with some profit - or go bankrupt for doing stupid business.

Similiar to how unrealistic cheap printers backed by ink sales
fail when third parties undercuts the ridiculously expensive ink.
Or give-away cellphones tied to an expensive provider, being
unlocked by third parties so a switch to a cheaper provider will work.


Don't get me wrong - I have nothing against Tivo - and nothing for
them either.  Them making a locked device is just a fun arms
race to see who can reprogram the bootloader key
or find some other clever way to load software. . .

> Guys, in fighting for "your rights", you should look a bit at *other* 
> peoples rights too. Including the rights of hw manufacturers, and the 
> service providers. Because this is all an eco-system, where in order to 
> actually succeed, you need to make _everybody_ succeed.
>   
Nothing against a hw manufacturers right - but of course they
have no particular right to succeed with give-away hw and
expensive service.  When that strategy fail due to people operating
the hw without buing service then the cure is to charge properly
for the hw, not to ask everybody to please stop hacking.

Assuming that nobody can change the box after the sale is
their risk, and there is nothing unfair about failure here.

Helge Hafting

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Wed, 20 Jun 2007, Helge Hafting wrote:

> Linus Torvalds wrote:
> > 
> > You do realize that Tivo makes all their money on the service, don't you?
> > The actual hardware they basically give away at cost, exactly to get the
> > service contracts. Not exactly a very unusual strategy in the high-tech
> > world, is it?
> >   
> Not unusual - but so what.
> This strategy may backfire if users find a way
> to use the cheap box for something without Tivo's service contract.

Yeah, I'm not at all trying to say that Tivo has the "right to make 
money". 

They can fail for all I care. 

But they do have the right to make their own choices, and try their own 
strategies. And people shouldn't complain about that. If somebody doesn't 
like the Tivo box, and the Tivo service requirements, just don't *buy* the 
damn thing, and don't sign up for the service.

Whining about Tivo's choices is just stupid.

And anybody who thinks others don't have the "right to choice", and then 
tries to talk about "freedoms" is a damn hypocritical moron.

		Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Dave Neuer]

On 6/20/07, Linus Torvalds <torvalds@linux-foundation.org> wrote:
>
> But they do have the right to make their own choices, and try their own
> strategies. And people shouldn't complain about that. If somebody doesn't
> like the Tivo box, and the Tivo service requirements, just don't *buy* the
> damn thing, and don't sign up for the service.

>
> And anybody who thinks others don't have the "right to choice", and then
> tries to talk about "freedoms" is a damn hypocritical moron.

One might say the same thing about someone who claims not to have a
moral right to force certain choices on others in some circumstances
(e.g. when those others have used copyrighted work in a product and
ought to understand that for some not insignificant portion of the
copyright holders, the terms implicitly included preserving certain
"freedoms" for downstream recipients) while reserving a very similar
moral right with others (e.g. potential murderers, theives,
tresspassers, distributors of proprietary derived works).

As I pointed out in a previous message in the thread, that some people
want the copyright holders -- who have economic power over the
hardware vendors who use Linux in their products -- to force a
desireable outcome for those without that power (i.e. the small number
of people who buy TiVOs or Linksys routers who actually care about
source availability and who get told "don't buy it" as if that will
change anything) is understandable.

I think that there's a legitimate concern about what types of
constraints it's valid for a copyright holder to try to enforce w/ a
license -- I think it's immoral for an employer to force an employee
to toil at a meaningless, soul-crushing job for the vast majority of
one's single, short existence if they could make it more enjoyable,
but I'd hate to see someone try to enforce that with a license (I'm
happy telling a person in that situation to "just quit" just as you
tell people "just don't buy it"). To call people who draw the line in
a different place than you hypocrites is BS.

Your pragmatic arguments make much more sense than your moralistic ones, IMHO.

Dave

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alan Cox]

> constraints it's valid for a copyright holder to try to enforce w/ a
> license -- I think it's immoral for an employer to force an employee
> to toil at a meaningless, soul-crushing job for the vast majority of
> one's single, short existence if they could make it more enjoyable,
> but I'd hate to see someone try to enforce that with a license (I'm
> happy telling a person in that situation to "just quit" just as you
> tell people "just don't buy it"). To call people who draw the line in
> a different place than you hypocrites is BS.

Very poor example. In many parts of the world "Just quit" is "just starve
to death".

And guess what - brand owners do use the rights to try and stop that kind
of abuse with varying degrees of success. The 'fair trade' logo is
controlled this way. Many companies do their best not to license
production of goods to sweat shops and child slaves. It's considered
morally correct and companies get pounded by the public and lose business
when they fail to police these restrictions.

So please DON'T equate the two. Tivo is a minor control argument about a
silly little TV recording box. Employee rights is a matter of life, death
and slavery for many many people.

Alan

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Dave Neuer]

On 6/20/07, Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:
> > To call people who draw the line in
> > a different place than you hypocrites is BS.
>
> Very poor example. In many parts of the world "Just quit" is "just starve
> to death".

> So please DON'T equate the two. Tivo is a minor control argument about a
> silly little TV recording box. Employee rights is a matter of life, death
> and slavery for many many people.

Well, I consider your eloquent counter-argument to have strengthened
my point that such "rights to moral coercion" arguments are inherently
subjective and highly context-dependent.

Dave

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Wed, 20 Jun 2007, Dave Neuer wrote: 
> > 
> > And anybody who thinks others don't have the "right to choice", and then
> > tries to talk about "freedoms" is a damn hypocritical moron.
> 
> One might say the same thing about someone who claims not to have a
> moral right to force certain choices on others in some circumstances
> (e.g. when those others have used copyrighted work in a product and
> ought to understand that for some not insignificant portion of the
> copyright holders, the terms implicitly included preserving certain
> "freedoms" for downstream recipients) while reserving a very similar
> moral right with others (e.g. potential murderers, theives,
> tresspassers, distributors of proprietary derived works).

I don't disagree that "morals" are something very personal, and you can 
thus never really argue on morals *except*for*your*own*behaviour*.

So I claim that for *me* the right choice is GPLv2 (or something similar). 
I think the GPLv3 is overreaching.

There's a very fundamental, and very basic rule that is often a good 
guideline. It's "Do unto others".

So the reason I *personally* like the GPLv2 is that it does unto others 
exactly what I wish they would do unto me.

It allows everybody do make that choice that I consider to be really 
important: the choice of how something _you_ designed gets used.

And it does that exactly by *limiting* the license to only that one work. 
Not trying to extend it past the work.

See?

The GPLv3 can never do that. Quite fundamentally, whenever you extend the 
"reach" of a license past just the derived work, you will *always* get 
into a situation where people who designed two different things get into a 
conflict when they meet. The GPLv2 simply avoids the conflict entirely, 
and has no problem at all with the "Do unto others as you would have them 
do unto you".

In a very real sense, the GPLv3 asks people to do things that I personally 
would refuse to do. I put Linux on my kids computers, and I limit their 
ability to upgrade it. Do I have that legal right (I sure do, I'm their 
legal guardian), but the point is that this is not about "legality", this 
is about "morality". The GPLv3 doesn't match what I think is morally where 
I want to be. I think it *is* ok to control peoples hardware. I do it 
myself.

So your arguments about "potential murderes", "thieves", "trespassers" and 
"distributors of proprietary derived works" is totally missing the point.

It's missing the point that "morals" are about _personal_ choices. You 
cannot force others to a certain moral standpoint. 

Laws (like copyright law) and legal issues, on the other hand, are 
fundamentally *not* about "personal" things, they are about interactions 
that are *not* personal. So laws need to be fundamnetally different from 
morals. A law has to take into account that different people have 
different moral background, and a law has to be _pragmatic_.

So trying to mix up a moral argument with a legal one is a fundamental 
mistake. They have two totally different and separate areas.

The GPLv2 is a *legal* license. It's not a "moral license" or a "spiritual 
guide". Its raison-d'etre is that pragmatic area where different peoples 
different moral rules meet.

In contrast, a persons *choice* to use the GPLv2 is his private choice.  
Totally different. My choice of the GPLv2 doesn't say anything about my 
choice of laws or legal issues. 

You don't have to agree with it - but exactly because it's his private 
choice, it's a place where the persons moral rules matter, in a way that 
they do *not* matter in legal issues.

So killing, thieving, and distributing proprietary derived works are about 
*legal* choices. Are they also "immoral"? Who knows. Sometimes killing is 
moral. Sometimes thievery can me moral. Sometimes distributing derived 
works can be moral. Morality != legality. They are two totally different 
things.

Only religious fanatics and totalitarian states equate "morality" with 
"legality". There's tons of examples of that from human history. The ruler 
is not just a king, he's a God, so disagreeing with him is immoral, but 
it's also illegal, and you can get your head cut off.

In fact, a lot of our most well-known heroes are the ones that actually 
saw the difference between morals and laws.

A German soldier who refused to follow orders was clearly the more "moral" 
one, wouldn't you say? Never mind law. Gandhi is famous for his peaceful 
civil disobedience - was that "immoral" or "illegal"? 

Or Robin Hood. A romantic tale, but one where the big fundamnetal part of 
the picture is the _difference_ between morality and legality.

Think about it.

Yes, there is obviously overlap, in that a lot of laws are there to 
protect things that people also consider "moral". But the fact that there 
is correlation should *not* cause anybody to think that they are at all 
about the same thing. 

> To call people who draw the line in a different place than you 
> hypocrites is BS.

That was *not* what I did.

I don't think it's hypocritical to prefer the GPLv3. That's a fine choice, 
it's just not *mine*.

What I called hypocritical was to do so in the name of "freedom", while 
you're at the same time trying to argue that I don't have the "freedom" to 
make my own choice.

See? THAT is hypocritical.

			Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Dave Neuer]

On 6/20/07, Linus Torvalds <torvalds@linux-foundation.org> wrote:
>
> On Wed, 20 Jun 2007, Dave Neuer wrote:
> > >
> > > And anybody who thinks others don't have the "right to choice", and then
> > > tries to talk about "freedoms" is a damn hypocritical moron.
> >
> > One might say the same thing about someone who claims not to have a
> > moral right to force certain choices on others in some circumstances
> > (e.g. when those others have used copyrighted work in a product and
> > ought to understand that for some not insignificant portion of the
> > copyright holders, the terms implicitly included preserving certain
> > "freedoms" for downstream recipients) while reserving a very similar
> > moral right with others (e.g. potential murderers, theives,
> > tresspassers, distributors of proprietary derived works).
>
> I don't disagree that "morals" are something very personal, and you can
> thus never really argue on morals *except*for*your*own*behaviour*.
>
> So I claim that for *me* the right choice is GPLv2 (or something similar).
> I think the GPLv3 is overreaching.
>
> There's a very fundamental, and very basic rule that is often a good
> guideline. It's "Do unto others".
>
> So the reason I *personally* like the GPLv2 is that it does unto others
> exactly what I wish they would do unto me.
>
> It allows everybody do make that choice that I consider to be really
> important: the choice of how something _you_ designed gets used.
>
> And it does that exactly by *limiting* the license to only that one work.
> Not trying to extend it past the work.
>
> See?
>
> The GPLv3 can never do that. Quite fundamentally, whenever you extend the
> "reach" of a license past just the derived work, you will *always* get
> into a situation where people who designed two different things get into a
> conflict when they meet. The GPLv2 simply avoids the conflict entirely,
> and has no problem at all with the "Do unto others as you would have them
> do unto you".
>
> In a very real sense, the GPLv3 asks people to do things that I personally
> would refuse to do. I put Linux on my kids computers, and I limit their
> ability to upgrade it. Do I have that legal right (I sure do, I'm their
> legal guardian), but the point is that this is not about "legality", this
> is about "morality". The GPLv3 doesn't match what I think is morally where
> I want to be. I think it *is* ok to control peoples hardware. I do it
> myself.
>
> So your arguments about "potential murderes", "thieves", "trespassers" and
> "distributors of proprietary derived works" is totally missing the point.
>
> It's missing the point that "morals" are about _personal_ choices. You
> cannot force others to a certain moral standpoint.
>
> Laws (like copyright law) and legal issues, on the other hand, are
> fundamentally *not* about "personal" things, they are about interactions
> that are *not* personal. So laws need to be fundamnetally different from
> morals. A law has to take into account that different people have
> different moral background, and a law has to be _pragmatic_.
>
> So trying to mix up a moral argument with a legal one is a fundamental
> mistake. They have two totally different and separate areas.
>
> The GPLv2 is a *legal* license. It's not a "moral license" or a "spiritual
> guide". Its raison-d'etre is that pragmatic area where different peoples
> different moral rules meet.
>
> In contrast, a persons *choice* to use the GPLv2 is his private choice.
> Totally different. My choice of the GPLv2 doesn't say anything about my
> choice of laws or legal issues.
>
> You don't have to agree with it - but exactly because it's his private
> choice, it's a place where the persons moral rules matter, in a way that
> they do *not* matter in legal issues.
>
> So killing, thieving, and distributing proprietary derived works are about
> *legal* choices. Are they also "immoral"? Who knows. Sometimes killing is
> moral. Sometimes thievery can me moral. Sometimes distributing derived
> works can be moral. Morality != legality. They are two totally different
> things.
>
> Only religious fanatics and totalitarian states equate "morality" with
> "legality". There's tons of examples of that from human history. The ruler
> is not just a king, he's a God, so disagreeing with him is immoral, but
> it's also illegal, and you can get your head cut off.
>
> In fact, a lot of our most well-known heroes are the ones that actually
> saw the difference between morals and laws.
>
> A German soldier who refused to follow orders was clearly the more "moral"
> one, wouldn't you say? Never mind law. Gandhi is famous for his peaceful
> civil disobedience - was that "immoral" or "illegal"?
>
> Or Robin Hood. A romantic tale, but one where the big fundamnetal part of
> the picture is the _difference_ between morality and legality.
>
> Think about it.

Oh, I have, professor.

>
> Yes, there is obviously overlap, in that a lot of laws are there to
> protect things that people also consider "moral". But the fact that there
> is correlation should *not* cause anybody to think that they are at all
> about the same thing.

No, they overlap in cases where the reason for the law is to enforce
some moral edict, and when they're not in sync it's often because some
person or group of people casuistically draw a moral distinction which
is absent in the law (likely because it conflicts with some other
group's moral values).

Your example of Robin Hood is instructive. Rich people quite likely
don't agree with you that taking from the rich to give to the poor is
moral, and they get more say (in a culturally-dependent fashion) when
it comes time to codify the morals into laws.

>
> > To call people who draw the line in a different place than you
> > hypocrites is BS.
>
> That was *not* what I did.

It is.

>
> I don't think it's hypocritical to prefer the GPLv3. That's a fine choice,
> it's just not *mine*.

I understand your oft-repeated (in this message alone) preference,
which revolves around your personal "bright line" boundary of moral
coercion at "I designed it."

>
> What I called hypocritical was to do so in the name of "freedom", while
> you're at the same time trying to argue that I don't have the "freedom" to
> make my own choice.
>
> See? THAT is hypocritical.

It is no more hypocritical than any other instance of restricting one
person's freedom to protect some freedom of some other person. The law
restricts your freedom to take things from my house in order to
preserve my freedom to own things. That you don't disagree with that
law indicates that you agree with the majority of people about where
to draw the line between one person's freedom and another in that
instance, not that you are a hypocrite or that people who steal are
hypocrites (they're hypocrites if they complain about people stealing
from them).

You have every right to prefer using your copyright to coerce people
to provide source but not keys used to make binaries from that source
run on hardware with which those binaries are distributed.

But people who prefer the GPLv3 in the name of freedom simply favor
making explicit what they felt was always implicit in earlier
versions, which is their wish that people who want to use what "they
designed" do so in ways which comport with freedoms which are more
important to them than the freedom of the hardware developer to design
their system unconstrained (like the freedom of the end user to make
their TiVO work some other way). They are no more hypocrites than you
are for apparently feeling that it's OK for Robin Hood to steal from
the rich to feed the poor.

>
>                         Linus
>

Dave

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 20, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:

> It allows everybody do make that choice that I consider to be really 
> important: the choice of how something _you_ designed gets used.

> And it does that exactly by *limiting* the license to only that one work. 
> Not trying to extend it past the work.

Actually, the two paragraphs above are contradictory, and the second
is not true in as much as it gives way to the former.

Consider an independent file contributed to Linux.  It is an
independent work.  But the moment it gets combined with Linux, the
only license you can use is GPLv2.

Consider a patch, or any modified version of Linux.  It's another
work.  But the license applies to it as well.

Similarly, the GPL affects patents that somehow cover the work: the
distributor can no longer enforce them against downstream users of the
software.

See?

> The GPLv3 can never do that.

It does it in just the same way that GPLv1 and GPLv2 do: "no further
restrictions".  That it has to make some of them explicit, such that
people understand they apply, or such that this provision can't be
trampled on by other laws that by default trample copyright law, is
just a legal implementation detail.

> It's missing the point that "morals" are about _personal_ choices. You 
> cannot force others to a certain moral standpoint. 

FWIW, I tend to consider morals more of a society issue than an
individual issue.  Morals encode what the society understands to be
the common good, and that's often something evolutionary, even with
biological roots.

Laws (as you say) try to reflect the morals of the society, but since
it's based on morals, it often lags behind.  Which is why some laws
become forgotten and no longer applied, even if still applicable in
theory.  And that's also why (as you alluded to in your message), when
laws actively diverge from morals, you observe a lot of civil
disobedience: think DRM, DMCA, non-benevolent dictatorships and other
abusive regimes.


I must say that it *is* lovely to watch you talk about morals in ways
that I agree so much with, even if I dissent in a some details.  This
has further increased my admiration for you.  Thank you.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 20, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:

> And anybody who thinks others don't have the "right to choice", and then 
> tries to talk about "freedoms" is a damn hypocritical moron.

Yeah, it is indeed possible to twist it such that it sounds bad.

The important point is that one's freedom ends where another's
starts.  Unlimited freedom would be freedom to trample over others'
freedoms too, and that's not right.  That's why freedom of choice has
to be used with care.  Even fundamental human rights sometimes clash
with each other.

We don't fight for the freedoms as goals in themselves.  We fight for
them because we understand they're essential for the common good.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Thursday 14 June 2007 23:22:48 Alexandre Oliva wrote:
> On Jun 14, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> > Faulty logic. The hardware doesn't *restrict* you from *MODIFYING*
> > any fscking thing.
>
> Ok, lemme try again:
>
> case 2'': tivo provides source, end user tries to improve it, realizes
> the hardware won't let him use the result of his efforts, and gives up

And there is nothing in the license that says that this has to be done. 
Claiming that it is a requirement because of the "spirit" of the license or 
that such was the "intent" of the license does not make it any less legal 
than it is. And, as I've taken the time to explain to you, lacking any clear 
statement, written at the exact same time as the license, a statement of 
intent or spirit cannot have any real legal weight when the text of a license 
is finally decided upon. The reason, in case you missed the mail in which I 
gave it, is that the author *cannot*, no matter the belief anyone may have in 
their honesty or the oaths they may swear, be trusted to have *not* changed 
his/her mind as to the intent and/or spirit of the license at any time after 
the license goes into use.

DRH

> > On Thursday 14 June 2007 18:45:07 Alexandre Oliva wrote:
> >> Where's the payback, or the payforward?
> >>
> >> And then, tit-for-tat is about equivalent retaliation, an eye for an
> >> eye.  Where's the retaliation here?
> >>
> >> If GPLv2 were tit-for-tat, if someone invents artifices to prevent the
> >> user from making the changes the user wants on the software, wouldn't
> >> it be "equivalent retaliation" to prevent the perpetrator from making
> >> the changes it wants on the software?



-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Jesper Juhl]

On 15/06/07, Alexandre Oliva <aoliva@redhat.com> wrote:
> On Jun 14, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
>
> > Faulty logic. The hardware doesn't *restrict* you from *MODIFYING*
> > any fscking thing.
>
> Ok, lemme try again:
>
> case 2'': tivo provides source, end user tries to improve it, realizes
> the hardware won't let him use the result of his efforts, and gives up
>
So?  The user still has the source and is free to use that in other
GPLv2 projects, that's the point. The point is not being able to run
it on any specific hardware, the point is having the source with the
same rights to modify it and distribute it.   And no, the right to
modify your copy of the source does not also mean you *have to* be
able to install it on the hardware it was originally designed for - it
only means you have the right to modify it and redistribute it.

-- 
Jesper Juhl <jesper.juhl@gmail.com>
Don't top-post  http://www.catb.org/~esr/jargon/html/T/top-post.html
Plain text mails only, please      http://www.expita.com/nomime.html

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 15, 2007, "Jesper Juhl" <jesper.juhl@gmail.com> wrote:

> On 15/06/07, Alexandre Oliva <aoliva@redhat.com> wrote:
>> On Jun 14, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
>> 
>> > Faulty logic. The hardware doesn't *restrict* you from *MODIFYING*
>> > any fscking thing.

>> case 2'': tivo provides source, end user tries to improve it, realizes
>> the hardware won't let him use the result of his efforts, and gives up

> So?  The user still has the source and is free to use that in other
> GPLv2 projects, that's the point.

This point of yours is a distraction from the argument in this
sub-thread.


These cases were Chris Friesen's attempt to show that GPLv2 was
tit-for-tat, and case 2'' shows it isn't, at least not in the sense he
tried to picture it:


On Jun 14, 2007, "Chris Friesen" <cfriesen@nortel.com> wrote:

> Alexandre Oliva wrote:

>> That's where Linus' theory of tit-for-tat falls apart.

> Nope.

> case 1:  Upstream provides source, tivo modifies and distributes it
> (to their customers).

> case 2: tivo provides source, end user modifies and distributes it
> (possibly to their customers, maybe to friends, possibly even to
> upstream).

> See?  Tit for tat.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Thursday 14 June 2007 17:27:27 Alexandre Oliva wrote:
> On Jun 14, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> > <somewhat sarcastic>
> > And the companies that produce devices that come with Linux and/or
> > other GPL'd software installed and place limits such that only
> > people that have purchased that hardware have access to the
> > "modified" source running on the device are following the letter,
> > and the spirit, of the GPL.
>
> WAIT, WAIT, THAT'S... :-)
>
> > Before you start yelling I'm wrong, think about it this way: they
> > make the source available to the people that they've given binary
> > versions to, and there is nothing stopping one of those people from
> > making the source available to the rest of the world.
>
> The *only* in your sentence betrayed you.
>
> If they place the limits such that nobody else can access the sources,
> they're in violation of the license.

Nope. There is *NO* requirement *ANYWHERE* in the GPL, no matter the version, 
that says you have to *DISTRIBUTE* the source to *ANYONE* except those that 
you have given a binary to. Go read the licenses.

>
> If they merely refrain from distributing the sources to others, but
> still enable the recipients to do so, this is not a violation of the
> license.

Exactly what I said. "only the people that have purchased the hardware have 
access to the modified sources"

That is *EXACTLY* what a number of companies have done - Acer (yes, the laptop 
company) has done that. They sell laptops running Linux, but unless you have 
purchased one of them you can't download the sources (or even replacement 
binaries) for the version of linux they put on their machines. (From Acer, 
that is)

However, as I also said, there is nothing stopping one of those people from 
making those "modified sources" available to the rest of the world. (I have 
yet to find someone that has done that with the Acer specific stuff, but...) 

> But then IANAL.
>
> > *AND* the GPL has never been about making the source available to
> > everyone - just to those that get the binaries.
>
> Exactly.  Not even to the upstream distributor.  That's where Linus'
> theory of tit-for-tat falls apart.

Yes, it does. However, the practicality is that there is nothing *stopping* 
the person upstream from getting a copy of the source and incorporating the 
modifications they contain in a new version.

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> On Thursday 14 June 2007 17:27:27 Alexandre Oliva wrote:
>> On Jun 14, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
>> > <somewhat sarcastic>
>> > And the companies that produce devices that come with Linux and/or
>> > other GPL'd software installed and place limits such that only
>> > people that have purchased that hardware have access to the
>> > "modified" source running on the device are following the letter,
>> > and the spirit, of the GPL.
>> 
>> WAIT, WAIT, THAT'S... :-)
>> 
>> > Before you start yelling I'm wrong, think about it this way: they
>> > make the source available to the people that they've given binary
>> > versions to, and there is nothing stopping one of those people from
>> > making the source available to the rest of the world.
>> 
>> The *only* in your sentence betrayed you.
>> 
>> If they place the limits such that nobody else can access the sources,
>> they're in violation of the license.

> Nope. There is *NO* requirement *ANYWHERE* in the GPL, no matter the version, 
> that says you have to *DISTRIBUTE* the source to *ANYONE* except those that 
> you have given a binary to. Go read the licenses.

I agree.  I even said so.

But the *only* gave me the impression that you were talking about
magic, or any other sufficiently advanced technology ;-), that would
enable the recipients to get the source code, but not usefully pass it
on.

> That is *EXACTLY* what a number of companies have done - Acer (yes,
> the laptop company) has done that. They sell laptops running Linux,
> but unless you have purchased one of them you can't download the
> sources (or even replacement binaries) for the version of linux they
> put on their machines. (From Acer, that is)

That's the sort of stuff that breaks the tit-for-tat premise.  GPL
indeed is not concerned about tit-for-tat.  It's concerned about
respect for the freedoms.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Paul Mundt]

On Thu, Jun 14, 2007 at 01:57:26PM -0300, Alexandre Oliva wrote:
> On Jun 14, 2007, Paul Mundt <lethal@linux-sh.org> wrote:
> > I don't see how you can claim that the vendor is infringing on your
> > freedom, _you_ made the decision to go out and buy the product knowing
> > that the vendor wasn't going to go out of their way to help you hack
> > the device.
> 
> But I also made this decision fully aware that the software included
> in the package was published under a license that said I was entitled
> to modify it.

And you certainly are free to do so. The vendor ships the product with
the binaries, and you get the source as a result. You can in turn modify
that source and do whatever you like with it. If the vendor is more
proactive, they may have even tried to get all of their changes merged by
the time the product hit the market, so they wouldn't be sitting on
anything "special" anyways.

This however has nothing to do with your ability to apply those changes
to the _hardware_. If the vendor doesn't want to, or is unable to support
third-party modifications on their product, they have the basic right to
make that decision, as you have the basic right not to buy the product if
this is something that's going to be a problem for you.

> More than once I purchased a device that claimed to have GNU/Linux
> software on it, only to find out that I couldn't use the freedoms,
> because the distributor was infringing the license in various ways.
> 
In this example, at _no time_ did the vendor infringe on the license.
They haven't given you an easy way to change the hardware, but they're
completely compliant both in terms of the letter and the spirit
(depending on how they work with the community) of the license.

If you're trying to pretend that GPLv2 had _anything_ to say about
hardware, you'd be wrong. In such a situation, there'd hardly be a "need"
(as you seem to see it) for GPLv3 at all. If you think this bizarre
coupling of the hardware/software paradigm is in any way constructive,
you're of course welcome to use the GPLv3, but this does not
retroactively change the terms of the GPLv2 simply because you saw this
as an area that was apparently "lacking".

And on the other hand, you're more than welcome to dual-license all of
your kernel changes under v2/v3 if you really feel that that's the best
way to go, just as I'm welcome to print out and burn the GPLv3 as a
symbolic gesture. Simply because some folks have no intention of ever
supporting v3 doesn't stop you from using it on any of your own changes.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Krzysztof Halasa]

Alexandre Oliva <aoliva@redhat.com> writes:

> BTW, what license is Linux licensed under?  It's GPLv2 plus userland
> exception, right?  (There's some additional module exception, right?)

Pure GPLv2.

Userland exception? Never heard of.
Module exception? Perhaps you mean "interpretation"?
-- 
Krzysztof Halasa

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Krzysztof Halasa]

Daniel Hazelton <dhazelton@enter.net> writes:

> Exactly. And I don't see anything about a TiVO (or any device that, like a 
> TiVO, requires binaries that run on it to be digitally signed) that stops
> you 
> from exercising the "freedoms" guaranteed by the GPL. As I said before, what 
> it does is stop you from violating the license on the hardware.

BTW: don't they sell their hardware (as well)? I think it should be
easy to replace the ROMs (EPROMs? flash ROMs?) using some diagnostic
clip and/or JTAG. Unless the CPU itself verifies ROM signatures,
they shouldn't matter.
-- 
Krzysztof Halasa

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bernd Petrovitsch]

On Wed, 2007-06-13 at 23:38 -0300, Alexandre Oliva wrote:
> On Jun 13, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> > On Wednesday 13 June 2007 19:49:23 Alexandre Oliva wrote:
> 
> > Exactly. They don't. What TiVO prevents is using that modified version on 
> > their hardware. And they have that right, because the Hardware *ISN'T* 
       ^^^^^^^^^^^^^^
BTW as soon as I bought that thing, it is *my* hardware and no longer
*theirs* (whoever "theirs" was).

> > covered by the GPL.
> 
> Indeed, TiVO has this legal right.  But then they must not use

Do they? At least in .at, it is usually impossible to (legally) limit
the rights of the *owner* a (tangible) thing (and if I bought it, I *am*
the owner and no one else) - even if you put it in the sales contract
since this is discussion about/within sales law.

One usual example is "you buy a car and neither the car producer nor the
(re)seller can restrict the brands of the tires you may use or the brand
of the fuel etc.".

And the same holds for pretty much everything. No one can forbid you to
open a TV set and fix it (or let it fix by whoever I choose to).

Yes, there are exceptions in several laws for specific things (e.g. for
really dangerous ones like airbags in cars) but in general, you are
allowed to do almost anything (including the simple destruction of it).

And yes, if you *rent* the thing, you are not the owner and this is a
totally different thing.

> software under the GPLv3 in it.  And, arguably, they must not use
> software under the GPLv2 either.

	Bernd
-- 
Firmix Software GmbH                   http://www.firmix.at/
mobil: +43 664 4416156                 fax: +43 1 7890849-55
          Embedded Linux Development and Services

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Thursday 14 June 2007 04:37:55 Bernd Petrovitsch wrote:
> On Wed, 2007-06-13 at 23:38 -0300, Alexandre Oliva wrote:
> > On Jun 13, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> > > On Wednesday 13 June 2007 19:49:23 Alexandre Oliva wrote:
> > >
> > > Exactly. They don't. What TiVO prevents is using that modified version
> > > on their hardware. And they have that right, because the Hardware
> > > *ISN'T*
>
>        ^^^^^^^^^^^^^^
> BTW as soon as I bought that thing, it is *my* hardware and no longer
> *theirs* (whoever "theirs" was).

eh. Perhaps I should have said that differently. And TiVO could handle it 
differently. I'm not going to argue about it anymore. It's pointless.

> > > covered by the GPL.
> >
> > Indeed, TiVO has this legal right.  But then they must not use
>
> Do they? At least in .at, it is usually impossible to (legally) limit
> the rights of the *owner* a (tangible) thing (and if I bought it, I *am*
> the owner and no one else) - even if you put it in the sales contract
> since this is discussion about/within sales law.
>
> One usual example is "you buy a car and neither the car producer nor the
> (re)seller can restrict the brands of the tires you may use or the brand
> of the fuel etc.".

No argument there. However, that is not to say that "you bought it, now you're 
free to do with it whatever you please" is always what the law says (at least 
in the US)

In the TiVO case there may be restrictions placed on the manufacturer for 
legal reasons or contractual reasons. Seeing as I'm not privy to the 
contracts between TiVO and the various production and broadcasting companies 
I can't comment on what contracts they have. As to the legal side there are 
restrictions in copyright law.

> And the same holds for pretty much everything. No one can forbid you to
> open a TV set and fix it (or let it fix by whoever I choose to).

I know of at least one company that will sell you the parts to repair your TV 
if its out of warranty.

DRH

> Yes, there are exceptions in several laws for specific things (e.g. for
> really dangerous ones like airbags in cars) but in general, you are
> allowed to do almost anything (including the simple destruction of it).
>
> And yes, if you *rent* the thing, you are not the owner and this is a
> totally different thing.
>
> > software under the GPLv3 in it.  And, arguably, they must not use
> > software under the GPLv2 either.
>
> 	Bernd



-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bernd Petrovitsch]

On Thu, 2007-06-14 at 05:05 -0400, Daniel Hazelton wrote:
> On Thursday 14 June 2007 04:37:55 Bernd Petrovitsch wrote:
[...]
> > > > covered by the GPL.
> > >
> > > Indeed, TiVO has this legal right.  But then they must not use
> >
> > Do they? At least in .at, it is usually impossible to (legally) limit
> > the rights of the *owner* a (tangible) thing (and if I bought it, I *am*
> > the owner and no one else) - even if you put it in the sales contract
> > since this is discussion about/within sales law.
> >
> > One usual example is "you buy a car and neither the car producer nor the
> > (re)seller can restrict the brands of the tires you may use or the brand
> > of the fuel etc.".
> 
> No argument there. However, that is not to say that "you bought it, now you're 
> free to do with it whatever you please" is always what the law says (at least 
> in the US)

Of course not (and I neither stated nor implied it) - there are lots of
laws which forbid killing other people etc.
But the seller of the car is not in the position to forbid anything
(which is not forbidden by the law), e.g. ha cannot forbid to replace
the motor or similar thing. I may loose guarantee or have to cope with
other consequences (if it is done badly),  but that is my problem and
decision.

> In the TiVO case there may be restrictions placed on the manufacturer for 
> legal reasons or contractual reasons. Seeing as I'm not privy to the 

Frankly, I really don't care that much about legal and contractual
reasons of the *manufacturer* (starting from waste disposal regulations
up to tax regulations, etc.) and they are irrelevant to me anyways.
At most I can have
*) legal restrictions (obviously coming from the law) on the *usage* of
     the device or
*) from a contract (obviously with the seller of the device since there
    is no other involved - and this contract may contain inapplicable
    clauses - e.g. sth. like "you are not allowed hear German music with
    this device").
And I don't have a contract with the manufacturer so there can't be any
limitation by the manufacturer.

> contracts between TiVO and the various production and broadcasting companies 
> I can't comment on what contracts they have. As to the legal side there are 

And they are pretty irrelevant anyways to everyone else.

> restrictions in copyright law.

ACK. But copyright law (at least the equivalent in .at and very
probably .de - and IMHO it is probably everywhere else similar simply
because copyright/authors rights laws was actually designed and written
to deal with music, literature, etc. which are intangible by nature)
simply doesn't apply to hardware as such (pun intended;-).

	Bernd
-- 
Firmix Software GmbH                   http://www.firmix.at/
mobil: +43 664 4416156                 fax: +43 1 7890849-55
          Embedded Linux Development and Services

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Thu, 14 Jun 2007, Bernd Petrovitsch wrote:

> On Wed, 2007-06-13 at 23:38 -0300, Alexandre Oliva wrote:
> > On Jun 13, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> > > On Wednesday 13 June 2007 19:49:23 Alexandre Oliva wrote:
> > 
> > > Exactly. They don't. What TiVO prevents is using that modified version on 
> > > their hardware. And they have that right, because the Hardware *ISN'T* 
>        ^^^^^^^^^^^^^^
> BTW as soon as I bought that thing, it is *my* hardware and no longer
> *theirs* (whoever "theirs" was).

You bought *their* design. It was your choice. 

And yes, you own the hardware, and you can hack it any which way you like 
(modulo laws and any other contracts you signed when you bought it). But 
they had the right to design it certain ways, and part of that design may 
be making it _harder_ for you to hack.

For example, they may have used glue to put the thing together rather than 
standard phillips screws. Or poured resin over some of the chips. All of 
which has been done (not necessarily with Linux, but this really is an 
issue that has nothing to do with Linux per se). Making the firmware or 
hardware harder to access or modify is their choice.

Your choice is whether you buy it, despite the fact that you know it's not 
necessarily all that easy to hack.

> > Indeed, TiVO has this legal right.  But then they must not use
> 
> Do they? At least in .at, it is usually impossible to (legally) limit
> the rights of the *owner* a (tangible) thing (and if I bought it, I *am*
> the owner and no one else) - even if you put it in the sales contract
> since this is discussion about/within sales law.

The "when I buy it, I own it" argument is a favourite of the GPLv3 shills, 
but it's irrelevant. The *design* was done long before you bought it, and 
yes, Tivo had the right to design and build it, any which way they wanted 
to. 

> One usual example is "you buy a car and neither the car producer nor the
> (re)seller can restrict the brands of the tires you may use or the brand
> of the fuel etc.".
> 
> And the same holds for pretty much everything. No one can forbid you to
> open a TV set and fix it (or let it fix by whoever I choose to).

You are missing the picture. Sure, you can do whatever you want to (within 
any applicable laws) _after_ you bought it. But that doesn't take away the 
right from the manufacturer to design it his way.

And you're also *wrong*. Tivo doesn't limit the brands of electricity it 
uses or anything idiotic like that. You can put after-market rubber bumps 
on the thing to make it look sleeker, and I seriously doubt that Tivo will 
do aythign at all. It's about going into the innards, and different car 
manufacturers make that harder too, for various reasons.

If the car manufacturer makes things harder to hack, it's your choice. For 
example, car hackers *do* actually prefer certain brands. Apparently the 
Subaru's are popular, and German cars are a pain to try to change. I'm 
told that even somethign as simple as upgrading the sound system is just 
_harder_ in a German car, apparently because they make things fit together 
so tightly, that doing after-market cabling is just much more of a 
problem.

Same goes for things like electronic engine controls. Look it up. Try 
chipping a car lately? For some, it's literally buying a chip online, and 
some fairly simple work. For others, it's almost impossible, and you have 
to take your car in to somebody who really knows what he's doing. And you 
know what? Exactly like with a Tivo, the car manufacturer won't have 
anything to do with the car afterwards. If you broke it by chipping it, 
you voided your warranty.

See? If you are actually looking for a car to hack on, you'd buy a car 
with that in mind. Do the exact same thing with your Tivo. Don't buy it if 
you want to hack it: buy a Neuros OSD device instead!  I'm serious: the 
Neuros people do *not* limit you, and in fact they encourage hacking. 
Instead of whining about Tivo, do something *positive*, and support Neuros 
for their better policies!

				Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Kevin Fox]

On Wed, 2007-06-13 at 20:42 -0400, Daniel Hazelton wrote:
<SNIP>
> >
> > Do you deny that TiVo prevents you (or at least a random customer)
> > from modifying the copy of Linux that they ship in their DVR?
> 
> Exactly. They don't. What TiVO prevents is using that modified version on 
> their hardware. And they have that right, because the Hardware *ISN'T* 
> covered by the GPL.

The hardware isn't directly covered by the GPL, correct. But, if they
want to use the software on the hardware, they have to comply with the
GPL. The software license can then influence hardware IF they want to
use it badly enough.

For example, the hardware is perfectly capable of being used to break
the terms of the GPL by being used to distribute a modified binary
without releasing the source. But the hardware's behavior is restricted
by the software for the betterment of all.

This whole argument is about the spirit of the GPL. Linus and others
think the spirit is one thing, the FSF guys think its something else.
Since the license is clearly owned by the FSF, I think they get the
final vote on what they "intended" it to be when they wrote it, no? If
they say they intended it to not allow Tivoization then believe them,
because they are the only ones that know what they were thinking when
they wrote it! The GPLv2 seems to allow it though. If Linus and friends
want to allow it, then they can stay with the GPLv2. For those who want
to disallow Tivoization, choose v3. No worries guys.

> Do you understand that, or do I need to break out the finger-puppets next ?

Guys, we are all friends here. No reason to be so insulting. Its just a
difference of opinion. People seem to be talking past each other instead
of to one another. This usually happens when people are basing their
underlying assumptions on different things and not listening to the
other. Please take a step back and think about it.

<SNIP>

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Thu, 14 Jun 2007, Kevin Fox wrote:
> 
> The hardware isn't directly covered by the GPL, correct. But, if they
> want to use the software on the hardware, they have to comply with the
> GPL.

Only with the GPLv3.

Again, don't confuse the *new* requirements in the GPLv3 with any "GPL 
requirements". They didn't exist before. The kernel never signed up to 
them. They are irrelevant for the discussion.

So hardware details have *nothing* to do with compying with the GPLv2.

Could you write *another* license that puts limitations on the hardware or 
environment that you have to comply with? Sure can. And the GPLv3 does 
that. But the GPLv2 does not, and that's a fundmanetal *improvement* over 
the GPLv3 in my opinion.

Do you like licenses that force the licensee to give money back?

So why do you like licenses that force the licensee to give access to 
hardware back? It's a form of "extra compensation" that the GPLv2 never 
had. The GPLv2 talks about giving access to the *source* code. The GPLv3 
talks about giving access to the *hardware*. 

Can people really not see the difference, and why I might think it's a 
fundamental difference, and why I might choose to say that the GPLv3 is a 
worse license?

And *why* would I ever downgrade to a worse license? There had better be 
some really pressing reason to choose the worse version of the GPL. And I 
just don't find that reason in the GPLv3 itself - although, as mentioned, 
the reason could become *external* (ie I might accept a worse license it 
it comes with external code attached to it that I think makes up for the 
license deficiency).

			Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:

> On Thu, 14 Jun 2007, Kevin Fox wrote:
>> 
>> The hardware isn't directly covered by the GPL, correct. But, if they
>> want to use the software on the hardware, they have to comply with the
>> GPL.

> Only with the GPLv3.

This is not true.  The terms of the GPLv2 that say you can't impose
further restrictions on the exercise of the freedoms apply to the
software under GPLv2 and GPLv3 just the same way.

> Do you like licenses that force the licensee to give money back?

> So why do you like licenses that force the licensee to give access to 
> hardware back?

I don't know where the 'back' in the second question amounts to, but
it definitely isn't about GPLv3.

In fact, the GPL isn't about giving anything back.  It's about passing
on.

So both requirements, as you phrased them, would be equally wrong.

So let's change the question to turn them into forms of passing on:

  Do you like licenses that force the licensee to pass money on?

  Do you like licenses that force the licensee to pass on access to
  hardware?

This is still bad.  This is still not what the GPLv3 is about.
There's no requirement to let the user go wild and do whatever she
likes on the hardware.

The only requirement is the one that was always there: to respect the
freedoms of the users of the software, i.e., let them modify and share
the software, not imposing any further restrictions, by whatever
means.

So the second question would be correctly phrased as 

  Do you like licenses that force the licensee to pass on the right to
  modify the software in the hardware containing it?

Or, reframing it:

  Do you like licenses that permit the licensee to deny others the
  right to modify the software in the hardware containing it?

> It's a form of "extra compensation" that the GPLv2 never had.

No, sir, it's still respect for the freedoms.  The same "in kind"
contribution as always.

> The GPLv2 talks about giving access to the *source* code.

It does.  But that's not all.  Even GPLv1 went further than that.

> Can people really not see the difference, and why I might think it's a 
> fundamental difference, and why I might choose to say that the GPLv3 is a 
> worse license?

Since someone brought liberal (Original BSD, Modified BSD, MIT, etc)
licenses into the picture, and you expressed dislike for them, let me
pick that up for a moment.

> The license doesn't encode my fundamental beliefs of "fairness".  I
> think the BSD license encourages a "everybody for himself"
> mentality, and doesn't encourage people to work together, and to
> merge.

And then you say what TiVO does is ok, saying:

> Oh, but you want to hack the hardware to accept it? That's a totally
> different issue. If so, buy a Neuros OSD box.

Sounds a lot like the very "everybody for himself" attitude you
dislike.

So can you please explain to me how enabling TiVO to deny others the
freedom that it received "in kind", failing to keep with the "in kind"
spirit of the GPL, encourage people to work together, and to merge?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Chris Friesen]

Alexandre Oliva wrote:

> So can you please explain to me how enabling TiVO to deny others the
> freedom that it received "in kind", failing to keep with the "in kind"
> spirit of the GPL, encourage people to work together, and to merge?

They're not denying others the freedom that they themselves received. 
Tivo took GPL'd software, modified it, and distributed it with their own 
custom hardware.  You have the right to take their changes, possibly 
modify them further, and distribute them (possibly with your own hardware).

The fact that you can't modify the software and load it back onto the 
tivo is irrelevent.  They are not restricting your distribution of the 
software in any way.  Rather, they're restricting the *running* of the 
software on their proprietary hardware platform.

Chris

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Thu, 14 Jun 2007, Alexandre Oliva wrote:
> 
> > Only with the GPLv3.
> 
> This is not true.  The terms of the GPLv2 that say you can't impose
> further restrictions on the exercise of the freedoms apply to the
> software under GPLv2 and GPLv3 just the same way.

The GPLv2 talks *only* about the software. You're making everything else 
up, and when I point out that your reading of the GPLv2 is insane, you 
just ignore my proofs of your internal inconsistency.

> So can you please explain to me how enabling TiVO to deny others the
> freedom that it received "in kind", failing to keep with the "in kind"
> spirit of the GPL, encourage people to work together, and to merge?

Because Tivo *IS NOT DENYING* those freedoms.

Tivo *respected* the freedoms, and gave source back, and gave you all the 
same rights you had to Linux originally, and to their modifications.

How stupid are you to not acknowledge that?

Tivo limited their *hardware*, not the software.

		Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:

> Tivo *respected* the freedoms, and gave source back, and gave you all the 
> same rights you had to Linux originally, and to their modifications.

> How stupid are you to not acknowledge that?

> Tivo limited their *hardware*, not the software.

Have you ever wondered *why* it limited the hardware?

Is it per chance such that I cannot modify the software that runs on
the hardware?

How is that respecting the freedoms?  How is this not imposing further
restrictions?


And, more importantly, how is it that permitting this makes for
*better* compliance with your tit-for-tat conceptions about the GPL?

I.e., if Tivoization is the only issue that you think makes GPLv3 a
worse license than GPLv2, and you like GPLv2 because of this
tit-for-tat, surely you should be able to explain why Tivoization
promotes this tit-for-tat notion better than GPLv3, right?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[david]

On Thu, 14 Jun 2007, Alexandre Oliva wrote:

> On Jun 14, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:
>
>> Tivo *respected* the freedoms, and gave source back, and gave you all the
>> same rights you had to Linux originally, and to their modifications.
>
>> How stupid are you to not acknowledge that?
>
>> Tivo limited their *hardware*, not the software.
>
> Have you ever wondered *why* it limited the hardware?
>
> Is it per chance such that I cannot modify the software that runs on
> the hardware?

if you cannot modify the software that runs on your Tivo hardware you 
haven't tried very hard.

true, they don't go out of their way to make it easy, but even if they 
didn't do the integrity checking of the system it still wouldn't be easy 
to load your own software on the tivo, there's no path to load the 
software without disassembling the hardware.

> How is that respecting the freedoms?  How is this not imposing further
> restrictions?

I think the software is all available at www.tivo.com/linux that provides 
you all the freedom that they got.

David Lang

>
> And, more importantly, how is it that permitting this makes for
> *better* compliance with your tit-for-tat conceptions about the GPL?
>
> I.e., if Tivoization is the only issue that you think makes GPLv3 a
> worse license than GPLv2, and you like GPLv2 because of this
> tit-for-tat, surely you should be able to explain why Tivoization
> promotes this tit-for-tat notion better than GPLv3, right?
>
>

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bernd Paysan]

[-- Attachment #1: Type: text/plain, Size: 1689 bytes --]

On Friday 15 June 2007 01:46, david@lang.hm wrote:
> if you cannot modify the software that runs on your Tivo hardware you
> haven't tried very hard.

Yes, but the GPLv2 clearly says that you don't have to try very hard. The 
preferred form of modification has to be distributed. I can run a 
decompiler or disassembler on a program, and I can even modify it in place 
with a hex editor (I have even modified programs in embedded ROMs by using 
focussed ion beam, so I know you can modify every program if you try hard 
enough). It's certainly possible to crack Tivo's firmware to accept my own 
signature, but it's *not* the preferred form of modification, the source 
code and Tivo's key for the signature.

Since Tivo's firmware only accepts a signed kernel, the combination of 
kernel+signature is the binary they ship. The kernel itself is useless, the 
signature as well. Therefore, you can imply that Tivo's key is part of 
the "other stuff" the GPLv2 mentions, because you need it to recreate the 
same code as Tivo did and shipped (compilers insert timestamps and such), 
and to modify that code. The source code is just a mean, the thing they 
shipped is the end (the binary), and they have to comply with the GPL for 
that binary - which by all means of practical understanding includes the 
signature.

"You can imply" means: It depends on court and legal system. I'm quite 
confident that in Germany, the legal system might favor the "GPLv2 does not 
allow tivoization" point of view, and in the USA, the legal sysem might do 
the opposite.

-- 
Bernd Paysan
"If you want it done right, you have to do it yourself"
http://www.jwdt.com/~paysan/

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Friday 15 June 2007 05:30:09 Bernd Paysan wrote:
> On Friday 15 June 2007 01:46, david@lang.hm wrote:
> > if you cannot modify the software that runs on your Tivo hardware you
> > haven't tried very hard.
>
> Yes, but the GPLv2 clearly says that you don't have to try very hard. The
> preferred form of modification has to be distributed. I can run a
> decompiler or disassembler on a program, and I can even modify it in place
> with a hex editor (I have even modified programs in embedded ROMs by using
> focussed ion beam, so I know you can modify every program if you try hard
> enough). It's certainly possible to crack Tivo's firmware to accept my own
> signature, but it's *not* the preferred form of modification, the source
> code and Tivo's key for the signature.

How is a signing key part of the "preferred form for modification"? It isn't a 
requirement to *modify* anything, just to *replace* something. (And I am 
*NOT* going to explain why "replace != modify" again)

> Since Tivo's firmware only accepts a signed kernel, the combination of
> kernel+signature is the binary they ship. The kernel itself is useless, the
> signature as well. Therefore, you can imply that Tivo's key is part of
> the "other stuff" the GPLv2 mentions, because you need it to recreate the
> same code as Tivo did and shipped (compilers insert timestamps and such),
> and to modify that code. The source code is just a mean, the thing they
> shipped is the end (the binary), and they have to comply with the GPL for
> that binary - which by all means of practical understanding includes the
> signature.

I can find no such requirement in the GPLv2. In fact, it actually says that 
you don't even have to be able to *USE* the program. See section 12 of the 
GPL if you don't believe me.

> "You can imply" means: It depends on court and legal system. I'm quite
> confident that in Germany, the legal system might favor the "GPLv2 does not
> allow tivoization" point of view, and in the USA, the legal sysem might do
> the opposite.

In light of the d-link case, I'm pretty certain that the German Courts 
interpretation of the GPLv2 makes "Tivoization" a violation. In the US I can 
say that the result would be "GPLv2 does not disallow tivoization". As I've 
pointed out in other posts, the GPLv2 actually *limits* itself to three 
specific "activities". Whether it was intended to "incidentally" cover other 
things or not, it does *clearly* state what it's scope is. If that scope *IS* 
*NOT* the intent of the person and/or person who authored the license, that 
text *SHOULD* *NOT* exist.

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Michael Poole]

Daniel Hazelton writes:

> On Friday 15 June 2007 05:30:09 Bernd Paysan wrote:
>> On Friday 15 June 2007 01:46, david@lang.hm wrote:
>> > if you cannot modify the software that runs on your Tivo hardware you
>> > haven't tried very hard.
>>
>> Yes, but the GPLv2 clearly says that you don't have to try very hard. The
>> preferred form of modification has to be distributed. I can run a
>> decompiler or disassembler on a program, and I can even modify it in place
>> with a hex editor (I have even modified programs in embedded ROMs by using
>> focussed ion beam, so I know you can modify every program if you try hard
>> enough). It's certainly possible to crack Tivo's firmware to accept my own
>> signature, but it's *not* the preferred form of modification, the source
>> code and Tivo's key for the signature.
>
> How is a signing key part of the "preferred form for modification"? It isn't a 
> requirement to *modify* anything, just to *replace* something. (And I am 
> *NOT* going to explain why "replace != modify" again)

The signing key determines a critical portion of the binary form that
was distributed.  You cannot produce that portion of the binary form
without the signing key.  Without that portion, the binary form does
not perform the function for which it is distributed.  If you think
such an input is not part of "the preferred form for modification", I
have a bridge to sell you.

The work that the GPL protects a recipient's right to modify and
redistribute is not the source code -- it is each form the user
receives.

>> Since Tivo's firmware only accepts a signed kernel, the combination of
>> kernel+signature is the binary they ship. The kernel itself is useless, the
>> signature as well. Therefore, you can imply that Tivo's key is part of
>> the "other stuff" the GPLv2 mentions, because you need it to recreate the
>> same code as Tivo did and shipped (compilers insert timestamps and such),
>> and to modify that code. The source code is just a mean, the thing they
>> shipped is the end (the binary), and they have to comply with the GPL for
>> that binary - which by all means of practical understanding includes the
>> signature.
>
> I can find no such requirement in the GPLv2. In fact, it actually says that 
> you don't even have to be able to *USE* the program. See section 12 of the 
> GPL if you don't believe me.

Section 12 of the GPL(v2) is a warranty and liability disclaimer.  It
is not an absolution of license obligations.  It limits the liability
of a distributor to the end user, not to copyright owners.

Michael Poole

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Al Viro]

On Thu, Jun 14, 2007 at 09:06:31AM -0700, Kevin Fox wrote:
 
> This whole argument is about the spirit of the GPL. Linus and others
> think the spirit is one thing, the FSF guys think its something else.

What the fsck it is, linux-kernel or bleeding Council of Nikea?

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Thursday 14 June 2007 12:06:31 Kevin Fox wrote:
> On Wed, 2007-06-13 at 20:42 -0400, Daniel Hazelton wrote:
> <SNIP>
>
> > > Do you deny that TiVo prevents you (or at least a random customer)
> > > from modifying the copy of Linux that they ship in their DVR?
> >
> > Exactly. They don't. What TiVO prevents is using that modified version on
> > their hardware. And they have that right, because the Hardware *ISN'T*
> > covered by the GPL.
>
> The hardware isn't directly covered by the GPL, correct. But, if they
> want to use the software on the hardware, they have to comply with the
> GPL. The software license can then influence hardware IF they want to
> use it badly enough.

Until GPLv3 there was no requirement that the modified code be able to operate 
on any given device - even the one its designed for. Claiming otherwise seems 
stupid to me, almost ridiculous.

> For example, the hardware is perfectly capable of being used to break
> the terms of the GPL by being used to distribute a modified binary
> without releasing the source. But the hardware's behavior is restricted
> by the software for the betterment of all.

But this can be done *NOW* - has been done by at least one company, IIRC. 
(and, IIRC again, they didn't so much as "not release the modified version" 
as "not release all tools/scripts/installation/build instructions")

> This whole argument is about the spirit of the GPL. Linus and others
> think the spirit is one thing, the FSF guys think its something else.
> Since the license is clearly owned by the FSF, I think they get the
> final vote on what they "intended" it to be when they wrote it, no? If
> they say they intended it to not allow Tivoization then believe them,
> because they are the only ones that know what they were thinking when
> they wrote it! The GPLv2 seems to allow it though. If Linus and friends
> want to allow it, then they can stay with the GPLv2. For those who want
> to disallow Tivoization, choose v3. No worries guys.
>
> > Do you understand that, or do I need to break out the finger-puppets next
> > ?
>
> Guys, we are all friends here. No reason to be so insulting. Its just a
> difference of opinion. People seem to be talking past each other instead
> of to one another. This usually happens when people are basing their
> underlying assumptions on different things and not listening to the
> other. Please take a step back and think about it.

I noticed that ten or twenty messages after I made that comment. In truth, the 
reason I made it was, and is, because I am tired of explaining the fact that 
there is no "one" interpretation of the GPLv2 - or any license - *UNLESS* it 
has been ruled on by a court. And even then, the courts ruling only applies 
to the parts of the license that were in contention before it.

DRH

> <SNIP>



-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Chris Adams]

Once upon a time, Alexandre Oliva  <aoliva@redhat.com> said:
>  if you distribute copies of such a program, [...]
>  you must give the recipients all the rights that you have
>
>So, TiVo includes a copy of Linux in its DVR.  
>
>TiVo retains the right to modify that copy of Linux as it sees fit.
>
>It doesn't give the recipients the same right.

Sure it does; you received a program (the kernel) and you can modify it.
You also received hardware; they don't support modification of that.
Nowhere in the license does it say they have to, because the license
only covers the program.

Or are you claiming that putting software on hardware makes the result a
derivative work?  I think it falls under the "mere aggregation" clause.

What if TiVo had put the kernel in a burned-in ROM (not flash, or on a
flash ROM with no provision for reprogramming it)?  Would that also
violate the "spirit" of the GPL?  Must any device that wishes to include
GPL code include additional hardware to support replacing that code
(even if that hardware is otherwise superfluous)?

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alan Cox]

> What if TiVo had put the kernel in a burned-in ROM (not flash, or on a
> flash ROM with no provision for reprogramming it)?  Would that also
> violate the "spirit" of the GPL?  Must any device that wishes to include
> GPL code include additional hardware to support replacing that code
> (even if that hardware is otherwise superfluous)?

This is an area the GPLv3 tries to clarify and for good reason. 

Of course these days in the US someone would probably sue arguing that a
ROM is protection scheme ;)

Alan

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alan Cox]

> What if TiVo had put the kernel in a burned-in ROM (not flash, or on a
> flash ROM with no provision for reprogramming it)?  Would that also
> violate the "spirit" of the GPL?  Must any device that wishes to include
> GPL code include additional hardware to support replacing that code
> (even if that hardware is otherwise superfluous)?

As a PS to the GPL3 comment here is the basic difference

ROM	-	I can't modify the code on the device
		The creator can't modify the code further on the device

Tivo	-	I can't modify the code on the device
		The owner can modify the code 

One is an implicit limitation of the hardware (just like I can't run
openoffice on a 4MB PC even though the license gives me the right to
try), the other is an artificial restriction.

One case is witholding freedom in the GPL sense by one party while
keeping it themselves, the other is a limitation of the system
inevitably imposed on everyone.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Forrest]

On Thu, Jun 14, 2007 at 02:52:48AM +0100, Alan Cox wrote:
> 
> As a PS to the GPL3 comment here is the basic difference
> 
> ROM	-	I can't modify the code on the device
> 		The creator can't modify the code further on the device
> 
> Tivo	-	I can't modify the code on the device
> 		The owner can modify the code
> 
> One is an implicit limitation of the hardware (just like I can't run
> openoffice on a 4MB PC even though the license gives me the right to
> try), the other is an artificial restriction.
> 
> One case is witholding freedom in the GPL sense by one party while
> keeping it themselves, the other is a limitation of the system
> inevitably imposed on everyone.

I've been following this discussion and I find this interesting.
Consider these two cases:

1.) I ship the device back to the manufacturer, they replace the ROM,
    and ship it back to me.

2.) I ship the device back to the manufacturer, they load new code
    into it, and ship it back to me.

How do these two differ?  Or is it now just a question of the ROM
being in a socket?  I can't see how the technicalities of how the
hardware is constructed can change the legality of the software.

-- 
Dan

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 13, 2007, Daniel Forrest <forrest@lmcg.wisc.edu> wrote:

> 1.) I ship the device back to the manufacturer, they replace the ROM,
>     and ship it back to me.

> 2.) I ship the device back to the manufacturer, they load new code
>     into it, and ship it back to me.

> How do these two differ?  Or is it now just a question of the ROM
> being in a socket?  I can't see how the technicalities of how the
> hardware is constructed can change the legality of the software.

I don't see that they differ.  If the software can be replaced, the
manufacturer ought to tell you how to do it.  It doesn't have to do it
for you, it doesn't have to give you the hardware tools needed to do
it, but if you're not able to start from the source code and the
information provided by the manufacturer and get to a modified version
of the software on the device, while the manufacturer could do it,
then the manufacturer is locking you in, and therefore you're not
free.  This is a clear violation of the spirit of the license, even if
the legalese might make room for some such misbehavior.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Michael Gerdau]

[-- Attachment #1: Type: text/plain, Size: 1765 bytes --]

> > As a PS to the GPL3 comment here is the basic difference
> > 
> > ROM	-	I can't modify the code on the device
> > 		The creator can't modify the code further on the device
> > 
> > Tivo	-	I can't modify the code on the device
> > 		The owner can modify the code
> > 
> > One is an implicit limitation of the hardware (just like I can't run
> > openoffice on a 4MB PC even though the license gives me the right to
> > try), the other is an artificial restriction.
> > 
> > One case is witholding freedom in the GPL sense by one party while
> > keeping it themselves, the other is a limitation of the system
> > inevitably imposed on everyone.
> 
> I've been following this discussion and I find this interesting.
> Consider these two cases:
> 
> 1.) I ship the device back to the manufacturer, they replace the ROM,
>     and ship it back to me.
> 
> 2.) I ship the device back to the manufacturer, they load new code
>     into it, and ship it back to me.
> 
> How do these two differ?  Or is it now just a question of the ROM
> being in a socket?  I can't see how the technicalities of how the
> hardware is constructed can change the legality of the software.

At first glance I think a construct where the manufacturer is obliged
to load _MY_ modified software in a timely fashion and at a reasonable
price into the device would fit my understanding of the GPL's spirit
though this leaves room for the definition of timely...

Best,
Michael
-- 
 Technosis GmbH, Geschäftsführer: Michael Gerdau, Tobias Dittmar
 Sitz Hamburg; HRB 89145 Amtsgericht Hamburg
 Vote against SPAM - see http://www.politik-digital.de/spam/
 Michael Gerdau       email: mgd@technosis.de
 GPG-keys available on request or at public keyserver

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alan Cox]

> I've been following this discussion and I find this interesting.
> Consider these two cases:
> 
> 1.) I ship the device back to the manufacturer, they replace the ROM,
>     and ship it back to me.
> 
> 2.) I ship the device back to the manufacturer, they load new code
>     into it, and ship it back to me.
> 
> How do these two differ?  Or is it now just a question of the ROM

Thats one thing I don't like about the GPL3 special casing.

> being in a socket?  I can't see how the technicalities of how the
> hardware is constructed can change the legality of the software.

In the replace/reflash the ROM case its about access to the righ tools -
I could do it myself, send it to another company to load my code etc.

In the Tivo case its about one company having the ability to make such
mods and blocking others from doing so.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bron Gondwana]

On Thu, Jun 14, 2007 at 02:52:48AM +0100, Alan Cox wrote:
> > What if TiVo had put the kernel in a burned-in ROM (not flash, or on a
> > flash ROM with no provision for reprogramming it)?  Would that also
> > violate the "spirit" of the GPL?  Must any device that wishes to include
> > GPL code include additional hardware to support replacing that code
> > (even if that hardware is otherwise superfluous)?
> 
> As a PS to the GPL3 comment here is the basic difference
> 
> ROM	-	I can't modify the code on the device
> 		The creator can't modify the code further on the device
> 
> Tivo	-	I can't modify the code on the device
> 		The owner can modify the code 

Tivo gets sick of the endless flamewars on lkml, signs a copy
of QNX, pushes it out to the hardware.  No more Linux on Tivo.

You also can't replace that but Tivo can.  As I see it the two
are completely orthagonal:

a) Can anyone but the manufacturer upload new software into a
   a device without taking extreme measures (soldering a new
   public-key-containing-chip onto the board)

b) Is the software currently installed on a device licenced under
   a rule which requires the distributor to also distribute source
   code upon request.

Now I think it would reasonable to ask that the source code be able
to be built by [same compiler, same flags, same ...] to produce an
identical binary to the one running on the device so you can confirm
that it's exactly the same code.  That's separate from being able to
upload a changed binary.

Bron.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Bron Gondwana <brong@fastmail.fm> wrote:

> Tivo gets sick of the endless flamewars on lkml, signs a copy
> of QNX, pushes it out to the hardware.  No more Linux on Tivo.

What do we lose?

Do we actually get any benefit whatsoever from TiVO's choice of Linux
as the kernel for its device?

Do TiVO customers lose anything from the change from one non-Free
software to another?  (the Linux binary, as shipped in the TiVO, has
become non-Free)

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bron Gondwana]

On Thu, Jun 14, 2007 at 01:58:26AM -0300, Alexandre Oliva wrote:
> On Jun 14, 2007, Bron Gondwana <brong@fastmail.fm> wrote:
> 
> > Tivo gets sick of the endless flamewars on lkml, signs a copy
> > of QNX, pushes it out to the hardware.  No more Linux on Tivo.
> 
> What do we lose?
> 
> Do we actually get any benefit whatsoever from TiVO's choice of Linux
> as the kernel for its device?

Sure, if they make any changes or fixes to Linux.  Other than that,
only the same benefit that Microsoft get from Windows piracy - TiVo
employees become familiar with Linux and are more likely to use it
and maybe contribute more in another job later.

What we don't get is TiVo having a better kernel than everyone else
because they've put some work into extending it without giving that
work back.

I see stuff in arch/powerpc/kernel/ which is Copyright "TiVo, Inc"
and more recent stuff in usb/net/asix.c and usb/net/mcs7830.c which
is more than I've ever contributed to the kernel, despite making
extensive use and even selling services where I ran servers with
Linux on them but didn't allow my customers to change the kernel
on the servers if there was some feature they wanted to play with.

> Do TiVO customers lose anything from the change from one non-Free
> software to another?  (the Linux binary, as shipped in the TiVO, has
> become non-Free)

Not particularly, no.  Other than maybe some nice features that TiVo
gains from being able to use Linux.

Bron.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Bron Gondwana <brong@fastmail.fm> wrote:

> On Thu, Jun 14, 2007 at 01:58:26AM -0300, Alexandre Oliva wrote:

>> Do we actually get any benefit whatsoever from TiVO's choice of Linux
>> as the kernel for its device?

> Sure, if they make any changes or fixes to Linux.  Other than that,
> only the same benefit that Microsoft get from Windows piracy - TiVo
> employees become familiar with Linux and are more likely to use it
> and maybe contribute more in another job later.

Now, what if TiVO actually permitted all of its customers to make
changes or fixes to Linux, and become familiar with it and become more
likely to use it and maybe contribute more later?

Would we lose more or gain more?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bill Nottingham]

Alexandre Oliva (aoliva@redhat.com) said: 
> > Sure, if they make any changes or fixes to Linux.  Other than that,
> > only the same benefit that Microsoft get from Windows piracy - TiVo
> > employees become familiar with Linux and are more likely to use it
> > and maybe contribute more in another job later.
> 
> Now, what if TiVO actually permitted all of its customers to make
> changes or fixes to Linux, and become familiar with it and become more
> likely to use it and maybe contribute more later?

a) there's nothing that prevents a Tivo user from changing or fixing
Linux completely outside of the Tivo
b) the 'interesting' bits that someone would modify the Tivo to change
*aren't actually the bits that everyone is kvetching about here*

Bill

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Bill Nottingham <notting@redhat.com> wrote:

> Alexandre Oliva (aoliva@redhat.com) said: 
>> > Sure, if they make any changes or fixes to Linux.  Other than that,
>> > only the same benefit that Microsoft get from Windows piracy - TiVo
>> > employees become familiar with Linux and are more likely to use it
>> > and maybe contribute more in another job later.
>> 
>> Now, what if TiVO actually permitted all of its customers to make
>> changes or fixes to Linux, and become familiar with it and become more
>> likely to use it and maybe contribute more later?

> a) there's nothing that prevents a Tivo user from changing or fixing
> Linux completely outside of the Tivo

But how about inside the TiVO, so as to use Linux and the rest of the
GNU/Linux distro put in there for an even better DVR experience?

Sure, this might still be accomplished on another hardware platform.
But the TiVO already has all the hardware there, and you already have
all the software ready to work on it.  Except that you can't change
it.  You'd have to waste time and money just to get to the same status
on another hardware platform.

What do we gain?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bill Nottingham]

Alexandre Oliva (aoliva@redhat.com) said: 
> But how about inside the TiVO, so as to use Linux and the rest of the
> GNU/Linux distro put in there for an even better DVR experience?
> 
> Sure, this might still be accomplished on another hardware platform.
> But the TiVO already has all the hardware there, and you already have
> all the software ready to work on it.  Except that you can't change
> it.  You'd have to waste time and money just to get to the same status
> on another hardware platform.
> 
> What do we gain?

Nothing. But that's not the terms it was licensed under, and no matter
what someone may claim about the *spirit* of the license, adding clauses
that restrict how you can deploy GPL software for use is a fundamental
enough change to the practical aspect of the license that it's no wonder
that people will choose not to use it.

If the designers of the license are more interested in vendettas against
those using the software in a way they didn't see beforehand (come on,
explicitly trying to define 'consumer product'?) in order to accomplish
pyrrhic victories (people moving to other platforms instead of using your
newly licensed code), that's fine, it's their choice. But not everyone
will want to follow that choice.

Bill

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bernd Schmidt]

Alexandre Oliva wrote:
> On Jun 14, 2007, Bron Gondwana <brong@fastmail.fm> wrote:
> 
>> Tivo gets sick of the endless flamewars on lkml, signs a copy
>> of QNX, pushes it out to the hardware.  No more Linux on Tivo.
> 
> What do we lose?
> 
> Do we actually get any benefit whatsoever from TiVO's choice of Linux
> as the kernel for its device?

Do they contribute back any code that makes Linux better?
If Tivo doesn't, what about other vendors who may be in a similar situation?


Bernd

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 13, 2007, Chris Adams <cmadams@hiwaay.net> wrote:

> Once upon a time, Alexandre Oliva  <aoliva@redhat.com> said:
>> if you distribute copies of such a program, [...]
>> you must give the recipients all the rights that you have

>> So, TiVo includes a copy of Linux in its DVR.  

>> TiVo retains the right to modify that copy of Linux as it sees fit.

>> It doesn't give the recipients the same right.

> Sure it does; you received a program (the kernel) and you can modify it.

If I take the software I received, build it and install it on the same
hardware, it won't run.  Something is missing in the source code I
received, I guess..

If I make changes to the source code, build it, and install it on that
same computer, it won't run.  How is that being able to modify *that*
copy of the program?

If TiVo makes the same changes, builds tehm, and installs it on my
computer, it will run just fine.  How are they passing on the right
they had to me?

> You also received hardware; they don't support modification of that.

They don't have to support them.  They don't have to help me if it
breaks.  But if they can do it and I can't, they're failing to comply
with the spirit of the GPL.

> Nowhere in the license does it say they have to, because the license
> only covers the program.

They can't distribute the program while imposing restrictions on
modification not present in the software license itself.

> Or are you claiming that putting software on hardware makes the result a
> derivative work?  I think it falls under the "mere aggregation" clause.

I tend to agree, in this particular case, but IANAL.  I don't rule out
derivative works in future attempts to find loopholes in the GPL.

> What if TiVo had put the kernel in a burned-in ROM

Then they wouldn't have the ability to change it any more, so there
wouldn't be such a right to pass on to the users.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Wed, 13 Jun 2007, Alexandre Oliva wrote:
> 
> So, TiVo includes a copy of Linux in its DVR.  

Stop right there.

You seem to make the mistake to think that software is something physical.

> TiVo retains the right to modify that copy of Linux as it sees fit.

No. If you were logical (which you are not), you would admit that 
 (a) physical property is very different from intellectual property (the 
     FSF seems to admit that when it suits their needs, not otherwise)
 (b) They never modified "a copy" of Linux - they simply replaced it with
     "another copy" of Linux. The only thing that actually got *modified* 
     was their hardware!

The first copy didn't "morph" into a second copy. There was no "physical" 
software that was molded.  They do need to follow the GPLv2, since clearly 
they _do_ distribute Linux, but you have all the same rights as they do 
with regard to the *software*. 

The fact that they maintained some control of the *hardware* (and some 
software they wrote too) they designed is _their_ choice.

What Tivo did and do, is to distribute hardware that can *contain* a copy 
of Linux (or just about anything else, for that matter - again, there's 
a difference between physical and intellectual property).

And their hardware (and firmware) will run some integrity checks on 
*whatever* copies of software they have.  This is all totally outside 
Linux itself.

Btw, according to your _insane_ notion of "a copy" of software, you can 
never distribute GPL'd software on a CD-ROM, since you've taken away the 
right of people to modify that CD-ROM by burning and fixating it. So 
according to your (obvously incorrect) reading of the GPLv2, every time 
Red Hat sends anybody a CD-ROM, they have restricted peoples right to 
modify the software on that CD-ROM bymaking it write-only.

See? Your reading of the license doesn't _work_. Mine does. What I say is 
that when you distribute software, you don't distribute "a copy" of 
software, you distribute the _information_ about the software, so that 
others can take it and modify it. And notice? My reading of the license 
must be the correct one, since my reading actually makes sense, unlike 
yours.

And yes, when Tivo distributes Linux, they give everybody else all the 
same rights they have - with respect to Linux! No, not with respect to 
their hardware, but that's a totally different thing, and if you cannot 
wrap your mind around the difference between "the software that is on a 
CD" and the "piece of plastic that is the CD", and see that when you 
replace "CD" with any other medium, the equation doesn't change, I don't 
know what to say.

> It doesn't give the recipients the same right.
> 
> Oops.
> 
> Sounds like a violation of the spirit to me.

Only if you extend the license to the *hardware*. Oops. Which it never did 
before. 

In other words, you basically try to change the rules. The GPLv2 clearly 
states that it's about software, not hardware. All the language you quoted 
talks about software.

In other words, the only way to argue that I'm wrong is to try to twist 
the meanings of the words, and say that words only mean one specific thing 
that _you_ claim are their meaning.

And I'm saying you act like Humpty Dumpty when you do. You can argue that 
way all you like, but your argument is nonsensical. It's akin to the 
argument that "God is perfect. Perfect implies existence. Therefore God 
exists".

That kind of argument only works if you *define* the words to suit your 
argument. But it's a logical fallacy.

And I'm saying that the GPLv2 can mroe straightforwardly be read the way I 
read it - to talk about software, and to realize that software is not "a 
copy", it's a more abstract thing. You get Linux when you buy a Tivo (or 
preferably - don't buy it, since you don't like it), and that means that 
they have to give you access to and control over the SOFTWARE. But nowhere 
in the GPL (in the preamble or anywhere else) does it talk about giving 
you control over the HARDWARE, and the only way you can twist the GPLv2 to 
say that is by trying to re-define what the words mean.

And then you call *me* confused? After you yourself admitted that the FSF 
actually agrees with me, and that what Tivo did was not a license 
violation?

Trust me, I'm not the confused person here.

I'm perfectly fine with other people wanting to extend the license to 
cover the hardware, but I am *not* perfectly fine with people then trying 
to claim I'm confused just because I don't agree with them.

Face it: the GPLv3 is a _new_ license. Making funamentally _different_ and 
_new_ restrictions that do not exist in the GPLv2, and do not exist in the 
preamble. Any language attempts to make it appear otherwise are just 
sophistry.

And btw, just to make it clear: as far as I'm concerned, you can read the 
preamble and the word "freedom" and "rigths" _your_ way. I'm not objecting 
to that at all. If you read it so that you think it's wrong to distribute 
GPL'd software on a CD-ROM, that's really not my problem. You do whatever 
you want to, and think the license means whatever you want to.

What I'm objecting to is how you claim that anybody that doesn't follow 
your interpretation is "confused". When clearly even the FSF lawyers agree 
that my interpretation was _correct_, and I don't think your 
interpretation even makes sense!

			Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 13, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:

> On Wed, 13 Jun 2007, Alexandre Oliva wrote:
>> 
>> So, TiVo includes a copy of Linux in its DVR.  

> Stop right there.

> You seem to make the mistake to think that software is something physical.

Err, no.  Software, per legal definitions in Brazil, US and elsewhere,
require some physical support.  That's the hard disk in the TiVO DVR,
in this case.  I don't see how this matters, though.

>> TiVo retains the right to modify that copy of Linux as it sees fit.

>  (b) They never modified "a copy" of Linux - they simply replaced it with
>      "another copy" of Linux. The only thing that actually got *modified* 
>      was their hardware!

Per this reasoning, nobody never modifies software.  When you open a
source file in your editor, you make changes to it, then save it,
you're not modifying it, you're replacing it with another copy, and
the only thing that actually got modified was the hardware.

Maybe look what "modify" means in legal context?

Then refer to the GPL:

  2. You may modify your copy or copies of the Program or any portion
  of it, thus forming a work based on the Program,

> And their hardware (and firmware) will run some integrity checks on 
> *whatever* copies of software they have.  This is all totally outside 
> Linux itself.

Agreed.  But as it turns out they use these checks to stop people from
modifying the copy of Linux they ship in the device, and this
restriction is a GPL violation because they don't provide information
you need to build a functioning modified version.

> Btw, according to your _insane_ notion of "a copy" of software, you can 
> never distribute GPL'd software on a CD-ROM, since you've taken away the 
> right of people to modify that CD-ROM by burning and fixating it.

You don't retain that right yourself.  When you pass that copy on, you
pass it on with all the rights that you have.  No problem here.  This
is no different from the software on ROM.

> And I'm saying that the GPLv2 can mroe straightforwardly be read the way I 
> read it - to talk about software, and to realize that software is not "a 
> copy", it's a more abstract thing.

If you choose to disregard the legal meaning of the legal terms used
in the GPLv2, you may have a point.

> that means that they have to give you access to and control over the
> SOFTWARE.

Yes.  That's all I'm saying.  You just can't use the hardware to take
that control away.  That would be a violation of the license.

> Face it: the GPLv3 is a _new_ license. Making funamentally _different_ and 
> _new_ restrictions that do not exist in the GPLv2,

This is true.

> and do not exist in the preamble.

This is not true.

The spirit remains the same: let people modify and share the software.

If the binary you got can't be created out of the corresponding
sources, something is missing.  If it won't run without this missing
bit, you're missing functional portions of the source code.  This all
means the hardware is being used to impose a restriction on
modification of the software, which is against the spirit of the GPL,
and quite likely against its letter as well.

If you don't want it to be so, you can always add an additional
permission that clarifies this bit, such that TiVO and you will be
happy.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Matt Keenan]

Alexandre Oliva wrote:
> On Jun 13, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:
>
>   
>> On Wed, 13 Jun 2007, Alexandre Oliva wrote:
>>     
>>> So, TiVo includes a copy of Linux in its DVR.  
>>>       
>
>   
>> Stop right there.
>>     
>
>   
>> You seem to make the mistake to think that software is something physical.
>>     
>
> Err, no.  Software, per legal definitions in Brazil, US and elsewhere,
> require some physical support.  That's the hard disk in the TiVO DVR,
> in this case.  I don't see how this matters, though.
>
>   
I'm now intrigued, where are these (Brazilian and US) definitions
stipulated, and under what authority?

Matt

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Michael Poole]

Matt Keenan writes:

> Alexandre Oliva wrote:
>>
>> Err, no.  Software, per legal definitions in Brazil, US and elsewhere,
>> require some physical support.  That's the hard disk in the TiVO DVR,
>> in this case.  I don't see how this matters, though.
>>
>>   
> I'm now intrigued, where are these (Brazilian and US) definitions
> stipulated, and under what authority?

In the US, 17 USC 101 (the "Definitions" section of the title dealing
with Copyright) makes this definition:

    A "computer program" is a set of statements or instructions to be
    used directly or indirectly in a computer in order to bring about
    a certain result.

As its purpose is to outline the scope of copyright law, this
definition is made under the authority granted to Congress by Article
I, Section 8 of the United States Constitution.

Michael Poole

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Matt Keenan]

Michael Poole wrote:
> Matt Keenan writes:
>
>   
>> Alexandre Oliva wrote:
>>     
>>> Err, no.  Software, per legal definitions in Brazil, US and elsewhere,
>>> require some physical support.  That's the hard disk in the TiVO DVR,
>>> in this case.  I don't see how this matters, though.
>>>
>>>   
>>>       
>> I'm now intrigued, where are these (Brazilian and US) definitions
>> stipulated, and under what authority?
>>     
>
> In the US, 17 USC 101 (the "Definitions" section of the title dealing
> with Copyright) makes this definition:
>
>     A "computer program" is a set of statements or instructions to be
>     used directly or indirectly in a computer in order to bring about
>     a certain result.
>
> As its purpose is to outline the scope of copyright law, this
> definition is made under the authority granted to Congress by Article
> I, Section 8 of the United States Constitution.
>
>   
But where is the part that says it "requires some physical support"? It
says what it is; "a set of statements or instructions", how it should be
used; "to be used directly or indirectly in a computer", and what
purpose it serves; "in order to bring about a certain result", but it
doesn't seem to indicate that it "requires physical support" aka needing
some physical representation. I suspect this argument boils down to the
philosophical debate of whether ideas (in this case software) can be
truely devoid of the physical.

Matt

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Michael Poole]

Matt Keenan writes:

> Michael Poole wrote:
>> Matt Keenan writes:
>>
>>   
>>> Alexandre Oliva wrote:
>>>     
>>>> Err, no.  Software, per legal definitions in Brazil, US and elsewhere,
>>>> require some physical support.  That's the hard disk in the TiVO DVR,
>>>> in this case.  I don't see how this matters, though.
>>>>
>>>>   
>>>>       
>>> I'm now intrigued, where are these (Brazilian and US) definitions
>>> stipulated, and under what authority?
>>>     
>>
>> In the US, 17 USC 101 (the "Definitions" section of the title dealing
>> with Copyright) makes this definition:
>>
>>     A "computer program" is a set of statements or instructions to be
>>     used directly or indirectly in a computer in order to bring about
>>     a certain result.
>>
>> As its purpose is to outline the scope of copyright law, this
>> definition is made under the authority granted to Congress by Article
>> I, Section 8 of the United States Constitution.
>>
>>   
> But where is the part that says it "requires some physical support"? It
> says what it is; "a set of statements or instructions", how it should be
> used; "to be used directly or indirectly in a computer", and what
> purpose it serves; "in order to bring about a certain result", but it
> doesn't seem to indicate that it "requires physical support" aka needing
> some physical representation. I suspect this argument boils down to the
> philosophical debate of whether ideas (in this case software) can be
> truely devoid of the physical.

Sets of statements or instructions that cannot "be used directly or
indirectly in a computer in order to bring about a certain result"
are, for the purposes of copyright law, not software.  "A computer" is
a physical device.  It always has been a physical device, except when
"computer" referred to a person who performed computations -- and that
meaning fell out of common use 40 years ago.  Any suggestion that the
requirement to be usable on a physical device is significantly
different from "require[s] some physical support" is laughably stupid.

17 USC 102 requires that copyright protection only subsists in works
that are "fixed in any tangible medium of expression" -- which
obviously includes paper and hard drives, and has been ruled to
include volatile program memory (the 9th Circuit's holding to this
effect in MAI Systems Corp. v. Peak Computer, Inc. is what inspired
the addition of 17 USC 117(c)).  If the set of instructions exist only
in transmission or in someone's head, they are not protected by
copyright law.

Michael Poole

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Matt Keenan <tank.en.mate@gmail.com> wrote:

> Alexandre Oliva wrote:

>> Err, no.  Software, per legal definitions in Brazil, US and elsewhere,
>> require some physical support.  That's the hard disk in the TiVO DVR,
>> in this case.  I don't see how this matters, though.

> I'm now intrigued, where are these (Brazilian and US) definitions
> stipulated, and under what authority?

http://www.planalto.gov.br/ccIVIL_03/LEIS/L9609.htm

LEI Nº 9.609 , DE 19 DE FEVEREIRO DE 1998.

Art. 1º Programa de computador é a expressão de um conjunto organizado
de instruções em linguagem natural ou codificada, contida em suporte
                                                  ^^^^^^^^^^^^^^^^^^
físico de qualquer natureza, de emprego necessário em máquinas
^^^^^^^^^^^^^^^^^^^^^^^^^^^
automáticas de tratamento da informação, dispositivos, instrumentos ou
equipamentos periféricos, baseados em técnica digital ou análoga, para
fazê-los funcionar de modo e para fins determinados.


-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bernd Paysan]

[-- Attachment #1: Type: text/plain, Size: 2096 bytes --]

On Thursday 14 June 2007 01:49, Alexandre Oliva wrote:
> Oh, good, let's take this one.
>
>   if you distribute copies of such a program, [...]
>   you must give the recipients all the rights that you have
>
> So, TiVo includes a copy of Linux in its DVR.
>
> TiVo retains the right to modify that copy of Linux as it sees fit.
>
> It doesn't give the recipients the same right.
>
> Oops.
>
> Sounds like a violation of the spirit to me.
>
> Sounds like plugging this hole would retain the same spirit.

Note that Harald Welte has already managed to force Siemens to unlock 
a "tivoized" Linux router with the GPLv2 in Germany. German contract law 
values intention when the contract has no specific clause that deals with 
the issue, and in German law, an accepted license is a contract.

So the fact that tivoizing Linux is against the spirit of the GPLv2 is a 
court-proof fact, not just some speculation.

What about if your GPL program ends up in a piece of hardware (e.g. a ROM, 
or an embedded ROM, or if it's some GPL code from OpenCores, as gate 
netlist in silicon)? My interpretation is that you need a permission from 
the author for doing that, unless there's an easy way to replace it with a 
modified copy (e.g. if you put the OpenCores stuff into an FPGA, replacing 
the configuration PROM would do it).

Some people have difficulties with intentions of contracts rather than 
direct rules. That may be due to different rules in different countries. In 
continental Europe, contract law usually bases on Code Napoleon, and 
there, "good faith" is an important principle (and "good faith" means that 
the intention is more important than the actual coded practices). In the 
roman law that was used before and has survived in countries who didn't let 
Napoleon in (like the UK and the USA), it's slightly different. But a 
contract or a license still is not a program where anything that isn't said 
explicitely isn't said at all.

-- 
Bernd Paysan
"If you want it done right, you have to do it yourself"
http://www.jwdt.com/~paysan/

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

RE: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Schwartz]

> What about if your GPL program ends up in a piece of hardware
> (e.g. a ROM,
> or an embedded ROM, or if it's some GPL code from OpenCores, as gate
> netlist in silicon)? My interpretation is that you need a permission from
> the author for doing that, unless there's an easy way to replace
> it with a
> modified copy (e.g. if you put the OpenCores stuff into an FPGA,
> replacing
> the configuration PROM would do it).

The GPL does not require it to be easy in fact to modify the piece of
software. It just requires that you have the right to modify it, that is,
that there be no legal obstacles in your way. You are entitled to the source
code in modifiable, understandable form. There are no legal restrictions,
other than those in the GPL and in the law, on what you can do with it.

What you are actually *able* to do, however, depends upon a wide variety of
factors way outside the scope of the GPL.

By the way, I have a lot of sympathy for the argument that *if* you provide
me a binary made from GPL'd code that required a key to produce that binary,
I am entitled to that key. The key is precisely analogous to any other piece
of source code -- it is mathematically 'combined' and 'processed' by tools
to produce the final, distributed executable. If there's some rational basis
for a legal difference between a signing key and a header file, I don't know
what it is.

DS

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bernd Paysan]

[-- Attachment #1: Type: text/plain, Size: 527 bytes --]

On Thursday 14 June 2007 22:47, David Schwartz wrote:
> The GPL does not require it to be easy in fact to modify the piece of
> software.

Yes it does, section 3: "The source code for a work means the preferred form 
of the work for making modifications to it." It then even lists that you 
need to provide all the scripts and stuff you use to make it easy for you.

Come on, *READ* the GPL, before you argue.

-- 
Bernd Paysan
"If you want it done right, you have to do it yourself"
http://www.jwdt.com/~paysan/

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

RE: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Schwartz]

> On Thursday 14 June 2007 22:47, David Schwartz wrote:

> > The GPL does not require it to be easy in fact to modify the piece of
> > software.

> Yes it does, section 3: "The source code for a work means the
> preferred form
> of the work for making modifications to it." It then even lists that you
> need to provide all the scripts and stuff you use to make it easy for you.

> Come on, *READ* the GPL, before you argue.

Nice job quoting me out of context. For the record, here is the context, and
it addresses your criticism already:

"The GPL does not require it to be easy in fact to modify the piece of
software. It just requires that you have the right to modify it, that is,
that there be no legal obstacles in your way. You are entitled to the source
code in modifiable, understandable form. There are no legal restrictions,
other than those in the GPL and in the law, on what you can do with it."

DS

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Lennart Sorensen]

On Wed, Jun 13, 2007 at 07:38:14PM -0300, Alexandre Oliva wrote:
> So let's go back to the preamble, that provides motivations and some
> guidance as to the interpretation of the legal text (i.e., the spirit
> of the license):
> 
>   [...] the GNU General Public License is intended to guarantee your
>   freedom to share and change free software--to make sure the software
>   is free for all its users. [...]
> 
>   [...] Our General Public Licenses are designed to make sure that you have
>   the freedom to distribute copies of free software (and charge for
>   this service if you wish), that you receive source code or can get
>   it if you want it, that you can change the software or use pieces of
>   it in new free programs; and that you know you can do these things.
> 
>   To protect your rights, we need to make restrictions that forbid
>   anyone to deny you these rights or to ask you to surrender the
>   rights.  These restrictions translate to certain responsibilities
>   for you if you distribute copies of the software, or if you modify
>   it.
> 
>   [...] if you distribute copies of such a program, whether gratis or
>   for a fee, you must give the recipients all the rights that you have
> 
> 
> Can anyone show me how any of the provisions of GPLv3 fails to meet
> this spirit?

Well much as I don't like what Tivo did with only allowing signed
kernels to run, I don't see anything in the above that says they can't
do that.  They let you have the code and make changes to it, they just
don't let you put that changed stuff on the device they build.  The
software is free, even though the hardware is locked down.  The GPL v3
really seems to change the spirit to try and cover usage and hardware
behaviour, while the spirit of the GPL v2 seemed to me at least to
simply be to allow people to copy and change and use the code, and pass
that on to people.  It didn't have anything to do with what they did
with it on hardware.  Nothing prevents you from taking tivos kernel
changes and building your own hardware to run that code on, and as such
the spirit of the GPL v2 seems fulfilled.  It covers freedom of the
source code and resulting binaries, not of the platform you run it on.
The GPL v3 has a much broader coverage of what it wants to control,
which to me means the spirit is different.

I don't have a tivo, I use mythtv on my own PC.  Tivo doesn't force you
to buy their hardware after all.

--
Len Sorensen

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, lsorense@csclub.uwaterloo.ca (Lennart Sorensen) wrote:

> They let you have the code and make changes to it,

Not to the software installed in the device.

What they do is like an author A who distributes a program to user B
under a non-Free Software license, and to user C under a Free Software
license.

C passes the program on to B under the same license.  Now B has two
copies of the program.  One is free, the other is not.

Except that TiVO had no right to distribute the program under non-Free
terms in the first place, because it was not the author, and the
license it had explicitly said it couldn't impose further
restrictions.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bongani Hlope]

On Thursday 14 June 2007 21:32:08 Alexandre Oliva wrote:
> On Jun 14, 2007, lsorense@csclub.uwaterloo.ca (Lennart Sorensen) wrote:
> > They let you have the code and make changes to it,
>
> Not to the software installed in the device.

So now you want access to all the software that is installed in their device? 
Could you explain that please? You do have access to the GPL code that they 
used. If you buy one of Google's Search Appliance, are you expecting to allow 
you to make changes to the software that is installed on that device?

>
> What they do is like an author A who distributes a program to user B
> under a non-Free Software license, and to user C under a Free Software
> license.
>
> C passes the program on to B under the same license.  Now B has two
> copies of the program.  One is free, the other is not.
>
> Except that TiVO had no right to distribute the program under non-Free
> terms in the first place, because it was not the author, and the
> license it had explicitly said it couldn't impose further
> restrictions.

Reread what you wrote here and see the complete lack of logic in your 
argument.

Author A are Linux developers who distribute their software under GPL 2, TiVO 
gets the software under the same license and distributes it to their end 
users. They then make the all the changes to the Linux Kernel available to 
their end users under the same terms that they got from the Linux kernel 
developers.

What freedom did they take away?

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Bongani Hlope <bonganilinux@mweb.co.za> wrote:

> On Thursday 14 June 2007 21:32:08 Alexandre Oliva wrote:
>> On Jun 14, 2007, lsorense@csclub.uwaterloo.ca (Lennart Sorensen) wrote:
>> > They let you have the code and make changes to it,
>> 
>> Not to the software installed in the device.

> So now you want access to all the software that is installed in
> their device?

What's under the GPL.

> If you buy one of Google's Search Appliance, are you expecting to
> allow you to make changes to the software that is installed on that
> device?

Arguably, if I purchased the device, I ought to be entitled to make
changes to it, yes.  But that's a distraction I'd rather not get into
ATM.

> They then make the all the changes to the Linux Kernel available to
> their end users under the same terms that they got from the Linux
> kernel developers.

> What freedom did they take away?

They prevent the user from installing and running modified versions of
the program on the box, while they can still do it themselves on the
same box.

I guess I must have repeated this at least a dozen times in this
thread, so I'll just refrain from repeating this point from now on.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Dave Neuer]

On 6/14/07, Lennart Sorensen <lsorense@csclub.uwaterloo.ca> wrote:
> Nothing prevents you from taking tivos kernel
> changes and building your own hardware to run that code on, and as such
> the spirit of the GPL v2 seems fulfilled.

Oh, come on: you're not serious, right? Something indeed prevents me
-- the fact that I'm not a hardware manufacturer, I don't have fabs,
outsource vendors to provide me w/ designs, ASICs, etc. Nor to I have
the money to pay one-off prices for various components if they're even
available in batches that small.

This argument seems totally disingenuous to me. The GPLv<3 was written
in a time when the majority of sotware to which the license was
applied was written for general purpose computers. The "user" was the
owner of the computer, and Freedom 0 was about letting that user RUN
modified copies of the software.

Things have changed a lot; we're surrounded by embedded computers, and
Freedom 0 seems to strongly imply I should have the right to run
modified versions of the Free Software I own on the hardware I OWN. Or
is the future of Open Source that you'll be able to hack on free
software as long as you work for Intel, Red Hat, TiVO, Google or OSDL?
Or own many-thousand-$$ fab printer?

Look, I totally respect Linus' and others' position that the license
is an inappropriate way to enforce what they feel are hardware design
decisions, but can we dispense w/ the silly argument that the intent
of the GPL is fullfilled as long as the user is allowed to modify the
software where modify means "imagine a world where they'd be able to
run" it?

Dave

RE: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Schwartz]

> Oh, come on: you're not serious, right? Something indeed prevents me
> -- the fact that I'm not a hardware manufacturer, I don't have fabs,
> outsource vendors to provide me w/ designs, ASICs, etc. Nor to I have
> the money to pay one-off prices for various components if they're even
> available in batches that small.
>
> This argument seems totally disingenuous to me. The GPLv<3 was written
> in a time when the majority of sotware to which the license was
> applied was written for general purpose computers. The "user" was the
> owner of the computer, and Freedom 0 was about letting that user RUN
> modified copies of the software.

And what about people who can't modify the Linux kernel? They don't know C.
They don't know how to use a shell. They're not familiar with UNIX operating
systems at all. Maybe they aren't smart enough to modify kernel code.

The GPL is about having the legal right to modify the software and being
able to put other people's distributed improvements back into the original
code base. It does not guarantee that you will actually be able to modify
the software and get it to work on some particular hardware.

I certainly understood the GPL as ensuring the right to get the source code
so that you could do something else with it. I never understood the GPL to
be about getting hardware to do something else just because it ran GPL'd
software.

DS

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Dave Neuer]

On 6/14/07, David Schwartz <davids@webmaster.com> wrote:
>
> And what about people who can't modify the Linux kernel? They don't know C.
> They don't know how to use a shell. They're not familiar with UNIX operating
> systems at all. Maybe they aren't smart enough to modify kernel code.

I learned C in part by modifying the Linux kernel and running the
modified kernel on hardware I own, and enabling precisely that kind of
tinkering is what the "spirit" of the GPL is about, as is quite plain
(to me) from the preamble.

>
> The GPL is about having the legal right to modify the software and being
> able to put other people's distributed improvements back into the original
> code base.

I agree that is what the letter of the GPLv<3 is about.

> It does not guarantee that you will actually be able to modify
> the software and get it to work on some particular hardware.

Please don't conflate my endorsement of the "spirit" of the GPL with
Alexandre's assertion that the GPLv2 forbids TiVOisation. I don't
agree with him. My point is that people arguing that the spirit of the
GPL doesn't revolve around the freedom of the end user to modify the
software *and* run modified copies seem to be missing the point. Linus
gets that, as he said in a previous message, he just doesn't
personally care about freedom defined that way.

Dave

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Lennart Sorensen]

On Thu, Jun 14, 2007 at 04:24:19PM -0400, Dave Neuer wrote:
> Oh, come on: you're not serious, right? Something indeed prevents me
> -- the fact that I'm not a hardware manufacturer, I don't have fabs,
> outsource vendors to provide me w/ designs, ASICs, etc. Nor to I have
> the money to pay one-off prices for various components if they're even
> available in batches that small.

Yes I am serious.  I wouldn't want to buy any such locked down hardware,
but that still doesn't mean that I don't think it fits within the spirit
of the GPLv2.

> This argument seems totally disingenuous to me. The GPLv<3 was written
> in a time when the majority of sotware to which the license was
> applied was written for general purpose computers. The "user" was the
> owner of the computer, and Freedom 0 was about letting that user RUN
> modified copies of the software.
> 
> Things have changed a lot; we're surrounded by embedded computers, and
> Freedom 0 seems to strongly imply I should have the right to run
> modified versions of the Free Software I own on the hardware I OWN. Or
> is the future of Open Source that you'll be able to hack on free
> software as long as you work for Intel, Red Hat, TiVO, Google or OSDL?
> Or own many-thousand-$$ fab printer?

I think it depends on the type of hardware.  Certainly I agree some
types of hardware really should not allow you to change the code on them
due to the potential risks from doing so.  Hence if a license starts to
get into the grey area that covers such things, it is getting onto some
thin ice that is probably should stay off.  You risk excluding things
you didn't intend to exclude while almost certainly still missing things
you would like to have excluded.  I agree that for many devices I could
buy, being able to change the code on it would be great, and that there
generally is no good reason to deny me from doing it, but I don't think
it is worth the risk to put such a requirement into the license, and I
certainly never read the GPLv2 to in any way imply such a thing.
Apparently from what I can see, Linus never read any such thing in it
either when he chose to use it.  In fact I think you have to already
have a very narrow preset view in order to read the GPLv2 in such as
way as to think it intended to prevent such things.

> Look, I totally respect Linus' and others' position that the license
> is an inappropriate way to enforce what they feel are hardware design
> decisions, but can we dispense w/ the silly argument that the intent
> of the GPL is fullfilled as long as the user is allowed to modify the
> software where modify means "imagine a world where they'd be able to
> run" it?

It seems many people really do feel that it is fulfilled.  They may
think it is a stupid hardware design and they may also chose not to buy
such hardware, but at the same time they can be perfectly willing to say
that as long as the modified sources are provided, that is good enough
since further development of the source can be done, never mind what you
can do with that particular locked down door stop the code was modified
to support.  Not everyone views the world through the eyes of RMS.

--
Len Sorensen

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Dmitry Torokhov]

On 6/14/07, Dave Neuer <mr.fred.smoothie@pobox.com> wrote:
> On 6/14/07, Lennart Sorensen <lsorense@csclub.uwaterloo.ca> wrote:
> > Nothing prevents you from taking tivos kernel
> > changes and building your own hardware to run that code on, and as such
> > the spirit of the GPL v2 seems fulfilled.
>
> Oh, come on: you're not serious, right? Something indeed prevents me
> -- the fact that I'm not a hardware manufacturer, I don't have fabs,
> outsource vendors to provide me w/ designs, ASICs, etc. Nor to I have
> the money to pay one-off prices for various components if they're even
> available in batches that small.
>

So your objection here is that one needs additional resources to do
excersise their rights. Well, what about spending time and money to
get education to be able to do programming work? Being able to
understand C and hardware, etc is also an additional restriction
imposed on an average person. Do you advocate that every copy of GPL
program should be accompanied with an engineer who would explain how
it all works?

-- 
Dmitry

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Dave Neuer]

On 6/14/07, Dmitry Torokhov <dmitry.torokhov@gmail.com> wrote:
> On 6/14/07, Dave Neuer <mr.fred.smoothie@pobox.com> wrote:
> > On 6/14/07, Lennart Sorensen <lsorense@csclub.uwaterloo.ca> wrote:
> > > Nothing prevents you from taking tivos kernel
> > > changes and building your own hardware to run that code on, and as such
> > > the spirit of the GPL v2 seems fulfilled.
> >
> > Oh, come on: you're not serious, right? Something indeed prevents me
> > -- the fact that I'm not a hardware manufacturer, I don't have fabs,
> > outsource vendors to provide me w/ designs, ASICs, etc. Nor to I have
> > the money to pay one-off prices for various components if they're even
> > available in batches that small.
> >
>
> So your objection here is that one needs additional resources to do
> excersise their rights. Well, what about spending time and money to
> get education to be able to do programming work?

Come on, again w/ the bullshit. TiVO does not try to prevent me from
getting a CS degree, or buying a C reference. They _do_ prevent me
from running modified code on my TiVO box.

> Being able to
> understand C and hardware, etc is also an additional restriction
> imposed on an average person.

Not imposed by TiVO.

> Do you advocate that every copy of GPL
> program should be accompanied with an engineer who would explain how
> it all works?

No, just that hardware vendors not lock me out of _my_ hardware if
they've benefitted from code which was intended to be modifiable by
end users.

Dave

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bron Gondwana]

On Thu, Jun 14, 2007 at 05:25:19PM -0400, Dmitry Torokhov wrote:
> On 6/14/07, Dave Neuer <mr.fred.smoothie@pobox.com> wrote:
>> On 6/14/07, Lennart Sorensen <lsorense@csclub.uwaterloo.ca> wrote:
>> > Nothing prevents you from taking tivos kernel
>> > changes and building your own hardware to run that code on, and as such
>> > the spirit of the GPL v2 seems fulfilled.
>>
>> Oh, come on: you're not serious, right? Something indeed prevents me
>> -- the fact that I'm not a hardware manufacturer, I don't have fabs,
>> outsource vendors to provide me w/ designs, ASICs, etc. Nor to I have
>> the money to pay one-off prices for various components if they're even
>> available in batches that small.
>>
>
> So your objection here is that one needs additional resources to do
> excersise their rights. Well, what about spending time and money to
> get education to be able to do programming work? Being able to
> understand C and hardware, etc is also an additional restriction
> imposed on an average person. Do you advocate that every copy of GPL
> program should be accompanied with an engineer who would explain how
> it all works?

Yes please.  Can she be spunky as well?  ta.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Manu Abraham]

Lennart Sorensen wrote:

> Well much as I don't like what Tivo did with only allowing signed
> kernels to run, I don't see anything in the above that says they can't

Well, it is not Tivo alone -- look at http://aminocom.com/ for an
example. If you want the kernel sources pay USD 50k and we will provide
the kernel sources, was their attitude.

> do that.  They let you have the code and make changes to it, they just
> don't let you put that changed stuff on the device they build.  The
> software is free, even though the hardware is locked down.  The GPL v3
> really seems to change the spirit to try and cover usage and hardware
> behaviour, while the spirit of the GPL v2 seemed to me at least to
> simply be to allow people to copy and change and use the code, and pass
> that on to people.  It didn't have anything to do with what they did
> with it on hardware.  Nothing prevents you from taking tivos kernel
> changes and building your own hardware to run that code on, and as such
> the spirit of the GPL v2 seems fulfilled.  It covers freedom of the
> source code and resulting binaries, not of the platform you run it on.
> The GPL v3 has a much broader coverage of what it wants to control,
> which to me means the spirit is different.
> 
> I don't have a tivo, I use mythtv on my own PC.  Tivo doesn't force you
> to buy their hardware after all.

Well, it is not Tivo alone, a large chunk of the vendors do that. The
vendors who actually do it the clean way are just few and can be counted
very easily.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alan Cox]

> Well, it is not Tivo alone -- look at http://aminocom.com/ for an
> example. If you want the kernel sources pay USD 50k and we will provide
> the kernel sources, was their attitude.

GPLv2 deals with that case, and they can (and should) be sued for it
[except that US copyright law is designed for large music companies not
people]

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Manu Abraham]

Alan Cox wrote:
>> Well, it is not Tivo alone -- look at http://aminocom.com/ for an
>> example. If you want the kernel sources pay USD 50k and we will provide
>> the kernel sources, was their attitude.
> 
> GPLv2 deals with that case, and they can (and should) be sued for it
> [except that US copyright law is designed for large music companies not
> people]


Their argument was that the mentioned sources contain propreitary closed
stuff from IBM/AMCC for the PPC 405/440 and or for the NXP (MIPS based)
chips. Even if the GPLv2 deals with it, well haven't reached anywhere
with it, inspite of talks with them. So most of the users just probably
stopped talking sense with them, just like me.

Have some of those Amino STB's, the software on it being buggy,
including myself many others wanted to fix those bugs, but then people
had to pay for their annual support to get the fixes. People who were
able to fix also were denied the same since there is no source
available. But if you wanted the sources, then you pay for the sources.
If you don't pay for their sources, then pay for their
Bronze/Silver/Gold Support schemes, where people pay through their nose.

For a specific case with which i wanted to attach a USB based device to
the box, they stated: we can port in a driver that which exists in the
vanilla kernel, to their device but just that they need to be paid for
that to be done, eventhough if someone else was willing to do that job,
but that wasn't possible because of no sources.

In either way, if you buy their devices, it is just that, you keep
paying us, if you want your device your work as expected.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 20, 2007, Manu Abraham <abraham.manu@gmail.com> wrote:

> Alan Cox wrote:
>>> Well, it is not Tivo alone -- look at http://aminocom.com/ for an
>>> example. If you want the kernel sources pay USD 50k and we will provide
>>> the kernel sources, was their attitude.
>> 
>> GPLv2 deals with that case, and they can (and should) be sued for it
>> [except that US copyright law is designed for large music companies not
>> people]

> Their argument was that the mentioned sources contain propreitary
> closed stuff from IBM/AMCC for the PPC 405/440 and or for the NXP
> (MIPS based) chips.

As you probably know, this is not a valid excuse to distribute the
software under conditions that disrespect its license.

It doesn't mean you can force them to give you the source code, it
only means the copyright holder can stop them from distributing the
software this way.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Lennart Sorensen]

On Tue, Jun 19, 2007 at 07:28:22PM +0400, Manu Abraham wrote:
> Well, it is not Tivo alone -- look at http://aminocom.com/ for an
> example. If you want the kernel sources pay USD 50k and we will provide
> the kernel sources, was their attitude.

Hmm, set top boxes are often rented from the cable company rather than
sold.  Stupid grey area for sure.  At least tivo does give you the
sources, without demanding more money.  Rather big difference.

> Well, it is not Tivo alone, a large chunk of the vendors do that. The
> vendors who actually do it the clean way are just few and can be counted
> very easily.

Well at least where I work we don't try to lock down the hardware, we do
contribute our changes and bug fixes to upstream when it makes sense
(and where our changes wouldn't make sense for upstream, they are still
clearly included with the sources we have.)  If a customer wants a copy
of the sources, they will get a nice DVD, although strangely none have
asked for one yet.

--
Len Sorensen

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Manu Abraham]

Lennart Sorensen wrote:
> On Tue, Jun 19, 2007 at 07:28:22PM +0400, Manu Abraham wrote:
>> Well, it is not Tivo alone -- look at http://aminocom.com/ for an
>> example. If you want the kernel sources pay USD 50k and we will provide
>> the kernel sources, was their attitude.
> 
> Hmm, set top boxes are often rented from the cable company rather than
> sold.  Stupid grey area for sure.  At least tivo does give you the
> sources, without demanding more money.  Rather big difference.

I am not talking about the rented aspect, since these STB's are usually
sold rather than rented.

>> Well, it is not Tivo alone, a large chunk of the vendors do that. The
>> vendors who actually do it the clean way are just few and can be counted
>> very easily.
> 
> Well at least where I work we don't try to lock down the hardware, we do
> contribute our changes and bug fixes to upstream when it makes sense
> (and where our changes wouldn't make sense for upstream, they are still
> clearly included with the sources we have.)  If a customer wants a copy
> of the sources, they will get a nice DVD, although strangely none have
> asked for one yet.


Providing the changes back itself is a great thing altogether.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Lennart Sorensen]

On Thu, Jun 21, 2007 at 10:56:33AM +0400, Manu Abraham wrote:
> Providing the changes back itself is a great thing altogether.

It also makes sense.  If the changes are accepted back, the community at
large will keep the changes maintained.  Less work for me to do when
going to newer code versions later.  And even better, it may help
someone else out too.

A company is likely to like the reduced maintenance burden part, but I
think the other part is even better.  After all we saved time not having
to write everything our selves, so helping others save time seems only
fair.  The GPL may only require giving the sources to the people who
buys the product, but there isn't really any benefit to us in doing only
that.

--
Len Sorensen

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Wed, 13 Jun 2007, Alexandre Oliva wrote:
> 
> Look, there was room for misunderstandings in earlier drafts of the
> license.  Based on the public comments, the wording was improved.  I'd
> like to think the issues that arose from misunderstandings of the
> earlier drafts are no longer an issue.  Is it not so?

No. The anti-DRM language is still there, and no, it was never a 
misunderstanding. Now it's been limited to "consumer devices" (after I 
pointed out some of the _obvious_ problems with the original language), 
and the only people who called anything a "misunderstanding" were the ones 
that tried to point to *other* points in the license altogether (ie there 
was also a "drm section", which didn't really seem to say anything much at 
all).

Rms calls it "tivoization", but that's a word he has made up, and a term I 
find offensive, so I don't choose to use it. It's offensive because Tivo 
never did anything wrong, and the FSF even acknowledged that. The fact 
that they do their hardware and have some DRM issues with the content 
producers and thus want to protect the integrity of that hardware.

The kernel license covers the *kernel*. It does not cover boot loaders and 
hardware, and as far as I'm concerned, people who make their own hardware 
can design them any which way they want. Whether that means "booting only 
a specific kernel" or "sharks with lasers", I don't care.

> Keeping on making false claims about the license drafts can be one of
> two things: misunderstandings, out of ambiguity in the text or
> preconceptions, or ill intentions.  I'd rather believe it's the
> former.

No, it was not the former. And I think the whole "the kernel developers 
misunderstand the license" crap that the FSF was saying (several times) 
was very trying to confuse the issue: the FSF knew damn well which part of 
the license was obnoxious, they just tried to confuse the issue by 
pointing to *another* part of the license.

And you're just parrotting their idiotic line.

> Now, of course you can look at the licenses and decide that you never
> agreed with the spirit of the GPL in the first place, and that GPLv2
> models better your intentions than GPLv3.

And this is again the same *disease*. You claim that I "misunderstood" the 
"spirit of the GPL".

Dammit, the GPL is a license. I understand it quite well. Probably better 
than most. The fact that the FSF then noticed that there were *other* 
things that they wanted to do, and that were *not* covered by the GPLv2, 
does *not* mean that they can claim that others "misunderstood" the 
license.

I understood it perfectly fine, and it fit my needs. So tell me: who is 
the more confused one: the one who chose the license fifteen years ago, 
and realized what it means legally, and still stands behind it? I don't 
think so.

> Your assessment about sharing of code between Linux and OpenSolaris
> very much makes it seem like that the spirit of sharing, of letting
> others run, study, modify and share the code as long as they respect
> others' freedoms, has never been what moved you.  Rather, you seem to
> perceive the GPL as demanding some form of payback, of contribution,
> rather than the respect for others' freedoms that it requires.  In
> fact, you said something along these lines yourself many months ago.

I have said *exactly* that many many times.

The beauty of the GPLv2 is exactly that it's a "tit-for-tat" license, and 
you can use it without having to drink the kool-aid.

I've said that over and over again. It's the "spirit of the GPLv2". It's 
what has made it such a great license, that lots of people (and companies) 
can use, is very fundamentally that it's fair.

The fact that the FSF sees *another* spirit to it is absolutely not a 
reason to say that I'm "confused". Quite frankly, apparently I'm _less_ 
confused than they are, since I saw the GPLv2 for what it was, and they 
did not - and as a result they felt they needed to extend upon it, because 
the license didn't actually match what they thought it would do.

> In fact, the spirit has always been described in its preamble, and it 
> didn't change at all: it's all about respecting others' freedoms.

That's a lot of bullshit. You are apparently the grand poobah, and can 
decide _which_ freedoms and for _what_ others' that matter.

I respect peoples freedoms too. I just disagree with the FSF on what that
slippery word means.

The fact that you are unable to even apparently fathom this fundamental 
issue, and that the FSF thinks that they own the definition of "freedom" 
is _your_ problem.

You're acting like some Alice-in-Wonderland character, saying that your 
definition of words is the only one that matter. And that others are 
"confused". Read up on your humpty-dumpty some day. 

I'm damn fed up with the FSF being the "protector of freedoms", and also 
feeling that they can define what those freedoms mean.

The GPLv2 is a *legal*license*. And no, the FSF doesn't get to define what 
the words mean to suit their agenda.

		Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alan Cox]

> find offensive, so I don't choose to use it. It's offensive because Tivo 
> never did anything wrong, and the FSF even acknowledged that. The fact 

Not all of us agree with this for the benefit of future legal
interpretation.

> The GPLv2 is a *legal*license*. And no, the FSF doesn't get to define what 
> the words mean to suit their agenda.

Agreed - everyone contributed to the kernel based upon the GPLv2. Lots of
different reasons, lots of different viewpoints about GPL2 v  GPL3, DRM ,
Treacherous Computing, etc. The commonality is not political, not a
grand plan, not a grand unified social agenda but a bunch of people for
whom the GPLv2 was an acceptable license for furthering their intentions
whether that is education for all, a shared commons or just making a
quick buck

Alan

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Wed, 13 Jun 2007, Alan Cox wrote:
>
> > find offensive, so I don't choose to use it. It's offensive because Tivo 
> > never did anything wrong, and the FSF even acknowledged that. The fact 
> 
> Not all of us agree with this for the benefit of future legal
> interpretation.

Well, even the FSF lawyers did, but one of the reasons I never wanted to 
do the copyright assignments(*) is exactly because I think people need to 
make their own judgments on what the GPLv2 means. In the end, the only 
thing that really matters is what a judge says (after appeals etc), and 
the fact is, any license will always have gray areas where people disagree 
about interpretation.

And I actually am of the very firm opinion that a world with gray areas 
(and purple, and pink, and green) is a hell of a lot better than one where 
everything is black-and-white. Only lawyers want a black-and-white world.

So I would actually *encourage* other people to sue over their GPLv2 
interpretations, as they have done in Germany (and as IBM has done in the 
US). I'd sue based on _my_ reading of it, but hey, while my opinion is 
obviously always correct, I recognize that I live in a world where not 
everybody else always sees that.

[ (*) Obviously, the *biggest* reason not to do copyright assignments is 
  that they are just a total pain in the ass to do, and cause tons of 
  totally pointless paperwork. So "Linus is lazy and not interested in 
  being a lawyer" is obviously the primary reason for the lack of 
  assignments. I'm just much happier with people owning their own code 
  outright. ]

Of course, I also realize that suing people over license violations is a 
big pain in the ass, and in that sense while I "encourage" people to 
assert their own copyrights, I would obviously also say that it's almost 
certainly not worth doing if it's in a "gray" area. But that, in the end, 
has to be the copyright owners own decision!

> > The GPLv2 is a *legal*license*. And no, the FSF doesn't get to define what 
> > the words mean to suit their agenda.
> 
> Agreed - everyone contributed to the kernel based upon the GPLv2. Lots of
> different reasons, lots of different viewpoints about GPL2 v  GPL3, DRM ,
> Treacherous Computing, etc. The commonality is not political, not a
> grand plan, not a grand unified social agenda but a bunch of people for
> whom the GPLv2 was an acceptable license for furthering their intentions
> whether that is education for all, a shared commons or just making a
> quick buck

Indeed. And it's _fine_ to even be in it "just to make a quick buck". We 
do want all kinds of input. I think the community is much healthier having 
lots of different reasons for people wanting to be involved, rather than 
concentrating on just some specific reason.

For some it's the technology. For some it's the license. For some it's 
just a thing to pass boredom. Others like to learn. Whatever. It's all 
good!

		Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 13, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:

> On Wed, 13 Jun 2007, Alan Cox wrote:

>> > find offensive, so I don't choose to use it. It's offensive because Tivo 
>> > never did anything wrong, and the FSF even acknowledged that. The fact 

>> Not all of us agree with this for the benefit of future legal
>> interpretation.

> Well, even the FSF lawyers did,

Or rather they didn't think an attempt to enforce that in the US would
prevail (or so I'm told).  That's not saying what TiVo did was right,
and that's not saying that what TiVo did was permitted by the license.
Only courts of law can do that.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Wednesday 13 June 2007 19:15:42 Alexandre Oliva wrote:
> On Jun 13, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:
> > On Wed, 13 Jun 2007, Alan Cox wrote:
> >> > find offensive, so I don't choose to use it. It's offensive because
> >> > Tivo never did anything wrong, and the FSF even acknowledged that. The
> >> > fact
> >>
> >> Not all of us agree with this for the benefit of future legal
> >> interpretation.
> >
> > Well, even the FSF lawyers did,
>
> Or rather they didn't think an attempt to enforce that in the US would
> prevail (or so I'm told).  That's not saying what TiVo did was right,
> and that's not saying that what TiVo did was permitted by the license.
> Only courts of law can do that.

Wrong! Anyone with half a brain can make the distinction. What TiVO did is 
entirely legal - they fully complied with the GPLv2. Note that what they 
*DON'T* allow people to do is run whatever version of whatever software they 
want on their hardware. They have that right - its the "Free Software 
Foundation" and the GPL - regardless of version - is a *SOFTWARE* license. 
TiVO never stopped people from copying, modifying or distributing the code - 
what they did was say "The code is GPL'd, the hardware is restricted" - 
ie: "You can do what you want with the code, but you can only run compiled 
version of it that we provide on our hardware". Why is that legal? Because 
TiVO produces the hardware and sells it to you with a certain *LICENSE* - 
because it does contain hardware covered under any number of patents. That 
license grants you the right to use the patents - in this case algorithms - 
provided you comply with the terms of the license. (Just like the GPL gives 
you the right to copy, modify and distribute GPL'd code as long as you comply 
with its terms)

If you believe otherwise then you are sadly mistaken. Now stop parroting the 
FSF's worn and tired tripe.

DRH
PS: Looking at your .sig I guess maybe you can't do that without getting 
kicked out of the FSF-LA

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Adrian Bunk]

On Wed, Jun 13, 2007 at 07:46:15PM -0400, Daniel Hazelton wrote:
> On Wednesday 13 June 2007 19:15:42 Alexandre Oliva wrote:
> > On Jun 13, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:
> > > On Wed, 13 Jun 2007, Alan Cox wrote:
> > >> > find offensive, so I don't choose to use it. It's offensive because
> > >> > Tivo never did anything wrong, and the FSF even acknowledged that. The
> > >> > fact
> > >>
> > >> Not all of us agree with this for the benefit of future legal
> > >> interpretation.
> > >
> > > Well, even the FSF lawyers did,
> >
> > Or rather they didn't think an attempt to enforce that in the US would
> > prevail (or so I'm told).  That's not saying what TiVo did was right,
> > and that's not saying that what TiVo did was permitted by the license.
> > Only courts of law can do that.
> 
> Wrong! Anyone with half a brain can make the distinction. What TiVO did is 
> entirely legal - they fully complied with the GPLv2. Note that what they 
> *DON'T* allow people to do is run whatever version of whatever software they 
> want on their hardware. They have that right - its the "Free Software 
> Foundation" and the GPL - regardless of version - is a *SOFTWARE* license. 
>...

The GPLv2 says:

"For an executable work, complete source code means all the source code 
for all modules it contains, plus any associated interface definition 
files, plus the scripts used to control compilation and installation of 
the executable."

The question is whether this includes private keys.
Different people have different opinions regarding this issue.

If "the complete source code" includes private keys, the GPLv2 requires 
them to give any costumer the private keys.

Fact is that Harald Welte did in several cases successfully convince 
vendors that private keys are part of the source code if they are 
required for running the compiled binary on some hardware.

AFAIK there haven't been any court rulings on this issue, and it could 
even be that courts in different countries will decide differently.

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Wednesday 13 June 2007 20:44:19 Adrian Bunk wrote:
> On Wed, Jun 13, 2007 at 07:46:15PM -0400, Daniel Hazelton wrote:
> > On Wednesday 13 June 2007 19:15:42 Alexandre Oliva wrote:
> > > On Jun 13, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:
> > > > On Wed, 13 Jun 2007, Alan Cox wrote:
> > > >> > find offensive, so I don't choose to use it. It's offensive
> > > >> > because Tivo never did anything wrong, and the FSF even
> > > >> > acknowledged that. The fact
> > > >>
> > > >> Not all of us agree with this for the benefit of future legal
> > > >> interpretation.
> > > >
> > > > Well, even the FSF lawyers did,
> > >
> > > Or rather they didn't think an attempt to enforce that in the US would
> > > prevail (or so I'm told).  That's not saying what TiVo did was right,
> > > and that's not saying that what TiVo did was permitted by the license.
> > > Only courts of law can do that.
> >
> > Wrong! Anyone with half a brain can make the distinction. What TiVO did
> > is entirely legal - they fully complied with the GPLv2. Note that what
> > they *DON'T* allow people to do is run whatever version of whatever
> > software they want on their hardware. They have that right - its the
> > "Free Software Foundation" and the GPL - regardless of version - is a
> > *SOFTWARE* license. ...
>
> The GPLv2 says:
>
> "For an executable work, complete source code means all the source code
> for all modules it contains, plus any associated interface definition
> files, plus the scripts used to control compilation and installation of
> the executable."
>
> The question is whether this includes private keys.
> Different people have different opinions regarding this issue.
>
> If "the complete source code" includes private keys, the GPLv2 requires
> them to give any costumer the private keys.
>
> Fact is that Harald Welte did in several cases successfully convince
> vendors that private keys are part of the source code if they are
> required for running the compiled binary on some hardware.

If the hardware was designed for the end-user to change the software running 
on it - including running software that it was never meant to run (ie: a 
complete webserver on cell phone) - then yes, the signing keys are a part of 
the source, as the software running on the device is designed to be updated 
by the user using the provided system.

If, on the other hand, the only "software updates" the user is expected to 
perform are the installation of newer versions of the existing code 
for "Security" or "Bug Fix" reasons then the signing keys aren't part of the 
source.

I haven't looked into what Harald Welte did, but I'd be surprised if someone 
tried following suit in America and had as much success.

>
> AFAIK there haven't been any court rulings on this issue, and it could
> even be that courts in different countries will decide differently.

Agreed.

DRH

>
> cu
> Adrian



-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Adrian Bunk]

On Wed, Jun 13, 2007 at 09:01:28PM -0400, Daniel Hazelton wrote:
> On Wednesday 13 June 2007 20:44:19 Adrian Bunk wrote:
> > On Wed, Jun 13, 2007 at 07:46:15PM -0400, Daniel Hazelton wrote:
> > > On Wednesday 13 June 2007 19:15:42 Alexandre Oliva wrote:
> > > > On Jun 13, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:
> > > > > On Wed, 13 Jun 2007, Alan Cox wrote:
> > > > >> > find offensive, so I don't choose to use it. It's offensive
> > > > >> > because Tivo never did anything wrong, and the FSF even
> > > > >> > acknowledged that. The fact
> > > > >>
> > > > >> Not all of us agree with this for the benefit of future legal
> > > > >> interpretation.
> > > > >
> > > > > Well, even the FSF lawyers did,
> > > >
> > > > Or rather they didn't think an attempt to enforce that in the US would
> > > > prevail (or so I'm told).  That's not saying what TiVo did was right,
> > > > and that's not saying that what TiVo did was permitted by the license.
> > > > Only courts of law can do that.
> > >
> > > Wrong! Anyone with half a brain can make the distinction. What TiVO did
> > > is entirely legal - they fully complied with the GPLv2. Note that what
> > > they *DON'T* allow people to do is run whatever version of whatever
> > > software they want on their hardware. They have that right - its the
> > > "Free Software Foundation" and the GPL - regardless of version - is a
> > > *SOFTWARE* license. ...
> >
> > The GPLv2 says:
> >
> > "For an executable work, complete source code means all the source code
> > for all modules it contains, plus any associated interface definition
> > files, plus the scripts used to control compilation and installation of
> > the executable."
> >
> > The question is whether this includes private keys.
> > Different people have different opinions regarding this issue.
> >
> > If "the complete source code" includes private keys, the GPLv2 requires
> > them to give any costumer the private keys.
> >
> > Fact is that Harald Welte did in several cases successfully convince
> > vendors that private keys are part of the source code if they are
> > required for running the compiled binary on some hardware.
> 
> If the hardware was designed for the end-user to change the software running 
> on it - including running software that it was never meant to run (ie: a 
> complete webserver on cell phone) - then yes, the signing keys are a part of 
> the source, as the software running on the device is designed to be updated 
> by the user using the provided system.
> 
> If, on the other hand, the only "software updates" the user is expected to 
> perform are the installation of newer versions of the existing code 
> for "Security" or "Bug Fix" reasons then the signing keys aren't part of the 
> source.

Are you an idiot, or do you just choose to ignore all proof that doesn't 
fit your preconceived beliefs?

The GPL doesn't give someone distributing the software the choice of how
much to limit the freedom of the user.

Either private keys required to run the kernel on the hardware are 
always considered part of "the complete source code" or they are never 
part of it.

> I haven't looked into what Harald Welte did, but I'd be surprised if someone 
> tried following suit in America and had as much success.
>...

Harald is in Germany, and he therefore takes legal action against people 
distributing products violating his copyright on the Linux kernel
in Germany at German courts based on German laws.

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Wednesday 13 June 2007 21:24:01 Adrian Bunk wrote:
> On Wed, Jun 13, 2007 at 09:01:28PM -0400, Daniel Hazelton wrote:
> > On Wednesday 13 June 2007 20:44:19 Adrian Bunk wrote:
> > > On Wed, Jun 13, 2007 at 07:46:15PM -0400, Daniel Hazelton wrote:
> > > > On Wednesday 13 June 2007 19:15:42 Alexandre Oliva wrote:
> > > > > On Jun 13, 2007, Linus Torvalds <torvalds@linux-foundation.org> 
wrote:
> > > > > > On Wed, 13 Jun 2007, Alan Cox wrote:
> > > > > >> > find offensive, so I don't choose to use it. It's offensive
> > > > > >> > because Tivo never did anything wrong, and the FSF even
> > > > > >> > acknowledged that. The fact
> > > > > >>
> > > > > >> Not all of us agree with this for the benefit of future legal
> > > > > >> interpretation.
> > > > > >
> > > > > > Well, even the FSF lawyers did,
> > > > >
> > > > > Or rather they didn't think an attempt to enforce that in the US
> > > > > would prevail (or so I'm told).  That's not saying what TiVo did
> > > > > was right, and that's not saying that what TiVo did was permitted
> > > > > by the license. Only courts of law can do that.
> > > >
> > > > Wrong! Anyone with half a brain can make the distinction. What TiVO
> > > > did is entirely legal - they fully complied with the GPLv2. Note that
> > > > what they *DON'T* allow people to do is run whatever version of
> > > > whatever software they want on their hardware. They have that right -
> > > > its the "Free Software Foundation" and the GPL - regardless of
> > > > version - is a *SOFTWARE* license. ...
> > >
> > > The GPLv2 says:
> > >
> > > "For an executable work, complete source code means all the source code
> > > for all modules it contains, plus any associated interface definition
> > > files, plus the scripts used to control compilation and installation of
> > > the executable."
> > >
> > > The question is whether this includes private keys.
> > > Different people have different opinions regarding this issue.
> > >
> > > If "the complete source code" includes private keys, the GPLv2 requires
> > > them to give any costumer the private keys.
> > >
> > > Fact is that Harald Welte did in several cases successfully convince
> > > vendors that private keys are part of the source code if they are
> > > required for running the compiled binary on some hardware.
> >
> > If the hardware was designed for the end-user to change the software
> > running on it - including running software that it was never meant to run
> > (ie: a complete webserver on cell phone) - then yes, the signing keys are
> > a part of the source, as the software running on the device is designed
> > to be updated by the user using the provided system.
> >
> > If, on the other hand, the only "software updates" the user is expected
> > to perform are the installation of newer versions of the existing code
> > for "Security" or "Bug Fix" reasons then the signing keys aren't part of
> > the source.
>
> Are you an idiot, or do you just choose to ignore all proof that doesn't
> fit your preconceived beliefs?

Nope. Merely stating a distinction. Either a device is distributed, like the 
common PC, that is designed for the user to change and update the software 
on, or, like the PS2 it isn't designed for that. If I find a way to update my 
PS2 to run Linux and find that it doesn't want to start the "Linux Firmware" 
because I'm lacking a signing key...

In the case of a device that internally runs Linux (or any other GPL'd 
software) and wasn't designed for the end-user to change the software running 
on it then the signing keys aren't part of the source. OTOH, if I sell a PC 
running Linux that requires the kernel be signed then the signing keys *are* 
part of the source, since a PC is designed for the end-user to change the 
software running on it.

BTW, nice use of irony with that line. Makes me regret letting my fingers get 
ahead of my brain.

> The GPL doesn't give someone distributing the software the choice of how
> much to limit the freedom of the user.

Never claimed it did. I just wasn't as specific as I should have been when 
giving my examples.

> Either private keys required to run the kernel on the hardware are
> always considered part of "the complete source code" or they are never
> part of it.

No. It all depends on the use-case. If the hardware is designed for the user 
to install their own, custom versions of the code on then the signing keys 
are part of the source as defined by the GPLv2.

If, OTOH, the hardware was never meant for the end-user to install custom 
versions of the software on, then while the signing keys are still 
*technically* part of the source, in practice they are not. Why? Because in 
most of those cases the end-user isn't granted the right to install and run 
custom binaries on the hardware. If the manufacturer provided the signing 
keys they'd be facilitating the commission of a crime. (call it "Breach of 
Contract")

> > I haven't looked into what Harald Welte did, but I'd be surprised if
> > someone tried following suit in America and had as much success.
> >...
>
> Harald is in Germany, and he therefore takes legal action against people
> distributing products violating his copyright on the Linux kernel
> in Germany at German courts based on German laws.

I know this. As I said, I doubt that anyone who tried this in America would 
have the success he has had.

DRH

> cu
> Adrian



-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Adrian Bunk]

On Wed, Jun 13, 2007 at 09:40:13PM -0400, Daniel Hazelton wrote:
> On Wednesday 13 June 2007 21:24:01 Adrian Bunk wrote:
> > On Wed, Jun 13, 2007 at 09:01:28PM -0400, Daniel Hazelton wrote:
> > > On Wednesday 13 June 2007 20:44:19 Adrian Bunk wrote:
> > > > On Wed, Jun 13, 2007 at 07:46:15PM -0400, Daniel Hazelton wrote:
> > > > > On Wednesday 13 June 2007 19:15:42 Alexandre Oliva wrote:
> > > > > > On Jun 13, 2007, Linus Torvalds <torvalds@linux-foundation.org> 
> wrote:
> > > > > > > On Wed, 13 Jun 2007, Alan Cox wrote:
> > > > > > >> > find offensive, so I don't choose to use it. It's offensive
> > > > > > >> > because Tivo never did anything wrong, and the FSF even
> > > > > > >> > acknowledged that. The fact
> > > > > > >>
> > > > > > >> Not all of us agree with this for the benefit of future legal
> > > > > > >> interpretation.
> > > > > > >
> > > > > > > Well, even the FSF lawyers did,
> > > > > >
> > > > > > Or rather they didn't think an attempt to enforce that in the US
> > > > > > would prevail (or so I'm told).  That's not saying what TiVo did
> > > > > > was right, and that's not saying that what TiVo did was permitted
> > > > > > by the license. Only courts of law can do that.
> > > > >
> > > > > Wrong! Anyone with half a brain can make the distinction. What TiVO
> > > > > did is entirely legal - they fully complied with the GPLv2. Note that
> > > > > what they *DON'T* allow people to do is run whatever version of
> > > > > whatever software they want on their hardware. They have that right -
> > > > > its the "Free Software Foundation" and the GPL - regardless of
> > > > > version - is a *SOFTWARE* license. ...
> > > >
> > > > The GPLv2 says:
> > > >
> > > > "For an executable work, complete source code means all the source code
> > > > for all modules it contains, plus any associated interface definition
> > > > files, plus the scripts used to control compilation and installation of
> > > > the executable."
> > > >
> > > > The question is whether this includes private keys.
> > > > Different people have different opinions regarding this issue.
> > > >
> > > > If "the complete source code" includes private keys, the GPLv2 requires
> > > > them to give any costumer the private keys.
> > > >
> > > > Fact is that Harald Welte did in several cases successfully convince
> > > > vendors that private keys are part of the source code if they are
> > > > required for running the compiled binary on some hardware.
> > >
> > > If the hardware was designed for the end-user to change the software
> > > running on it - including running software that it was never meant to run
> > > (ie: a complete webserver on cell phone) - then yes, the signing keys are
> > > a part of the source, as the software running on the device is designed
> > > to be updated by the user using the provided system.
> > >
> > > If, on the other hand, the only "software updates" the user is expected
> > > to perform are the installation of newer versions of the existing code
> > > for "Security" or "Bug Fix" reasons then the signing keys aren't part of
> > > the source.
> >
> > Are you an idiot, or do you just choose to ignore all proof that doesn't
> > fit your preconceived beliefs?
> 
> Nope. Merely stating a distinction. Either a device is distributed, like the 
> common PC, that is designed for the user to change and update the software 
> on, or, like the PS2 it isn't designed for that. If I find a way to update my 
> PS2 to run Linux and find that it doesn't want to start the "Linux Firmware" 
> because I'm lacking a signing key...
> 
> In the case of a device that internally runs Linux (or any other GPL'd 
> software) and wasn't designed for the end-user to change the software running 
> on it then the signing keys aren't part of the source. OTOH, if I sell a PC 
> running Linux that requires the kernel be signed then the signing keys *are* 
> part of the source, since a PC is designed for the end-user to change the 
> software running on it.
> 
> BTW, nice use of irony with that line. Makes me regret letting my fingers get 
> ahead of my brain.
> 
> > The GPL doesn't give someone distributing the software the choice of how
> > much to limit the freedom of the user.
> 
> Never claimed it did. I just wasn't as specific as I should have been when 
> giving my examples.
> 
> > Either private keys required to run the kernel on the hardware are
> > always considered part of "the complete source code" or they are never
> > part of it.
> 
> No. It all depends on the use-case. If the hardware is designed for the user 
> to install their own, custom versions of the code on then the signing keys 
> are part of the source as defined by the GPLv2.
> 
> If, OTOH, the hardware was never meant for the end-user to install custom 
> versions of the software on, then while the signing keys are still 
> *technically* part of the source, in practice they are not. Why? Because in 
> most of those cases the end-user isn't granted the right to install and run 
> custom binaries on the hardware. If the manufacturer provided the signing 
> keys they'd be facilitating the commission of a crime. (call it "Breach of 
> Contract")
>...

Repetition doesn't let wrong things become true.

Where does the GPLv2 talk about the distinction you are trying to make 
based on distributor intentions?     

We are talking about the GPLv2 licence text, not about what you would
personally prefer.

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Wednesday 13 June 2007 22:08:27 Adrian Bunk wrote:
> On Wed, Jun 13, 2007 at 09:40:13PM -0400, Daniel Hazelton wrote:
> > On Wednesday 13 June 2007 21:24:01 Adrian Bunk wrote:
> > > On Wed, Jun 13, 2007 at 09:01:28PM -0400, Daniel Hazelton wrote:
> > > > On Wednesday 13 June 2007 20:44:19 Adrian Bunk wrote:
> > > > > On Wed, Jun 13, 2007 at 07:46:15PM -0400, Daniel Hazelton wrote:
> > > > > > On Wednesday 13 June 2007 19:15:42 Alexandre Oliva wrote:
> > > > > > > On Jun 13, 2007, Linus Torvalds <torvalds@linux-foundation.org>
> >
> > wrote:
> > > > > > > > On Wed, 13 Jun 2007, Alan Cox wrote:
> > > > > > > >> > find offensive, so I don't choose to use it. It's
> > > > > > > >> > offensive because Tivo never did anything wrong, and the
> > > > > > > >> > FSF even acknowledged that. The fact
> > > > > > > >>
> > > > > > > >> Not all of us agree with this for the benefit of future
> > > > > > > >> legal interpretation.
> > > > > > > >
> > > > > > > > Well, even the FSF lawyers did,
> > > > > > >
> > > > > > > Or rather they didn't think an attempt to enforce that in the
> > > > > > > US would prevail (or so I'm told).  That's not saying what TiVo
> > > > > > > did was right, and that's not saying that what TiVo did was
> > > > > > > permitted by the license. Only courts of law can do that.
> > > > > >
> > > > > > Wrong! Anyone with half a brain can make the distinction. What
> > > > > > TiVO did is entirely legal - they fully complied with the GPLv2.
> > > > > > Note that what they *DON'T* allow people to do is run whatever
> > > > > > version of whatever software they want on their hardware. They
> > > > > > have that right - its the "Free Software Foundation" and the GPL
> > > > > > - regardless of version - is a *SOFTWARE* license. ...
> > > > >
> > > > > The GPLv2 says:
> > > > >
> > > > > "For an executable work, complete source code means all the source
> > > > > code for all modules it contains, plus any associated interface
> > > > > definition files, plus the scripts used to control compilation and
> > > > > installation of the executable."
> > > > >
> > > > > The question is whether this includes private keys.
> > > > > Different people have different opinions regarding this issue.
> > > > >
> > > > > If "the complete source code" includes private keys, the GPLv2
> > > > > requires them to give any costumer the private keys.
> > > > >
> > > > > Fact is that Harald Welte did in several cases successfully
> > > > > convince vendors that private keys are part of the source code if
> > > > > they are required for running the compiled binary on some hardware.
> > > >
> > > > If the hardware was designed for the end-user to change the software
> > > > running on it - including running software that it was never meant to
> > > > run (ie: a complete webserver on cell phone) - then yes, the signing
> > > > keys are a part of the source, as the software running on the device
> > > > is designed to be updated by the user using the provided system.
> > > >
> > > > If, on the other hand, the only "software updates" the user is
> > > > expected to perform are the installation of newer versions of the
> > > > existing code for "Security" or "Bug Fix" reasons then the signing
> > > > keys aren't part of the source.
> > >
> > > Are you an idiot, or do you just choose to ignore all proof that
> > > doesn't fit your preconceived beliefs?
> >
> > Nope. Merely stating a distinction. Either a device is distributed, like
> > the common PC, that is designed for the user to change and update the
> > software on, or, like the PS2 it isn't designed for that. If I find a way
> > to update my PS2 to run Linux and find that it doesn't want to start the
> > "Linux Firmware" because I'm lacking a signing key...
> >
> > In the case of a device that internally runs Linux (or any other GPL'd
> > software) and wasn't designed for the end-user to change the software
> > running on it then the signing keys aren't part of the source. OTOH, if I
> > sell a PC running Linux that requires the kernel be signed then the
> > signing keys *are* part of the source, since a PC is designed for the
> > end-user to change the software running on it.
> >
> > BTW, nice use of irony with that line. Makes me regret letting my fingers
> > get ahead of my brain.
> >
> > > The GPL doesn't give someone distributing the software the choice of
> > > how much to limit the freedom of the user.
> >
> > Never claimed it did. I just wasn't as specific as I should have been
> > when giving my examples.
> >
> > > Either private keys required to run the kernel on the hardware are
> > > always considered part of "the complete source code" or they are never
> > > part of it.
> >
> > No. It all depends on the use-case. If the hardware is designed for the
> > user to install their own, custom versions of the code on then the
> > signing keys are part of the source as defined by the GPLv2.
> >
> > If, OTOH, the hardware was never meant for the end-user to install custom
> > versions of the software on, then while the signing keys are still
> > *technically* part of the source, in practice they are not. Why? Because
> > in most of those cases the end-user isn't granted the right to install
> > and run custom binaries on the hardware. If the manufacturer provided the
> > signing keys they'd be facilitating the commission of a crime. (call it
> > "Breach of Contract")
> >...
>
> Repetition doesn't let wrong things become true.
>
> Where does the GPLv2 talk about the distinction you are trying to make
> based on distributor intentions?
>
> We are talking about the GPLv2 licence text, not about what you would
> personally prefer.

The GPLv2 doesn't have to cover this distinction to make it a reality. This 
distinction is *EXACTLY* the type of distinction a lawyer will make when 
arguing the point.

Yes, it's artificial. Yes, it does appear to violate the GPLv2 - *IF* you read 
the text of such in a specific manner. 

However, the GPL, until version 3, *NEVER* guaranteed the right to run a given 
piece of software on *ANY* hardware - not the hardware it *COMES* on. 

And please, I repeated myself only because your reply seemed to imply that you 
didn't understand the statement I had made. Since you have now informed me, 
in a backhanded way, informed me that my interpretation of your response was 
wrong, I will not repeat myself again.

Also note that I have re-examined the facts, in light of new information 
presented in this discussion, and have come to the conclusion that devices 
like the TiVO, in keeping the signing keys private (because of 
the "Facilitation of a Crime" thing I noted earlier), is violating the GPL, 
but not in the manner almost everyone is arguing. The violation is, rather, 
with the clause about the license being null and void in event of laws 
impacting the delivery of the source. (Because, as I also stated earlier, the 
signing keys are part of the source. Since, in some cases, the license on the 
hardware prevents running modified binaries (a reason for the digital 
signing) companies will keep said keys private - doing otherwise can (and I 
can assure you that some lawyer will do this) be construed as "Facilitating 
the Commission of a crime". In this case, it'd be "Breach of Contract" - 
IANAL, but IIRC, licenses fall under contract law))

DRH

>
> cu
> Adrian



-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Adrian Bunk]

On Wed, Jun 13, 2007 at 10:43:14PM -0400, Daniel Hazelton wrote:
> On Wednesday 13 June 2007 22:08:27 Adrian Bunk wrote:
> > On Wed, Jun 13, 2007 at 09:40:13PM -0400, Daniel Hazelton wrote:
> > > On Wednesday 13 June 2007 21:24:01 Adrian Bunk wrote:
>...
> > > > Either private keys required to run the kernel on the hardware are
> > > > always considered part of "the complete source code" or they are never
> > > > part of it.
> > >
> > > No. It all depends on the use-case. If the hardware is designed for the
> > > user to install their own, custom versions of the code on then the
> > > signing keys are part of the source as defined by the GPLv2.
> > >
> > > If, OTOH, the hardware was never meant for the end-user to install custom
> > > versions of the software on, then while the signing keys are still
> > > *technically* part of the source, in practice they are not. Why? Because
> > > in most of those cases the end-user isn't granted the right to install
> > > and run custom binaries on the hardware. If the manufacturer provided the
> > > signing keys they'd be facilitating the commission of a crime. (call it
> > > "Breach of Contract")
> > >...
> >
> > Repetition doesn't let wrong things become true.
> >
> > Where does the GPLv2 talk about the distinction you are trying to make
> > based on distributor intentions?
> >
> > We are talking about the GPLv2 licence text, not about what you would
> > personally prefer.
> 
> The GPLv2 doesn't have to cover this distinction to make it a reality. This 
> distinction is *EXACTLY* the type of distinction a lawyer will make when 
> arguing the point.
>...

Reality check:

Harald convinced companies that they have to provide the private keys 
required to run the Linux kernel they ship on their hardware.

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Wednesday 13 June 2007 22:56:40 Adrian Bunk wrote:
> On Wed, Jun 13, 2007 at 10:43:14PM -0400, Daniel Hazelton wrote:
> > On Wednesday 13 June 2007 22:08:27 Adrian Bunk wrote:
> > > On Wed, Jun 13, 2007 at 09:40:13PM -0400, Daniel Hazelton wrote:
> > > > On Wednesday 13 June 2007 21:24:01 Adrian Bunk wrote:
> >
> >...
> >
> > > > > Either private keys required to run the kernel on the hardware are
> > > > > always considered part of "the complete source code" or they are
> > > > > never part of it.
> > > >
> > > > No. It all depends on the use-case. If the hardware is designed for
> > > > the user to install their own, custom versions of the code on then
> > > > the signing keys are part of the source as defined by the GPLv2.
> > > >
> > > > If, OTOH, the hardware was never meant for the end-user to install
> > > > custom versions of the software on, then while the signing keys are
> > > > still *technically* part of the source, in practice they are not.
> > > > Why? Because in most of those cases the end-user isn't granted the
> > > > right to install and run custom binaries on the hardware. If the
> > > > manufacturer provided the signing keys they'd be facilitating the
> > > > commission of a crime. (call it "Breach of Contract")
> > > >...
> > >
> > > Repetition doesn't let wrong things become true.
> > >
> > > Where does the GPLv2 talk about the distinction you are trying to make
> > > based on distributor intentions?
> > >
> > > We are talking about the GPLv2 licence text, not about what you would
> > > personally prefer.
> >
> > The GPLv2 doesn't have to cover this distinction to make it a reality.
> > This distinction is *EXACTLY* the type of distinction a lawyer will make
> > when arguing the point.
> >...
>
> Reality check:
>
> Harald convinced companies that they have to provide the private keys
> required to run the Linux kernel they ship on their hardware.

In Germany, not America. I should have qualified my statement to make it clear 
I mean "In America". Sorry about the confusion.

DRH

>
> cu
> Adrian



-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Michael Gerdau]

[-- Attachment #1: Type: text/plain, Size: 497 bytes --]

> In Germany, not America. I should have qualified my statement to make it clear 
> I mean "In America". Sorry about the confusion.

You shouldn't say "America" when you mean the "US".

Best wishes,
Michael
-- 
 Technosis GmbH, Geschäftsführer: Michael Gerdau, Tobias Dittmar
 Sitz Hamburg; HRB 89145 Amtsgericht Hamburg
 Vote against SPAM - see http://www.politik-digital.de/spam/
 Michael Gerdau       email: mgd@technosis.de
 GPG-keys available on request or at public keyserver

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Thursday 14 June 2007 01:39:13 Michael Gerdau wrote:
> > In Germany, not America. I should have qualified my statement to make it
> > clear I mean "In America". Sorry about the confusion.
>
> You shouldn't say "America" when you mean the "US".

Sorry, I slipped. I'm still trying to rid myself of the uniquely "US" belief 
that "America" == "USA". Thanks for the reminder.

DRH

>
> Best wishes,
> Michael



-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Valdis.Kletnieks]

[-- Attachment #1: Type: text/plain, Size: 1125 bytes --]

On Thu, 14 Jun 2007 04:56:40 +0200, Adrian Bunk said:

> Reality check:
> 
> Harald convinced companies that they have to provide the private keys 
> required to run the Linux kernel they ship on their hardware.

No, the *real* reality check:

The operative words here are "convinced companies" - as opposed to "convinced
a judge to rule that private keys are required to be disclosed". (I just
checked around on gpl-violations.org, and I don't see any news items that say
they actually generated citable case law on the topic of keys...)

Harald convinced companies that it was easier/cheaper/faster to provide the
private keys than to continue in a long legal battle with an uncertain outcome.
If the company estimates the total loss due to keys being released is US$100K,
but the costs of taking it to court are estimated at US$200K, it's obviously
a win (lesser loss, actually) for the company to just fold.

Incidentally, this same logic is what drives the average successful patent
troll lawsuit - the sued company will buy a license for $25K, just because
they know that fighting the lawsuit will cost $100K and up.



[-- Attachment #2: Type: application/pgp-signature, Size: 226 bytes --]

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Theodore Tso]

On Thu, Jun 14, 2007 at 12:00:17AM -0400, Valdis.Kletnieks@vt.edu wrote:
> Incidentally, this same logic is what drives the average successful patent
> troll lawsuit - the sued company will buy a license for $25K, just because
> they know that fighting the lawsuit will cost $100K and up.

You're off by a factor of 10-50.  The usual estimates I've heard from
people who ought to know is the minimum ante for fighting a patent
lawsuit is $1 million to $5 million.  Lawyer time and expert witness
time to give the judge a granduate education in the technologies
involved is *expensive* (since the judge may be really smart, but most
judges have no engineering background to speak of, so you have to
explain the technologies involved in terms that make sense to someone
with an honors education with a Bachelor of Arts degree).

Basically, in the US, you get the best justice money can buy.  :-)

							- Ted

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[jimmy bahuleyan]

Theodore Tso wrote:
> Basically, in the US, you get the best justice money can buy.  :-)

that has to be one of the best one-liners ever! :)

> 
> 							- Ted

-jb
-- 
Tact is the art of making a point without making an enemy.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Adrian Bunk]

On Thu, Jun 14, 2007 at 12:00:17AM -0400, Valdis.Kletnieks@vt.edu wrote:
> On Thu, 14 Jun 2007 04:56:40 +0200, Adrian Bunk said:
> 
> > Reality check:
> > 
> > Harald convinced companies that they have to provide the private keys 
> > required to run the Linux kernel they ship on their hardware.
> 
> No, the *real* reality check:
> 
> The operative words here are "convinced companies" - as opposed to "convinced
> a judge to rule that private keys are required to be disclosed". (I just
> checked around on gpl-violations.org, and I don't see any news items that say
> they actually generated citable case law on the topic of keys...)
> 
> Harald convinced companies that it was easier/cheaper/faster to provide the
> private keys than to continue in a long legal battle with an uncertain outcome.
> If the company estimates the total loss due to keys being released is US$100K,
> but the costs of taking it to court are estimated at US$200K, it's obviously
> a win (lesser loss, actually) for the company to just fold.
>...

Here in Germany, the rules at court are roughly "the loser pays 
everything including the costs of the winner", so if a big company is 
sure they will win at court there's no reason not to go there.

And if they did the effort of using private keys to only allow running 
an official firmware, they must have seen an advantage from doing so.

I'm not saying it legally clear the other way round, my statement was 
an answer to Daniel's emails claiming it was clear what such companies 
do was legal.

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Thu, 14 Jun 2007, Adrian Bunk wrote:
> 
> Here in Germany, the rules at court are roughly "the loser pays 
> everything including the costs of the winner", so if a big company is 
> sure they will win at court there's no reason not to go there.

Well, the thing is (and I've said this before), a lawsuit is (and _should_ 
be) very much a last resort.

I think that the Open Source community (and the FSF too) is much better 
off *not* concentrating so much on "legal rules" of what can and cannot be 
done, and instead spend much more effort on showing people why the whole 
"Open Source" thing actually works.

And in fact, I think that's _exactly_ what Linux has been doing for the 
last decade!

A lot of companies are actually doing the Right Thing (tm).

Not because of anybody "forcing" them, but because they have literally 
bought into the whole "Open Source can do things better" mentality. 

In fact, the whole "coercive" approach is counter-productive. It makes 
people dislike you. It makes companies _resist_ open source, rather than 
see it as a potential ally. 

And no, I'm not speaking out of my *ss. Anybody who goes back fifteen 
years and looks at how the FSF was acting wrt the GPL (v2, back then), and 
how many friends - and enemies - they were making, should see that as a 
big clue. Linux really *did* change the landscape - for the better (*). By 
being much less contrary.

So look at Intel in the open source space. They're doing well. Look at 
Sun. They aren't _forced_ to open-source, they see others open-sourcing, 
and they see that it works damn well.

In the "Tivo space", look at Neuros. 

In other words, we're just *much* better off with a friendly license and 
not trying to force people to choose sides, than with the rabid idealism 
that was - and still is - the FSF. The FSF always makes for this horrible 
"you're with us, or you're against us" black-and-white mentality, where 
there are "evil" companies (Tivo) and "good" companies (although I dunno 
if the FSF really sees anybody as truly "good").

I'd much rather just see "individuals" and "companies". They're not evil 
or good, they are all in it for their own reasons (and their reasons are 
*NOT* the same reasons they are for me, you, or anybody else), and we 
should show them that the whole "Open Source" approach really does work 
for them.

It's totally pointless to try to "force" people to be good. That's like 
"curing" gay people. Not going to happen. 

			Linus

(*) Not just Linux, of course, but I do claim that this is actually an 
area where Linux was a big influence. Not the only one, but a major 
player.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Sean]

On Thu, 14 Jun 2007 09:01:32 -0700 (PDT)
Linus Torvalds <torvalds@linux-foundation.org> wrote:

> In other words, we're just *much* better off with a friendly license and 
> not trying to force people to choose sides, than with the rabid idealism 
> that was - and still is - the FSF. The FSF always makes for this horrible 
> "you're with us, or you're against us" black-and-white mentality, where 
> there are "evil" companies (Tivo) and "good" companies (although I dunno 
> if the FSF really sees anybody as truly "good").

Linus,

If you really believe that then why didn't you choose a BSD license
for Linux?  You didn't say "completely free, no restrictions attached,
people will follow because they'll see it's best, we just won't buy
products that use Linux in a way with which we disagree".

Instead you chose a license which enforced the so called tit-for-tat
policy you think is fair.  But people who prefer the BSD license may
think you're a moron for forcing your political agenda (ie. tit-for-tat)
on users of your code.  The point of all that being, you _do_ believe
in enforcing restrictions or you wouldn't like the GPL v2.

So you draw the line of "fairness" and belief that people will
do-the-right-thing somewhere short of the BSD license.  Why is it
so hard then to accept that the FSF draws the line short of the
GPLv2 after having gained practical experience with it
since its release?

You can argue till the cows come home the belief that _your_
restrictions are more fair, moral and reasonable than theirs.
But at the end of the day it's all just a matter of opinion about
what constitutes fair and reasonable.  You think its a fair trade
that you get code back, the FSF think its fair that people can hack
and run the code anywhere its used..  It all comes down to the
author of the code getting to attach whatever restrictions they
choose.

Sean

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Thu, 14 Jun 2007, Sean wrote:
> 
> If you really believe that then why didn't you choose a BSD license
> for Linux? 

Because I think the GPLv2 is a great license.

And I don't like the FSF's radical world-view, but I am able to separate 
the license (the GPLv2) from the author and source of the license (rms and 
the FSF).

Why do people always confuse the two? The GPLv2 stands on its own. The 
fact that I disagree with the FSF on how to act has _zero_ relevance for 
my choice of license. 

The BSD license, as far as I'm concerned, is _horrible_ for any project I 
would use. I have actually released code under it, but never a "project". 
I've given some code of mine that I don't care about that much to the BSD 
projects, just because I didn't think that code really mattered, and I 
thought it would be stupid and small-minded not to let the BSD's use it.

But for a project I actually care about, I would never choose the BSD 
license. The license doesn't encode my fundamental beliefs of "fairness". 
I think the BSD license encourages a "everybody for himself" mentality, 
and doesn't encourage people to work together, and to merge.

Let me put this in source management terms, since I've also been working 
on a source control management project for the last few years: the BSD 
license encourages "branching", but the fact is, branching is not really 
all that interesting. What's interesting is "merging": the branching is 
just a largely irrelevant prerequisite to be able to merge.

The GPLv2 encourages *merging*. Again, the right to "branch" needs to be 
there in order for merges to be possible, but the right to branch is 
actually much less important than the right to "merge".

See? 

So I'm a *big* believer in the GPLv2. I think the GPLv2 is an almost 
perfect license. That doesn't mean that I have to agree with the FSF on 
everything else.

> Instead you chose a license which enforced the so called tit-for-tat
> policy you think is fair.  But people who prefer the BSD license may
> think you're a moron for forcing your political agenda (ie. tit-for-tat)
> on users of your code.

Oh, and some people did and do.

And you know what? That's PERFECTLY OK!

I think that the BSD license is wrong for me. Does that mean that people 
who choose the BSD license are wrong to do so? No. For *them* the choices 
that the BSD license makes may be the right ones!

> The point of all that being, you _do_ believe in enforcing restrictions 
> or you wouldn't like the GPL v2.

..  but I think that the software license I choose should be about the 
software, and about giving back in kind.

And the GPLv2 is _perfect_ for that.

And the GPLv3 is horrible.

And you know what? YOU can choose the GPLv3 for your projects. I'm not 
saying anything else. I'm saying that no, I was _not_ confused when I 
chose the GPLv2. I thought it was a good license 15 years ago. I thought 
it was a good license 10 years ago. I thought it was a good license five 
years ago. And I think it's a good license today.

Because it fundamnetally does what I think is fair. 

In a way that the GPLv3 DOES NOT.

		Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:

> ..  but I think that the software license I choose should be about the 
> software, and about giving back in kind.

> And the GPLv2 is _perfect_ for that.

> And the GPLv3 is horrible.

Is there anything other than TiVOization to justify these statements?


Also, can you elaborate on what you mean about 'giving back in kind'?
(I suspect this is related with the tit-for-tat reasoning, that you've
failed to elaborate on before)


The only thing the GPL demands is respect for others' freedoms, as in,
"I, the author, respect your freedoms, so you, the licensee, must
respect others' freedoms as well".  Is this the "in kind" you're
talking about?  Or are you mistaken about the actual meaning of even
GPLv2?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Thu, 14 Jun 2007, Alexandre Oliva wrote:
> 
> Is there anything other than TiVOization to justify these statements?

Do you need anything else?

But if by the question you mean "would you think the GPLv3 is fine without 
the new language in section 6 about the 'consumer devices'", then the 
answer is that yes, I think that the current GPLv3 draft looks fine apart 
from that.

> Also, can you elaborate on what you mean about 'giving back in kind'?
> (I suspect this is related with the tit-for-tat reasoning, that you've
> failed to elaborate on before)

I've *not* failed to elaborate on that before. Not at all. 

Just google for

	torvalds tit-for-tat

and you'll see a lot of my previous postings. Trying to claim that this is 
somehow "new" is ludicrous. In fact, some of the google hits you find are 
from 2004, *loong* before the current GPLv3 discussion.

So your "failed to elaborate" is not a failure on my side. 

Giving back "in kind" is obvious. I give you source code to do with as you 
see fit. I just expect you to give back in kind: source code for me to do 
with as I see fit, under the same license I gave you source code.

How hard is that to accept?

I don't ask for money. I don't ask for sexual favors. I don't ask for 
access to the hardware you design and sell. I just ask for the thing I 
gave you: source code that I can use myself.

I really don't think my "tit-for-tat" or "give back in kind" is that hard 
to understand, is it?

And no, it's not a new concept. Neither is the fact that I've never agreed 
with the FSF's agenda about "freedom" (as defined by _them_ - I have a 
notion of "freedom" myself, and the FSF doesn't get to define it for me).

I don't call Linux "Free Software". I haven't called it that for close to 
ten years! Because I think the term "Open Source" is a lot better.

> The only thing the GPL demands is respect for others' freedoms, as in,
> "I, the author, respect your freedoms, so you, the licensee, must
> respect others' freedoms as well".  Is this the "in kind" you're
> talking about?  Or are you mistaken about the actual meaning of even
> GPLv2?

I respect your freedom to design products around Linux. You can do 
whatever you damn well please - I just ask that you give the software back 
in a usable form. That's all I ask for.

And that's all the GPLv2 asks for. 

Which is why I selected the GPLv2 in the first place, and why I *still* 
think the GPLv2 is a wonderful license!

So I claim that the "freedoms" that the GPLv2 embodies are *greater* than 
the "freedoms" embodied in the GPLv3.

		Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:

> On Thu, 14 Jun 2007, Alexandre Oliva wrote:
>> 
>> Is there anything other than TiVOization to justify these statements?

> Do you need anything else?

No, I'm quite happy that this is all.

> But if by the question you mean "would you think the GPLv3 is fine without 
> the new language in section 6 about the 'consumer devices'", then the 
> answer is that yes, I think that the current GPLv3 draft looks fine apart 
> from that.

Then would you consider relicensing Linux under GPLv3 + additional
permission for Tivoization?

>> Also, can you elaborate on what you mean about 'giving back in kind'?
>> (I suspect this is related with the tit-for-tat reasoning, that you've
>> failed to elaborate on before)

> I've *not* failed to elaborate on that before. Not at all. 

> Just google for

> 	torvalds tit-for-tat

> and you'll see a lot of my previous postings. Trying to claim that this is 
> somehow "new" is ludicrous.

I didn't.  But I've provided evidence that your prior musings on this
topic were wrong.  I wanted to give you an opportunity to review your
position under this new light.  I see you haven't changed it at all.

> Giving back "in kind" is obvious. I give you source code to do with as you 
> see fit. I just expect you to give back in kind: source code for me to do 
> with as I see fit, under the same license I gave you source code.

> How hard is that to accept?

Forgive me if I find this a bit hard, because that's *not* what the
GPL says.

Where do you think the GPL say that you get the source code back?

> I don't ask for money. I don't ask for sexual favors. I don't ask for 
> access to the hardware you design and sell. I just ask for the thing I 
> gave you: source code that I can use myself.

See, that's not what the license says.

The license says what you ask for is respect for other users'
freedoms.  Nothing whatsoever for you.  Only for users.

Freedom is in "in kind" payment, and it's not even a retribution, a
payback: it's payforward, or paysideways.

Do you understand why I find your reasoning hard to accept?

> And no, it's not a new concept. Neither is the fact that I've never agreed 
> with the FSF's agenda about "freedom" (as defined by _them_ - I have a 
> notion of "freedom" myself, and the FSF doesn't get to define it for me).

We don't have to agree on our individual definitions of freedom.  But
we're talking about a specific license that assigns a specific meaning
to the term "freedoms", and that's all this is about.

> I don't call Linux "Free Software". I haven't called it that for close to 
> ten years! Because I think the term "Open Source" is a lot better.

I can appreciate that you think it's better, but unfortunately it
appears to be playing a significant role in confusing your
interpretation of the GPL.  The GPL is not just about making the
source code visible, or even modifyable by others.  It's about
respecting others' freedoms.  No matter how badly you prefer Open
Source over Free Software, how badly you'd rather disregard the
freedoms in the spirit and in the legal terms of the GPL, you chose a
license designed to protect those freedoms, not only the ability to
see and modify source code.

>> The only thing the GPL demands is respect for others' freedoms, as in,
>> "I, the author, respect your freedoms, so you, the licensee, must
>> respect others' freedoms as well".  Is this the "in kind" you're
>> talking about?  Or are you mistaken about the actual meaning of even
>> GPLv2?

> I just ask that you give the software back in a usable form. That's
> all I ask for.

I'm afraid that's not what the GPLv2 says.  There's no provision
whatsoever about giving anything back.  Not in the spirit, not in the
legal terms.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Sam Ravnborg]

On Thu, Jun 14, 2007 at 04:46:36PM -0300, Alexandre Oliva wrote:

> > Giving back "in kind" is obvious. I give you source code to do with as you 
> > see fit. I just expect you to give back in kind: source code for me to do 
> > with as I see fit, under the same license I gave you source code.
> 
> > How hard is that to accept?
> 
> Forgive me if I find this a bit hard, because that's *not* what the
> GPL says.

What part of the word "expect" did you not understand?
And whats your point here anyway?

	Sam

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Sam Ravnborg <sam@ravnborg.org> wrote:

> On Thu, Jun 14, 2007 at 04:46:36PM -0300, Alexandre Oliva wrote:
>> > Giving back "in kind" is obvious. I give you source code to do with as you 
>> > see fit. I just expect you to give back in kind: source code for me to do 
>> > with as I see fit, under the same license I gave you source code.
>> 
>> > How hard is that to accept?
>> 
>> Forgive me if I find this a bit hard, because that's *not* what the
>> GPL says.

> What part of the word "expect" did you not understand?

http://lkml.org/lkml/2006/9/24/246

  It asks everybody - regardless of circumstance - for the same thing.
  It asks for the effort that was put into improving the software to
  be given back to the common good.  You can use the end result any
  way you want (and if you want to use it for "bad" things, be my
  guest), but we ask the same exact thing of everybody - give your
  modifications back.

> And whats your point here anyway?

The the GPL doesn't do that.  It encourages that.  But what it asks
for is respect for the freedoms it defends WRT the software licensed
under it.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Sam Ravnborg]

On Thu, Jun 14, 2007 at 05:42:44PM -0300, Alexandre Oliva wrote:
> On Jun 14, 2007, Sam Ravnborg <sam@ravnborg.org> wrote:
> 
> > On Thu, Jun 14, 2007 at 04:46:36PM -0300, Alexandre Oliva wrote:
> >> > Giving back "in kind" is obvious. I give you source code to do with as you 
> >> > see fit. I just expect you to give back in kind: source code for me to do 
> >> > with as I see fit, under the same license I gave you source code.
> >> 
> >> > How hard is that to accept?
> >> 
> >> Forgive me if I find this a bit hard, because that's *not* what the
> >> GPL says.
> 
> > What part of the word "expect" did you not understand?
> 
> http://lkml.org/lkml/2006/9/24/246
> 
>   It asks everybody - regardless of circumstance - for the same thing.
>   It asks for the effort that was put into improving the software to
>   be given back to the common good.  You can use the end result any
>   way you want (and if you want to use it for "bad" things, be my
>   guest), but we ask the same exact thing of everybody - give your
>   modifications back.
> 
> > And whats your point here anyway?
> 
> The the GPL doesn't do that.  It encourages that.  But what it asks
> for is respect for the freedoms it defends WRT the software licensed
> under it.

Reading the above you are writing exact the same as Linus here
but refusing to accept it and without usign the same words.

In the end of the day both things says: "please give back"
And the rest is just worthless nitpicking - in my local
language is it "flueknepperi".

	Sam

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Thursday 14 June 2007 16:42:44 Alexandre Oliva wrote:
> On Jun 14, 2007, Sam Ravnborg <sam@ravnborg.org> wrote:
> > On Thu, Jun 14, 2007 at 04:46:36PM -0300, Alexandre Oliva wrote:
> >> > Giving back "in kind" is obvious. I give you source code to do with as
> >> > you see fit. I just expect you to give back in kind: source code for
> >> > me to do with as I see fit, under the same license I gave you source
> >> > code.
> >> >
> >> > How hard is that to accept?
> >>
> >> Forgive me if I find this a bit hard, because that's *not* what the
> >> GPL says.
> >
> > What part of the word "expect" did you not understand?
>
> http://lkml.org/lkml/2006/9/24/246
>
>   It asks everybody - regardless of circumstance - for the same thing.
>   It asks for the effort that was put into improving the software to
>   be given back to the common good.  You can use the end result any
>   way you want (and if you want to use it for "bad" things, be my
>   guest), but we ask the same exact thing of everybody - give your
>   modifications back.
>
> > And whats your point here anyway?
>
> The the GPL doesn't do that.  It encourages that.  But what it asks
> for is respect for the freedoms it defends WRT the software licensed
> under it.

Logical fallacy. The two statements are semantically equivalent, and the draw 
and allure of "Open Source" is that the software continually gets better at 
doing its job, grows more features, etc... *ALL* because the modifications 
*DO* get "given back".

Because it is *VERY* hard to keep a modification *PRIVATE* and avoid 
the "distribution" clauses of the GPL the belief that it "doesn't require 
giving changes back" is technically and literally true, but is false in 
practice.

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Thu, 14 Jun 2007, Alexandre Oliva wrote:
> 
> Then would you consider relicensing Linux under GPLv3 + additional
> permission for Tivoization?

No. I'm not stupid.

The GPLv3 explicitly allows removing additional permissions.

So anybody who does "GPLv3 + additional permissions" is basically setting 
himself up for people taking those permissions away.

Since the Tivo kind of permission is in my opinion a *fundamental right* 
(or call if "freedom" if you want), then "GPLv3 + additional permissions" 
simply is not a viable alternative, since anybody could just decide to 
make improvements and strip those permissions.

The whole notion of "additional permissions" in the GPLv3 is totally 
pointless, since it's legally *exactly* the same as allowing dual 
licensing (which a license doesn't even have to spell out: you can 
dual-license *regardless* of the license!).

The reason for the "additional permissions" is just to make the LGPL go 
away, and become a sub-clause of the GPLv3.

If you really thought anything else, you're just uninformed and stupid, 
and didn't think things through.

			Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:

> On Thu, 14 Jun 2007, Alexandre Oliva wrote:
>> 
>> Then would you consider relicensing Linux under GPLv3 + additional
>> permission for Tivoization?

> No. I'm not stupid.

> The GPLv3 explicitly allows removing additional permissions.

So what?  You just refrain from accepting contributions that attempt
to remove them, and you'll keep TiVO happy.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Thu, 14 Jun 2007, Alexandre Oliva wrote:

> On Jun 14, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:
> > 
> > No. I'm not stupid.
> >
> > The GPLv3 explicitly allows removing additional permissions.
> 
> So what?  You just refrain from accepting contributions that attempt
> to remove them, and you'll keep TiVO happy.

You really aren't thinking, are you?

It's not about keeping Tivo happy. It's about keeping *me* happy. That's 
my primary (only) motivation for a license.

And let's go back to why I selected the GPLv2 in the first place, shall 
we?

I want to be able to use other peoples improvements. If they release 
improved versions of the software I started, I want to be able to merge 
those improvements if I want to.

Your *IDIOTIC* suggestion is explicitly against the whole POINT! By saying 
that I shouldn't accept contributions like that, you just INVALIDATED the 
whole point of the license in the first place!

Can you really not see that?

		Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:

> I want to be able to use other peoples improvements. If they release 
> improved versions of the software I started, I want to be able to merge 
> those improvements if I want to.

Hmm...  So, if someone takes one of the many GPLv2+ contributions and
makes improvements under GPLv3+, you're going to make an effort to
accept them, rather than rejecting them because they're under the
GPLv3?

> Your *IDIOTIC* suggestion is explicitly against the whole POINT! By saying 
> that I shouldn't accept contributions like that, you just INVALIDATED the 
> whole point of the license in the first place!

I understand.  I assumed you had some trust that people would abide by
your wish to permit TiVOization, and that authors of modifications
were entitled to make "whatever restrictions they wanted" on their
code.

Pardon me if I think your position is at least somewhat incoherent.
Can you help me make sense of it?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Thu, 14 Jun 2007, Alexandre Oliva wrote:
> 
> Hmm...  So, if someone takes one of the many GPLv2+ contributions and
> makes improvements under GPLv3+, you're going to make an effort to
> accept them, rather than rejecting them because they're under the
> GPLv3?

You *cannot* make GPLv3-only contributions to the kernel.

I'm sorry, but that's how it is. You can take some of the code that is 
GPLv2+ in the kernel, and MOVE IT TO ANOTHER PROJECT, and use them there. 
But not within the confines of the Linux kernel. Within the Linux kernel, 
the GPLv2 rules - and "GPLv2+" becomes just "GPLv2", since the GPLv3 is 
not compatible with v2.

This is no different from the fact that we have some drivers that are 
GPLv2/BSD licensed. Within the kernel, they are GPLv2. But on their own, 
you can choose to use them under the BSD license, make your changes to 
them, and release them commercially.

And correct - I cannot (and neither can anybody else) then accept those 
*non*GPLv2 changes back.

> I understand.  I assumed you had some trust that people would abide by
> your wish to permit TiVOization, and that authors of modifications
> were entitled to make "whatever restrictions they wanted" on their
> code.

Actually, normally I *do* have such a trust. It's why I have no problem 
with drivers that are dual-GPL/BSD, and in fact, I've told people that I 
don't want them to turn them into GPL-only, because that is simply not 
polite.

But I hold *myself* to higher standards than I hold others. And in 
particular, when it comes to people with a religious agenda, I don't 
expect them to be polite or take my feelings into account. I expect (from 
good history) that people with a license agenda will consider the license 
agenda more important than any hurt feelings, or any wishes of mine. 

> Pardon me if I think your position is at least somewhat incoherent.
> Can you help me make sense of it?

I'm giving up. I'm moving you to my "flamers" list, so that your emails go 
to a separate mailbox that I read weekly. I've wasted too much time with 
you, your arguments don't make sense, and you seem to refuse to even _try_ 
to understand my position, or respect the fact that my choice of license 
is MY choice, and that I actually have a brain of my own.

		Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:

> On Thu, 14 Jun 2007, Alexandre Oliva wrote:
>> 
>> Hmm...  So, if someone takes one of the many GPLv2+ contributions and
>> makes improvements under GPLv3+, you're going to make an effort to
>> accept them, rather than rejecting them because they're under the
>> GPLv3?

> You *cannot* make GPLv3-only contributions to the kernel.

I can make improvements to GPLv2+ files under GPLv3 (or rather will,
after GPLv3 is published).  And you wrote:

> I want to be able to use other peoples improvements. If they release
> improved versions of the software I started, I want to be able to
> merge those improvements if I want to.

So which is it?  Do you want to be able to use other people's
improvements, respecting the conditions you said they are legitimately
entitled to make, or is this not quite the whole story?

> But not within the confines of the Linux kernel. Within the Linux kernel, 
> the GPLv2 rules - and "GPLv2+" becomes just "GPLv2", since the GPLv3 is 
> not compatible with v2.

I understand this very well.  You'd have to get the kernel upgraded to
GPLv3 in order to accept the contribution.

Likewise for any other contribution under any other GPLv2-incompatible
license.

So, you see, your statement above, about wanting to be able to use
other people's improvements, cannot be taken without qualification.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Jeremy Maitin-Shepard]

Alexandre Oliva <aoliva@redhat.com> writes:

> On Jun 14, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:
>> On Thu, 14 Jun 2007, Alexandre Oliva wrote:
>>> 
>>> Hmm...  So, if someone takes one of the many GPLv2+ contributions and
>>> makes improvements under GPLv3+, you're going to make an effort to
>>> accept them, rather than rejecting them because they're under the
>>> GPLv3?

>> You *cannot* make GPLv3-only contributions to the kernel.

> I can make improvements to GPLv2+ files under GPLv3 (or rather will,
> after GPLv3 is published).

You can do that, but you won't be able to distribute those changes along
with the rest of the kernel.

-- 
Jeremy Maitin-Shepard

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Jeremy Maitin-Shepard <jbms@cmu.edu> wrote:

> Alexandre Oliva <aoliva@redhat.com> writes:
>> On Jun 14, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:
>>> On Thu, 14 Jun 2007, Alexandre Oliva wrote:
>>>> 
>>>> Hmm...  So, if someone takes one of the many GPLv2+ contributions and
>>>> makes improvements under GPLv3+, you're going to make an effort to
>>>> accept them, rather than rejecting them because they're under the
>>>> GPLv3?

>>> You *cannot* make GPLv3-only contributions to the kernel.

>> I can make improvements to GPLv2+ files under GPLv3 (or rather will,
>> after GPLv3 is published).

> You can do that, but you won't be able to distribute those changes along
> with the rest of the kernel.

I know.  Neither will Linus.  But he says he chose GPLv2 such that he
could, and the v2 is better than v3 in this regard.  What's wrong with
this picture?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Carlo Wood]

On Thu, Jun 14, 2007 at 11:31:13PM -0300, Alexandre Oliva wrote:
> I know.  Neither will Linus.  But he says he chose GPLv2 such that he
> could, and the v2 is better than v3 in this regard.  What's wrong with
> this picture?

I'm sure it's a rethorical question - but what is wrong is that
imho the clause that normally is added to any file under the GPL
("... version 2 or higher") is builtin in as a safeguard: If, for
whatever reason, in the future it is discovered that the GPL version
X has a flaw that was not forseen - then it can be 'patched' by
writing a successor, having that released by the FSF and allow
anyone to "upgrade" all GPL-ed software to the new license, as
such having avoided the problems of the said (fictious) flaw.

Therefore, it seems pretty weird to me that LONG before version 3
was written and released, someone purposely would choose to freeze
their software at version 2. Why make it "impossible" to use this
safe-guard? ["impossible" because if you get a signature from every
author where they transfer the authorship rights to you, you can
re-release everything under a different license anyway.  However,
the kernel has so many authors who never signed anything(?) that
this is not possible anymore].

I never knew it was possible to change the "version 2 or higher" into
"only version 2", but I am not a laywer and not into licenses at all,
and I am sure Linus had laywers look into this, so we can take this
as a fact.

The result is simple:

1) A lot of files in the kernel are fixed at version 2.
2) Version 3 is incompatible with version 2 (which I also take
   for granted, having read that in this thread).

Many (or at least important) authors of version-2 files do not
wish to change the license to one that allows it to be transformed
to GPL v3 (ie, add the clause "version 2 or higher". Therefore,
this will not happen.

The result is that it is impossible to accept/add patches that
can not be converted to GPL v2 (ie, which are explicitely version 2
or have the phrase "version 2 or higher").

I think this whole thread has only one purpose:

To "test" if the kernel source is indeed - by LAW - immutable
and fixed to version 2.

Of course, as an unwritten rule, everyone here will respect
Linus' wishes: he is our leader for things like this. But that
is irrelevant when it is about important things like this:
One day he will be gone - and if this license "upgrade" is only
stopped because people see him as a leader, then the discussion
will happen again and again and again - until the software IS
upgraded. This being a one-way lane (once upgraded, you can't
go back) I think it is LOGICAL to do this: 1) Don't upgrade
for as long as possible (possibly forever), 2) Have discussions
like this while Linus is still alive until it is cristal clear that
the kernel can or can not be "upgraded" to version 3 as a result
of the law. And, imho, as I pointed out above - this is already
clear to me (so this thread can end now as far as me is concerned).

Finally - I realize that most people who seem to oppose Linus -
ie Alexandre, and me included, are NOT (perse) in favour of version 3
(I know I am not). The most important purpose of the discussion
is to TEST if someone could force this "upgrade" in the future
when leadership is less clear. So, I think it's good that some
people are willing to take the side of version 3 and try to go
all the way to prove that it's better and/or possible to "upgrade"
etc - as if they REALLY want that.

-- 
Carlo Wood <carlo@alinoe.com>

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Fri, 15 Jun 2007, Carlo Wood wrote:
> 
> Therefore, it seems pretty weird to me that LONG before version 3
> was written and released, someone purposely would choose to freeze
> their software at version 2. Why make it "impossible" to use this
> safe-guard?

Because what you call "safe-guard" is just call "idiotic".

The fact is, anybody who cares about the copyright license he uses on 
software (and I sure do care!) should *never* give a blank check to 
somebody else that they don't trust to change that license for them.

In other words, the whole "v2 or later" language *only* makes sense if:

 - You are the FSF, and you *decide* what "or later" actually means

   In this case, you obviously don't have to trust "somebody else". You 
   just trust yourself.

 - You don't care about your choice of license.

   In this case, you might as well let somebody else make that decision 
   for you, although quite frankly, you might as well use something like 
   the BSD license and let many *more* people make that decision to 
   relicense for you.

 - ..or you trust the FSF implicitly.

   In this case, you're not an independent entity, you're just a lackey of 
   the FSF when it comes to the license.

If none of those conditions are true, you'd be *incompetent* to leave the 
"v2 or later".

And I'm not incompetent.  None of the above conditions holds true for me, 
so the "v2 or later" would be totally idiotic. Which is why Linux has 
*never* had that statement for *any* code I have ever written!

So the fact that *you* (and the FSF) call it a "safe-guard" by no means 
makes it so.

> I never knew it was possible to change the "version 2 or higher" into
> "only version 2", but I am not a laywer and not into licenses at all,
> and I am sure Linus had laywers look into this, so we can take this
> as a fact.

What you state above is that you didn't even *READ* the GPL, yet you make 
some arguments about it as if you had!

So let me educate you. You may not want to be educated, but I'll try 
anyway:

 - Please *read* the license that you discuss. If you don't read it, 
   there's no point in you making any arguments about it.

   And by reading it, I don't mean just mechanical "reading" in a 
   technical sense. I mean "read and think about it"

 - Once you have actually read it, you'll notice that there *is* no 
   "either version 2 of the License, or (at your option) any later 
   version" in the GPL v2 license AT ALL.

   Really. That phrase DOES NOT EXIST in the GPLv2. 

 - That phrase exists *outside* the license, in the *suggestion* of how 
   you migth want to use it! Notice how that license comes *after* the 
   "END OF TERMS AND CONDITIONS" thing, and in the section that describes 
   how you might *use* the license.

In other words, you are a perfect example of somebody who argues about the 
GPLv2 without ever having read or understood it!

I'm really tired of people claiming that the kernel license is a 
"modified" GPLv2 to remove the "v2 or later" language. It's *not* modified 
at all: that language simply DOES NOT EXIST in the GPLv2. It exists 
outside the license, in a suggestion on how you _might_ use it, and it's a 
suggestion that my code has *never* *ever* actually followed.

>From the very first version of Linux that used the GPLv2 (that would be 
Linux-0.95, iirc), Linux has *always* been under a unmodified GPLv2. Not 
"any later". Not "modified".  It has *always* stated the version 
explicitly (by *including* it verbatim - there was certainly never any 
ambiguity about versions), and it has *never* stated "v2 or later" in any 
source code I wrote!

Yes, afterwards, I have added clarifications, because it turns out that 
lawyers (and non-lawyers, for that matter) actually prefer having certain 
things clarified and written out, so that nobody can later claim that they 
misunderstood certain things.

But even those clarifications NEVER MODIFIED THE ACTUAL LICENSE. They are 
literally just clarifications. The license didn't change.

The clarifications are

 - the obvious *fact* that user-mode code using standard system calls is 
   not a derived work of the kernel.

   This is so obvious to anybody with any knowledge of what "derived work" 
   means that there is absolutely no legal question about this being a 
   fact, but the FSF had made such a name for the GPL being "viral" that 
   some people were still worried that I was crazy and would sue. They 
   probably weren't actually worried that they'd lose, but a lawsuit is 
   potentially quite damaging (to PR and in just legal expenses) even if 
   you do end winning, so to make it clear that I wasn't crazy, I added a 
   clarification about this very early on.

   Side note: while it's just a clarification, and as such doesn't 
   actually alter the license in *any* way, it's *also* a "public 
   statement of intent", and as such has legal meaning: it means that if I 
   were to stop taking my meds and go crazy, I *still* couldn't sue, 
   because of the so-called "doctrine of lathes" (and probably other 
   legal issues), since I have made it clear that you cannot be *expected* 
   to be sued over it.

   So please do realize that while a clarification does not change the 
   license in any way, shape, or form, it does actually have some real 
   legal implications as to whether you can effectively sue somebody. You 
   cannot first claim that certain uses are ok, and then suddenly change 
   your mind when people have made decisions based on your claim.

 - the above *fact* that the source code never followed the FSF suggestion 
   of "v2 or later" was clarified and made explicit.

   Again, this in no way changed the license itself (that "v2 or later" 
   language simply _does_not_exist_ in the GPLv2), but it *also* didn't 
   actually change the licensing of the code itself. Because the whole "v2 
   or later" language (in the "how to use") only takes effect 
   *if*it*exists* in the source code, and the whole "any version of your 
   choice" language (which *does* exist in the GPLv2) only takes effect if 
   there wasn't some explicit license version that was obvious (and 
   dammit, it doesn't get much more obvious than a COPYING file that 
   includes the whole license *verbatim*)

So I am 100% sure that the whole Linux kernel is "v2 only". It has 
*always* been "v2 only" since very early in 1992, and before that, it 
wasn't GPL'd at all.

Certain individual *parts* of the kernel can be used under other licenses 
(BSD, Mozilla, and yes, GPLv3+), but that is very much a localized issue, 
and does not in any way change the fact that the Linux kernel itself in 
its entirety is GPL version *2*. And nothing else.

And I'm also 100% sure that anybody who claims otherwise is totally 
ignorant of either copyright law, the GPLv2, or *both*.

				Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Sanjoy Mahajan]

>> "version 2 or higher"

> That phrase exists outside the license

That's true.  But sec. 9 of the GPLv2 says:

  If the Program does not specify a version number of this License, you
  may choose any version ever published by the Free Software Foundation.

So, by making the COPYING contain the v2 text, is the author
specifying a particular version?  If yes, then the sec. 9 provision
would be meaningless, since there would be no way to not specify a
version number.  

My understanding is that courts would presume that a license term has
a meaning, if it has a plausible reading.  And there such a reading:
that to specify a version, there needs to be (e.g. in the source
files) a statement like, "This file [or work] is licensed under the
GNU GPLv2."

Corrections, flames, etc. are welcome.

-Sanjoy

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Krzysztof Halasa]

Sanjoy Mahajan <sanjoy@mrao.cam.ac.uk> writes:

> So, by making the COPYING contain the v2 text, is the author
> specifying a particular version?  If yes, then the sec. 9 provision
> would be meaningless, since there would be no way to not specify a
> version number.

Of course the "published under terms of GPL." would do.
-- 
Krzysztof Halasa

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Sanjoy Mahajan]

Krzysztof Halasa <khc@pm.waw.pl> writes:
> > So, by making the COPYING contain the v2 text, is the author
> > specifying a particular version?  If yes, then the sec. 9 provision
> > would be meaningless, since there would be no way to not specify a
> > version number.
> 
> Of course the "published under terms of GPL." would do.

It would do, if you could stop there and say no more.  But:

  You may copy and distribute verbatim copies of the Program's source
  code as you receive it, in any medium, provided that you...give any
  other recipients of the Program a copy of this License along with the
  Program.  [GPLv2, section 1]

So you have to give recipients the license text from a particular
version of the GPL.  To make that the only version unde which the work
is licensed, you have to add something like "Licensed under the
GPLv2".  Otherwise sec. 9 says that you offer the work under any
version of the GPL, and the licensee can take his or her pick -- even
using v1 (!).

-Sanjoy

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Krzysztof Halasa]

Sanjoy Mahajan <sanjoy@mrao.cam.ac.uk> writes:

> So you have to give recipients the license text from a particular
> version of the GPL.  To make that the only version unde which the work
> is licensed, you have to add something like "Licensed under the
> GPLv2".  Otherwise sec. 9 says that you offer the work under any
> version of the GPL, and the licensee can take his or her pick -- even
> using v1 (!).

That's exactly what I meant by "not specifying the version".
-- 
Krzysztof Halasa

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Rob Landley]

On Thursday 14 June 2007 19:20:19 Alexandre Oliva wrote:
> > But not within the confines of the Linux kernel. Within the Linux kernel,
> > the GPLv2 rules - and "GPLv2+" becomes just "GPLv2", since the GPLv3 is
> > not compatible with v2.
>
> I understand this very well.  You'd have to get the kernel upgraded to
> GPLv3 in order to accept the contribution.

Why do you keep saying "upgraded" to GPLv3?  How is it an improvement to move 
from a small, simple, elegant, and tested implementation to something that's 
more complicated, less elegant, less coherent, totally untested, and full of 
numerous special cases?

Bumping a version number is not in indicator of quality, and spending over 
twice as much text to express the same legal principles is not an 
improvement.  So far, you haven't brought up a single reason to use v3 except 
for a higher version number.  (Not that I'm asking you to.)  You've just 
tried to argue that it isn't WORSE than the existing license.

Rob
-- 
"One of my most productive days was throwing away 1000 lines of code."
  - Ken Thompson.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Olivier Galibert]

On Thu, Jun 14, 2007 at 09:20:35PM -0400, Rob Landley wrote:
> Why do you keep saying "upgraded" to GPLv3?  How is it an improvement to move 
> from a small, simple, elegant, and tested implementation to something that's 
> more complicated, less elegant, less coherent, totally untested, and full of 
> numerous special cases?

Ahhh, but so much more entreprisy.  I never had realized before that
the DailyWTF applied to licenses too.

  OG.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Rob Landley <rob@landley.net> wrote:

> On Thursday 14 June 2007 19:20:19 Alexandre Oliva wrote:

>> I understand this very well.  You'd have to get the kernel upgraded to
>> GPLv3 in order to accept the contribution.

> Why do you keep saying "upgraded" to GPLv3?

Just because it has a higher version number.  Honest, no other
reason was implied.

I'm seriously not trying to push v3 here.  I got into this to try to
dispell myths and get a better grasp of the situation.

> Bumping a version number is not in indicator of quality,

Agreed.  Still, some people talk about upgrading from XP to Vista (ok,
no numbers here, but you get the idea), just like they talk about
upgrading from linux 2.4 to 2.6.

> So far, you haven't brought up a single reason to use v3

Sure, that was not my goal.  I wasn't even trying.

Would you like me to?

> You've just tried to argue that it isn't WORSE than the existing
> license.

Good, it's nice when people get the idea of what I'm trying to
accomplish.  I feared this had been lost in the noise.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Theodore Tso]

On Thu, Jun 14, 2007 at 08:20:19PM -0300, Alexandre Oliva wrote:
> 
> So, you see, your statement above, about wanting to be able to use
> other people's improvements, cannot be taken without qualification.

No.  Linus and other Linux kernels might *want* to take other people's
improvements, but thanks to Richard Stallman's choices for GPLv3, they
can *not* legally take other people's improvements without violating
the GPLv3 license.  That's not their fault, it's the fault of people
who wrote the GPLv3 license, promulgated the GPLv3 license, and who is
attempting to convince everyone that the GPLv3 license is the only
valid license for Right Thinking FSF automatons to use.

There are plenty of things that I might *want* to do, that I am
legally prohibited from doing.  that doesn't change the fact that I
might want to do it.  The fact that GPLv3 is incompatible with GPLv2
is a tragedy, in the Greek sense.

						- Ted

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Sean]

On Fri, 15 Jun 2007 01:24:32 -0400
Theodore Tso <tytso@mit.edu> wrote:

> No.  Linus and other Linux kernels might *want* to take other people's
> improvements, but thanks to Richard Stallman's choices for GPLv3, they
> can *not* legally take other people's improvements without violating
> the GPLv3 license.  That's not their fault, it's the fault of people
> who wrote the GPLv3 license, promulgated the GPLv3 license, and who is
> attempting to convince everyone that the GPLv3 license is the only
> valid license for Right Thinking FSF automatons to use.
> 
> There are plenty of things that I might *want* to do, that I am
> legally prohibited from doing.  that doesn't change the fact that I
> might want to do it.  The fact that GPLv3 is incompatible with GPLv2
> is a tragedy, in the Greek sense.

The _exact_ same arguments are made against the GPLv2 by the BSD folks.

Given that many people here defend the GPLv2 over BSD, it's ironic
the tone and level of vitriol shown against the v3 and such a
willingness to use the same arguments the BSD folks use against v2.

Both v2 and v3 enforce some restrictions that people who want to
participate must obey.  And _yes_ I acknowledge that v3 has _more_
restrictions.  But then, v2 has more restrictions than BSD and we're
more or less happy with that, aren't we?  In fact, many of us
believe it's a virtue that Linux has a more restrictive license
than that of the BSD's.

While this isn't an argument that we should happily accept more
restrictions, hopefully it will put things in a cheerier perspective.
We're not talking about a fundamental disagreement (ie. no restrictions
versus any restrictions); we're simply talking about _degree_ of
restriction.

There's no problem with people voicing honest disagreement with the v3,
but please lighten up a bit on FSF bashing and the Greek tragedy talk.

Sean.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Al Viro]

On Fri, Jun 15, 2007 at 02:16:46AM -0400, Sean wrote:

> There's no problem with people voicing honest disagreement with the v3,
> but please lighten up a bit on FSF bashing and the Greek tragedy talk.

<wry> Would you prefer a reference to Resistible Rise of Arturo Ui? </wry>

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bernd Paysan]

[-- Attachment #1: Type: text/plain, Size: 2235 bytes --]

On Friday 15 June 2007 07:24, Theodore Tso wrote:
> On Thu, Jun 14, 2007 at 08:20:19PM -0300, Alexandre Oliva wrote:
> > So, you see, your statement above, about wanting to be able to use
> > other people's improvements, cannot be taken without qualification.
>
> No.  Linus and other Linux kernels might *want* to take other people's
> improvements, but thanks to Richard Stallman's choices for GPLv3, they
> can *not* legally take other people's improvements without violating
> the GPLv3 license.  That's not their fault, it's the fault of people
> who wrote the GPLv3 license, promulgated the GPLv3 license, and who is
> attempting to convince everyone that the GPLv3 license is the only
> valid license for Right Thinking FSF automatons to use.

Ah no, it's their fault. The GPLv2 always was clear that there will be some 
future releases of the GPL, and that you should keep "upgrading" possible.

> There are plenty of things that I might *want* to do, that I am
> legally prohibited from doing.  that doesn't change the fact that I
> might want to do it.  The fact that GPLv3 is incompatible with GPLv2
> is a tragedy, in the Greek sense.

The GPLv2 tries hard to be compatible with any further versions of the GPL 
as possible, by allowing people to choose which license you take, and by 
making sure that no man in the middle can restrict this choice. If people 
deliberately select to use "GPLv2 only", who's to blame? RMS? Come on, 
that's bullshit. It's *Linus Torvalds* who made Linux incompatible with 
GPLv3, nobody else - ok, Al Viro with his tagged GPLv2 files (and honestly, 
I think this is just another Linus misinterpretation about the GPL, and he 
really didn't do it, because he couldn't).

This thread was fun, but I think all arguments have been repeated often 
enough. I try to give up. I suggest everyone who has some assertions about 
what the GPLv2 does read it through and find the place where it says so. 
Unfortunately, I haven't seen GPL citations from the Linus-fanboy curve, 
only suggestions that the GPL "does not say something" which it clearly 
does.

-- 
Bernd Paysan
"If you want it done right, you have to do it yourself"
http://www.jwdt.com/~paysan/

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Friday 15 June 2007 06:02:11 Bernd Paysan wrote:
> On Friday 15 June 2007 07:24, Theodore Tso wrote:
> > On Thu, Jun 14, 2007 at 08:20:19PM -0300, Alexandre Oliva wrote:
> > > So, you see, your statement above, about wanting to be able to use
> > > other people's improvements, cannot be taken without qualification.
> >
> > No.  Linus and other Linux kernels might *want* to take other people's
> > improvements, but thanks to Richard Stallman's choices for GPLv3, they
> > can *not* legally take other people's improvements without violating
> > the GPLv3 license.  That's not their fault, it's the fault of people
> > who wrote the GPLv3 license, promulgated the GPLv3 license, and who is
> > attempting to convince everyone that the GPLv3 license is the only
> > valid license for Right Thinking FSF automatons to use.
>
> Ah no, it's their fault. The GPLv2 always was clear that there will be some
> future releases of the GPL, and that you should keep "upgrading" possible.

Not true at all. The GPLv2 leaves it up to the person placing their work under 
the GPLv2 license that its up to them whether they want the license on 
their "covered work" to be able to be changed. That the boilerplate includes 
this clause is pretty pointless - anyone can easily remove the "or, at your 
option, any later version" clause and render section 9 meaningless as applies 
to their work.

> > There are plenty of things that I might *want* to do, that I am
> > legally prohibited from doing.  that doesn't change the fact that I
> > might want to do it.  The fact that GPLv3 is incompatible with GPLv2
> > is a tragedy, in the Greek sense.
>
> The GPLv2 tries hard to be compatible with any further versions of the GPL
> as possible, by allowing people to choose which license you take, and by
> making sure that no man in the middle can restrict this choice. If people
> deliberately select to use "GPLv2 only", who's to blame? RMS? Come on,
> that's bullshit. It's *Linus Torvalds* who made Linux incompatible with
> GPLv3, nobody else - ok, Al Viro with his tagged GPLv2 files (and honestly,
> I think this is just another Linus misinterpretation about the GPL, and he
> really didn't do it, because he couldn't).

Incorrect. Read section 9 of the GPLv2. It's pretty clear that the "any later 
version" clause is optional. Whats more is that since the modern linux kernel 
*IS* a "composite work" composed of Linus' original code with changes 
contributed by other people - Linus retains copyright to the work as a whole. 
This means that he can license it in any manner he chooses, as long as it 
doesn't affect the copyrights (or licensing) of the people that have 
contributed changes. I don't have to go to the US copyright law for this - 
Linus released Linux under the GPL, others made changes and sent them back 
saying "You let me have access to your code under the GPL, I've made some 
changes that make it better. You can have my changes under the GPL." QED: 
Linus still holds copyright to Linux and can license it in any way he 
chooses. This is limited because of the license he accepted when adding the 
changes back to his code. He may have locked the kernel, as a whole, to 
version 2 of the GPL - but that is his right. There is nothing he has done 
that has stopped people from having their code included that is still "v2 or 
later". 

> This thread was fun, but I think all arguments have been repeated often
> enough. I try to give up. I suggest everyone who has some assertions about
> what the GPLv2 does read it through and find the place where it says so.
> Unfortunately, I haven't seen GPL citations from the Linus-fanboy curve,
> only suggestions that the GPL "does not say something" which it clearly
> does.

Because there has been no need to quote the GPLv2 until it became clear that 
people were going to keep claiming it stated things it did not. Since then 
I've started quoting the relevant sections of it.

But I agree with you - the thread was fun. And then I realized that the 
discussion was going nowhere at all. So I'm going to answer the last few 
messages in my inbox and then start filtering messages with this topic off 
without reading them.

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Carlo Wood]

On Fri, Jun 15, 2007 at 06:33:51AM -0400, Daniel Hazelton wrote:
> Incorrect. Read section 9 of the GPLv2. It's pretty clear that the "any later 
> version" clause is optional. Whats more is that since the modern linux kernel 
> *IS* a "composite work" composed of Linus' original code with changes 
> contributed by other people - Linus retains copyright to the work as a whole. 

Huh - surely not to files added to the kernel that were written by
others from scratch!

> This means that he can license it in any manner he chooses, as long as it 
> doesn't affect the copyrights (or licensing) of the people that have 
> contributed changes. I don't have to go to the US copyright law for this - 
> Linus released Linux under the GPL, others made changes and sent them back 
> saying "You let me have access to your code under the GPL, I've made some 
> changes that make it better. You can have my changes under the GPL." QED: 
> Linus still holds copyright to Linux and can license it in any way he 
> chooses.

This is totally new to me - if this is true - I'd really like to be sure!
I always thought that it would be necessary to get signatures of each
and every contributor before you can change a license of a file. Why do
you think that the FSF demands written copyright-transfers with
signatures before you are allowed to submit a patch to any of their
largers projects? If they - as original copyright holder - could do
what you claim - they wouldn't need those signatures.

Having signed a copyright transfer for 'future' changes for gprof,
libiberty, readline, zlib, gcc, gdb, libstdc++, bfd, dejagnu, gas,
and binutils,
Carlo Wood <carlo@alinoe.com>

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Fri, 15 Jun 2007, Carlo Wood wrote:

> On Fri, Jun 15, 2007 at 06:33:51AM -0400, Daniel Hazelton wrote:
> > Incorrect. Read section 9 of the GPLv2. It's pretty clear that the "any later 
> > version" clause is optional. Whats more is that since the modern linux kernel 
> > *IS* a "composite work" composed of Linus' original code with changes 
> > contributed by other people - Linus retains copyright to the work as a whole. 
> 
> Huh - surely not to files added to the kernel that were written by
> others from scratch!

Actually, yes. Even to those - when they are part of "the whole".

I'm sorry, but I've learnt more about copyright law, and talked to more 
lawyers about licensing that probably most of the rest of the people 
involved in this discussion have *combined*.

And yes, at least under US copyright law, and at least if you see Linux as 
a "collective work" (which is arguably the most straightforward reading og 
copyright law, but perhaps not the only one) I am actually the sole owner 
of copyright in the *collective* work of the Linux kernel.

The way "collective works" work, there are two separate copyrights: there 
is the copyright in the "separate contribution", which is vests ininitally 
in the author of that contribution (unless he signs over his copyrights, 
often by virtue of working for somebody else).

And then there is the copyright in the "collective work", which would be 
me.

Of course, owning coyright in the "collective work" doesn't actually give 
me complete control anyway. I cannot relicense things in ways that go 
against the rules of the individual works. But in a very real sense, yes, 
I actually do own a certain (*limited*) copyright over even the parts that 
have not been explicitly signed over to me.

And yes, there are other potential ways to describe Linux, and in the end, 
it doesn't really matter. Because the way the GPLv2 works, it makes it 
clear that as long as a piece is a part of the whole, it has to be 
licensed under the GPLv2 and nothing else.

And btw, just to make you feel safe - I cannot do anything about that, 
even if I *do* own the copyright in the collective, because of the 
limitations on what that colletive work copyright implies (it says that I 
have the right to reproduce and distribute, but I don't have the right to 
*modify* except as given to me by the original author!)

So don't worry. I *technically* have certain special rights, but I 
practically speaking gave up most all of those rights by accepting code 
from others under the GPLv2 - since in order to do that, I had to agree to 
be bound by the GPLv2 license myself.

> This is totally new to me - if this is true - I'd really like to be sure!

It would be generally held to be true at least in the US, but it doesn't 
really matter.

> I always thought that it would be necessary to get signatures of each
> and every contributor before you can change a license of a file.

You are mostly correct. The "mostly" comes because I would not say "every 
contributor", but would clarify it by saying "every copyright holder". The 
difference? Not all contributions are necessarily copyrightable. If you 
send in trivial one-liners, we will credit you for them, but that does not 
automatically mean that you necessarily own copyright in something.

But yes, somebody who wrote an original file (that has some artistic 
expression, and isn't just a list of PCI ID's, for example) will be the 
copyright owner in that file. Some *very* few people have actually sent me 
paperwork to transfer the ownership of copyrights, but they seem to have 
done that because they were just used to doing it with the FSF, and I 
actually don't care. 

				Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Adrian Bunk]

On Fri, Jun 15, 2007 at 08:45:43AM -0700, Linus Torvalds wrote:
> 
> 
> On Fri, 15 Jun 2007, Carlo Wood wrote:
> 
> > On Fri, Jun 15, 2007 at 06:33:51AM -0400, Daniel Hazelton wrote:
> > > Incorrect. Read section 9 of the GPLv2. It's pretty clear that the "any later 
> > > version" clause is optional. Whats more is that since the modern linux kernel 
> > > *IS* a "composite work" composed of Linus' original code with changes 
> > > contributed by other people - Linus retains copyright to the work as a whole. 
> > 
> > Huh - surely not to files added to the kernel that were written by
> > others from scratch!
> 
> Actually, yes. Even to those - when they are part of "the whole".
> 
> I'm sorry, but I've learnt more about copyright law, and talked to more 
> lawyers about licensing that probably most of the rest of the people 
> involved in this discussion have *combined*.
> 
> And yes, at least under US copyright law, and at least if you see Linux as 
> a "collective work" (which is arguably the most straightforward reading og 
> copyright law, but perhaps not the only one) I am actually the sole owner 
> of copyright in the *collective* work of the Linux kernel.

US law is only relevant for < 5% of all people.

How valid would any action based on US copyright law be in other parts 
of the world?

> The way "collective works" work, there are two separate copyrights: there 
> is the copyright in the "separate contribution", which is vests ininitally 
> in the author of that contribution (unless he signs over his copyrights, 
> often by virtue of working for somebody else).
> 
> And then there is the copyright in the "collective work", which would be 
> me.
> 
> Of course, owning coyright in the "collective work" doesn't actually give 
> me complete control anyway. I cannot relicense things in ways that go 
> against the rules of the individual works. But in a very real sense, yes, 
> I actually do own a certain (*limited*) copyright over even the parts that 
> have not been explicitly signed over to me.
>...

Does this include GPLv2'ed code not intended to be used in the Linux 
kernel submitted by people other than the copyright holder for inclusion 
in the Linux kernel?

If yes, the FSF has exactly the same rights if taking a GPLv2 driver 
from the Linux kernel and including it in GNU Hurd.

> 				Linus

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Fri, 15 Jun 2007, Adrian Bunk wrote:
> 
> US law is only relevant for < 5% of all people.
> 
> How valid would any action based on US copyright law be in other parts 
> of the world?

I tried to explain that in the case of the Linux kernel, we really don't 
care, since in the end, what matters is the GPLv2, and I have bound myself 
to the terms of that license *regardless* of any US law.

So yes, US law is only relevant for < 5% of all people, but in the 
specific case of the kernel, even that US law isn't _really_ all thgat 
relevant at all, not *even* to those 5% ;)

So I *really* hope you took my explanation of why I actually have more 
rights than others as a nitpicking "legal detail", not as a "I own your 
very SOUL, bow down before me!" kind of thing.

But to answer your question by _another_ nitpicking answer, as the 
original author, I probably do have some special legal status even in 
Europe, and probably in other places too. The fact that others *extended* 
on my original work doesn't take away the special place of original 
authorship, even if the extended version has a totally different form (ie, 
a movie based on a book ends up still havign the original author of the 
book holding special rights - and in fact those rights are in some cases 
much *stronger* in Europe than they are in the US).

For example, Europe recognizes "moral rights" in original authorship, in 
ways that it is much harder to enforce (if at all) in the US.

But as mentioned, since I myself has bound myself to the GPLv2, that 
really is a pretty damn theoretical argument. When it comes to the kernel, 
I'm "Primus inter pares", if you wish.

("First among equals", for the non-latin-speaking world ;)

> Does this include GPLv2'ed code not intended to be used in the Linux 
> kernel submitted by people other than the copyright holder for inclusion 
> in the Linux kernel?

In the US sense of "within the compilation", probably yes. Ie that right 
is tied to *linux* as a compilation. In the European sense of "moral 
rights", no - that right is very much tied to original authorship (so I, 
as original author of Linux, have some rights with respect to Linux, but 
on the other hand, they, as the original authors of some specific code, 
have some specific rights wrt that code!)

> If yes, the FSF has exactly the same rights if taking a GPLv2 driver 
> from the Linux kernel and including it in GNU Hurd.

The FSF does indeed have special rights wrt Hurd, regardless of where the 
code in question has come from (as long as it came there *legally*, of 
course!). So when it comes to Hurd, they have the copyright in the 
compilation of that (and Red Hat as their copyright in their RHEL 
distribution - of which the kernel is just a small part!).

It's a confusing set of nesting dolls of different levels of copyrights, 
and the GPLv2 makes it all pretty much irrelevant, by being the thing that 
cuts through all the other licensing issues!

			Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Friday 15 June 2007 12:22:16 Adrian Bunk wrote:
> On Fri, Jun 15, 2007 at 08:45:43AM -0700, Linus Torvalds wrote:
> > On Fri, 15 Jun 2007, Carlo Wood wrote:
<snip>
> > The way "collective works" work, there are two separate copyrights: there
> > is the copyright in the "separate contribution", which is vests
> > ininitally in the author of that contribution (unless he signs over his
> > copyrights, often by virtue of working for somebody else).
> >
> > And then there is the copyright in the "collective work", which would be
> > me.
> >
> > Of course, owning coyright in the "collective work" doesn't actually give
> > me complete control anyway. I cannot relicense things in ways that go
> > against the rules of the individual works. But in a very real sense, yes,
> > I actually do own a certain (*limited*) copyright over even the parts
> > that have not been explicitly signed over to me.
> >...
>
> Does this include GPLv2'ed code not intended to be used in the Linux
> kernel submitted by people other than the copyright holder for inclusion
> in the Linux kernel?
>
> If yes, the FSF has exactly the same rights if taking a GPLv2 driver
> from the Linux kernel and including it in GNU Hurd.

Of course they do. That is defined by the license. They can include it with 
HURD, but that doesn't give them copyright to it, just the right to exercise 
the rights granted by the license.


DRH

>
> > 				Linus
>
> cu
> Adrian



-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Carlo Wood]

On Fri, Jun 15, 2007 at 08:45:43AM -0700, Linus Torvalds wrote:
> I'm sorry, but I've learnt more about copyright law, and talked to more 
> lawyers about licensing that probably most of the rest of the people 
> involved in this discussion have *combined*.

Which is why I have taken everything you said so far for granted
as being a fact -- no need to try to convince me of something ;)
In my case it suffices if you say that you are sure (and before
anyone says: you're nuts - what I REALLY mean is 'for the purpose
of this discussion'. For me, it makes no sense to waste more time
on discussing such an issue; I won't bet all my money it, of course).

[..snip..]
> > I always thought that it would be necessary to get signatures of each
> > and every contributor before you can change a license of a file.
> 
> You are mostly correct. The "mostly" comes because I would not say "every 
> contributor", but would clarify it by saying "every copyright holder". The 
> difference? Not all contributions are necessarily copyrightable. If you 
> send in trivial one-liners, we will credit you for them, but that does not 
> automatically mean that you necessarily own copyright in something.

I knew that.

> But yes, somebody who wrote an original file (that has some artistic 
> expression, and isn't just a list of PCI ID's, for example) will be the 
> copyright owner in that file. Some *very* few people have actually sent me 
> paperwork to transfer the ownership of copyrights, but they seem to have 
> done that because they were just used to doing it with the FSF, and I 
> actually don't care. 
> 
> 				Linus

The point is: can you, or can't you (legally) relicense the whole kernel
tree under the GPLv3 (or GPLv2+GPLv3)?

At first I thought that you cannot, because too many (significant) contributors
have been involved (and you will never get signatures from them all).
Then someone surprised me by claiming that the original author had
copyright for everything - even files added by others. To me, this
seemed to say: even if those contributors don't like it, the original
author can still sell the whole to some company under a proprietary
license (also still having the original under the GPL on the net, of
course), as he could do when he was the sole author.

Then you reply with:

> Carlo> Huh - surely not to files added to the kernel that were written by
> Carlo> others from scratch!
>
Linus> Actually, yes. Even to those - when they are part of "the whole".

But the rest of your reply made it a bit unclear again.

Assume you stopped taking your meds and next week you think that GPLv3
is THE thing for the kernel. Then could you legally, or can't you, go
ahead and change the license of the whole kernel to GPLv3? And if you
can't, then roughly how many files / authors are stopping you from
doing so?

If the answer is: I can't. Then I think you're a lucky bastard, and
have escaped years and years of discussions with people trying to
convince you that the GPLv3 is better ;)

I think that the question: can OTHERS "upgrade" the kernel to GPLv3
has been answered clearly now: No they can't. But if you can, you're
probably not done with dealing with people who want that to happen.

-- 
Carlo Wood <carlo@alinoe.com>

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Fri, 15 Jun 2007, Carlo Wood wrote:
> 
> The point is: can you, or can't you (legally) relicense the whole kernel
> tree under the GPLv3 (or GPLv2+GPLv3)?

No. My special rights do not actually give me those kinds of powers, 
exactly because I'm bound by my _other_ agreement (namely the GPLv2) to 
follow the license of the code that other people have sent me.

> At first I thought that you cannot, because too many (significant) contributors
> have been involved (and you will never get signatures from them all).
> Then someone surprised me by claiming that the original author had
> copyright for everything - even files added by others.

Both are true facts, but the "copyright for everything" is a *separate* 
kind of copyright, which does not include the right to relicense. It's 
literally the "copyright in the collective".

For examples of the US rules, see USC 17.2.201(c) ("Ownership of 
copyright" and " Contributions to Collective Works"), which spells out 
some limited special rights that I have (namely the right to reproduce and 
distribute).

Of course, US law being what it is, the USC is just part of the picture. 
US law is the strange kind of British law, where "case law" is in many 
ways more important than the written-down rules like the USC. So caveat 
emptor!

So I have limited special rights in the collective, but those rights are 
actually in almost every way *more* limited than the rights that the GPLv2 
gives to me (the "almost every way" is because quite frankly, I'm not 
entirely sure about certain special cases. In particular, if somebody 
tried to _revoke_ the rights to their code under the GPLv2, I suspect that 
my rights in the collective would protect me from that and allow me to 
still distribute the code in question, since _those_ rights cannot be 
revoked, and they are _mine_).

And btw: the above paragraph is *way* more legalistic detail than I am at 
all ready to state as "fact". It depends on too many things, and is 
largely speculative in nature.

But one thing is pretty clear and nonspeculative: *nobody* has the right 
to upgrade the kernel to GPLv3. Not me, not you, not anybody. Not without 
clearing it with every single person whose copyright is involved and who 
didn't already give that permission.

So only in the case of some really obscure and unclear situations, I _may_ 
have more rights than some other people, but trust me, but that is damn 
murky, and you'd better have a good lawyer state it, not just a programmer 
who has talked to too many lawyers..

			Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Rob Landley]

On Friday 15 June 2007 14:15:58 Linus Torvalds wrote:
> On Fri, 15 Jun 2007, Carlo Wood wrote:
> > The point is: can you, or can't you (legally) relicense the whole kernel
> > tree under the GPLv3 (or GPLv2+GPLv3)?
>
> No. My special rights do not actually give me those kinds of powers,
> exactly because I'm bound by my _other_ agreement (namely the GPLv2) to
> follow the license of the code that other people have sent me.
>
> > At first I thought that you cannot, because too many (significant)
> > contributors have been involved (and you will never get signatures from
> > them all). Then someone surprised me by claiming that the original author
> > had copyright for everything - even files added by others.
>
> Both are true facts, but the "copyright for everything" is a *separate*
> kind of copyright, which does not include the right to relicense. It's
> literally the "copyright in the collective".
>
> For examples of the US rules, see USC 17.2.201(c) ("Ownership of
> copyright" and " Contributions to Collective Works"), which spells out
> some limited special rights that I have (namely the right to reproduce and
> distribute).

http://www.copyright.gov/title17/92chap2.html#201

There are some really interesting bits in Chapter 1 too...

http://www.copyright.gov/title17

> entirely sure about certain special cases. In particular, if somebody
> tried to _revoke_ the rights to their code under the GPLv2,

There's no revocation clause in the license.  They can't.  (SCO would have 
tried to revoke Caldera's license to its contributions it if it were at all 
possible.)

> I suspect that 
> my rights in the collective would protect me from that and allow me to
> still distribute the code in question, since _those_ rights cannot be
> revoked, and they are _mine_).

Title 17 chapter 1 section 103(a) seems to lean against this.  I think it says 
that having rights to the collective is conditional on having had the rights 
to the components of that collective.

> So only in the case of some really obscure and unclear situations, I _may_
> have more rights than some other people, but trust me, but that is damn
> murky, and you'd better have a good lawyer state it, not just a programmer
> who has talked to too many lawyers..

Entirely agreed. :)

Rob
-- 
"One of my most productive days was throwing away 1000 lines of code."
  - Ken Thompson.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Friday 15 June 2007 09:02:54 Carlo Wood wrote:
> On Fri, Jun 15, 2007 at 06:33:51AM -0400, Daniel Hazelton wrote:
> > Incorrect. Read section 9 of the GPLv2. It's pretty clear that the "any
> > later version" clause is optional. Whats more is that since the modern
> > linux kernel *IS* a "composite work" composed of Linus' original code
> > with changes contributed by other people - Linus retains copyright to the
> > work as a whole.
>
> Huh - surely not to files added to the kernel that were written by
> others from scratch!

Even those.

> > This means that he can license it in any manner he chooses, as long as it
> > doesn't affect the copyrights (or licensing) of the people that have
> > contributed changes. I don't have to go to the US copyright law for this
> > - Linus released Linux under the GPL, others made changes and sent them
> > back saying "You let me have access to your code under the GPL, I've made
> > some changes that make it better. You can have my changes under the GPL."
> > QED: Linus still holds copyright to Linux and can license it in any way
> > he chooses.
>
> This is totally new to me - if this is true - I'd really like to be sure!
> I always thought that it would be necessary to get signatures of each
> and every contributor before you can change a license of a file. Why do
> you think that the FSF demands written copyright-transfers with
> signatures before you are allowed to submit a patch to any of their
> largers projects? If they - as original copyright holder - could do
> what you claim - they wouldn't need those signatures.

They don't. They demand the signature so that some contributor can't change 
their mind at a later date or even be able to give a proprietary software 
vendor the ability to use the GPL'd code in a non-GPL project.

> Having signed a copyright transfer for 'future' changes for gprof,
> libiberty, readline, zlib, gcc, gdb, libstdc++, bfd, dejagnu, gas,
> and binutils,
> Carlo Wood <carlo@alinoe.com>

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Fri, 15 Jun 2007, Bernd Paysan wrote:
> 
> Ah no, it's their fault. The GPLv2 always was clear that there will be some 
> future releases of the GPL, and that you should keep "upgrading" possible.

No. It is clear that you have the *option* of keep upgrading, but it is 
also equally clear that Linux has always decided *not* to exercise that 
option, exactly because I liked the GPLv2, not some "future upgrade".

I decided that long before I saw the GPLv3.

And I'm surprised by people who wonder why I did that.

I'm _intelligent_, dammit. That means that I can foresee the future to 
some degree, at least in the limited sense of what is a likely outcome of 
my actions.

Why are people surprised by the fact that I have foresight? I may be known 
for being an impolite bastard, but quite frankly, anybody who thinks I'm a 
_stupid_ impolite bastard must be missing a page.

You can disagree with my opinions. You can call me obstinate, impolite, 
and opinionated. But quite frankly, very few people have ever found me 
*stupid*.

So give me that - I'm not stupid. That means that I actually *can* predict 
the future to some fuzzy degree, and that people really should *not* be 
surprised by the fact that I never let the FSF control my choice of 
license.

> The GPLv2 tries hard to be compatible with any further versions of the GPL 
> as possible, by allowing people to choose which license you take, and by 
> making sure that no man in the middle can restrict this choice. If people 
> deliberately select to use "GPLv2 only", who's to blame?

There's no "blame". There's only credit.

Besides, you are wrong. The *default* for the GPLv2 in the presense of 
license information is *not* "v2 or later"

In order to get "GPLv2 or later", you actually have to explicitly specify 
it.

I just find it sad that so many people did that, often apparently just 
because they didn't actually read or understand the license.

		Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Jan Harkes]

On Fri, Jun 15, 2007 at 12:02:11PM +0200, Bernd Paysan wrote:
> > the GPLv3 license.  That's not their fault, it's the fault of people
> > who wrote the GPLv3 license, promulgated the GPLv3 license, and who is
> > attempting to convince everyone that the GPLv3 license is the only
> > valid license for Right Thinking FSF automatons to use.
> 
> Ah no, it's their fault. The GPLv2 always was clear that there will be some 
> future releases of the GPL, and that you should keep "upgrading" possible.
...
> enough. I try to give up. I suggest everyone who has some assertions about 
> what the GPLv2 does read it through and find the place where it says so. 

Ok, open your local copy of the GPL (you should have a copy in
/usr/share/common-licences/GPL if you use Debian).

Now read along with me...

|                    GNU GENERAL PUBLIC LICENSE
|                       Version 2, June 1991

Ok, we have that settled, this license describes GPL version 2,
specifically, no if's or later's. Bunch of legal mumbo jumbo follows
until we reach...

|                     END OF TERMS AND CONDITIONS

And we've reached the end of the license. Now it it followed up by some
helpful text for the reader...

|            How to Apply These Terms to Your New Programs
...
|  To do so, attach the following notices to the program.  It is safest
| to attach them to the start of each source file to most effectively
| convey the exclusion of warranty; and each file should have at least
| the "copyright" line and a pointer to where the full notice is found.
|
|    <one line to give the program's name and a brief idea of what it does.>
|    Copyright (C) <year>  <name of author>
|
|    This program is free software; you can redistribute it and/or modify
|    it under the terms of the GNU General Public License as published by
|    the Free Software Foundation; either version 2 of the License, or
|    (at your option) any later version.

Look, some example boiler plate that is suggested to be placed in the
source code _to most effectively convey_ that the code is GPL licensed.
So clearly there must be less effective ways, one of which may just be
to include a COPYING or LICENSE file along with the distributed source.

And the author of this example boiler place is clearly 'talking' to the
the copyright owner/licensor (i.e. not the licencee), giving suggestions
about adding contact information and even possible commands that could
be added to interactive programs. So that '(at your option)' may just be
a suggestion to the licensor, or maybe not. But that isn't really here
not there.

Files that don't have such boiler plate do not convey the terms of the
license as effectively, but they are still covered by the license that
was included with the kernel, which is clearly version 2.

Now ofcourse we were looking at the official GPLv2, so I ran a diff
against the COPYING file in the kernel tree. The only differences I can
see are the included clarification by Linus at the top, some formatting
differences and the fact that the FSF version seems to refer to the
'GNU Lesser General Public License', while Linus's version has,
'GNU Library General Public License'.

> Unfortunately, I haven't seen GPL citations from the Linus-fanboy curve, 
> only suggestions that the GPL "does not say something" which it clearly 
> does.

I hope this helps.

Jan

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 15, 2007, Theodore Tso <tytso@mit.edu> wrote:

> On Thu, Jun 14, 2007 at 08:20:19PM -0300, Alexandre Oliva wrote:
>> 
>> So, you see, your statement above, about wanting to be able to use
>> other people's improvements, cannot be taken without qualification.

> No.  Linus and other Linux kernels might *want* to take other people's
> improvements, but thanks to Richard Stallman's choices for GPLv3, they
> can *not* legally take other people's improvements without violating
> the GPLv3 license.

This argument is backwards.  It's because of Linus' choice for GPLv2
that he can't take improvements under the GPLv3.

Had he chosen any other GPLv3-compatible license, he could.

And the same applies to any other incompatible pair of licenses.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Thursday 14 June 2007 18:35:01 Alexandre Oliva wrote:
> On Jun 14, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:
> > I want to be able to use other peoples improvements. If they release
> > improved versions of the software I started, I want to be able to merge
> > those improvements if I want to.
>
> Hmm...  So, if someone takes one of the many GPLv2+ contributions and
> makes improvements under GPLv3+, you're going to make an effort to
> accept them, rather than rejecting them because they're under the
> GPLv3?

Doesn't matter at all. GPLv3 requires that any project incorporating GPLv3 
code be licensed under the GPLv3. Linus is, as he has shown, intelligent 
enough to know this. The *second* he actually accepted GPLv3 code into the 
kernel it would either be "change the license or start getting lawsuits for 
breach of the terms of the GPLv3".

> > Your *IDIOTIC* suggestion is explicitly against the whole POINT! By
> > saying that I shouldn't accept contributions like that, you just
> > INVALIDATED the whole point of the license in the first place!
>
> I understand.  I assumed you had some trust that people would abide by
> your wish to permit TiVOization, and that authors of modifications
> were entitled to make "whatever restrictions they wanted" on their
> code.
>
> Pardon me if I think your position is at least somewhat incoherent.
> Can you help me make sense of it?

You are making a distinction between "part" and "whole". When separate from 
the kernel the code can have whatever restrictions the creator pleases. If he 
has said "I want this in the "official" Linux Kernel" (ie: I want this in 
Linus' Linux Kernel source tree) then the creator of the code has stated a 
willingness to abide by Linus' decision about the whole work.

It's a moot point, though. The Linux Kernel is licensed under GPLv2, which 
means that *all* code in it has to be under the same license *and* that no 
code in it can have any restrictions *NOT* in the GPLv2.

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Thursday 14 June 2007 15:46:36 Alexandre Oliva wrote:
> On Jun 14, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:
> > On Thu, 14 Jun 2007, Alexandre Oliva wrote:
> >> Is there anything other than TiVOization to justify these statements?
> >
> > Do you need anything else?
>
> No, I'm quite happy that this is all.

So am I, actually :)

>
> > But if by the question you mean "would you think the GPLv3 is fine
> > without the new language in section 6 about the 'consumer devices'", then
> > the answer is that yes, I think that the current GPLv3 draft looks fine
> > apart from that.
>
> Then would you consider relicensing Linux under GPLv3 + additional
> permission for Tivoization?

With Al Viro, at least, specifying that his code has been released *strictly* 
under GPLv2, this is impossible.

> >> Also, can you elaborate on what you mean about 'giving back in kind'?
> >> (I suspect this is related with the tit-for-tat reasoning, that you've
> >> failed to elaborate on before)
> >
> > I've *not* failed to elaborate on that before. Not at all.
> >
> > Just google for
> >
> > 	torvalds tit-for-tat
> >
> > and you'll see a lot of my previous postings. Trying to claim that this
> > is somehow "new" is ludicrous.
>
> I didn't.  But I've provided evidence that your prior musings on this
> topic were wrong.  I wanted to give you an opportunity to review your
> position under this new light.  I see you haven't changed it at all.

It is wrong when you look at the text of the GPLv2 only. When you look at how 
the "Open Source" community works it is clear that the "tit-for-tat" nature 
is a reality. No, it isn't mandated by GPLv2, but that is the "spirit" of the 
GPLv2 that most people who work on Open Source projects follow.

If you want a *REAL* and *CONTINUING* violation of the GPL just look at Herr 
Schillings "cdrecord", in which he places additional restrictions on peoples 
ability to modify the code with statements in the same such as "You must 
leave this check in place" or "You have to leave this comment in place" - 
even when the comment isn't part of the "licensing statement" mandated by the 
GPLv2. (if you want specifics I'll go download the source and pull them out)

> > Giving back "in kind" is obvious. I give you source code to do with as
> > you see fit. I just expect you to give back in kind: source code for me
> > to do with as I see fit, under the same license I gave you source code.
> >
> > How hard is that to accept?
>
> Forgive me if I find this a bit hard, because that's *not* what the
> GPL says.
>
> Where do you think the GPL say that you get the source code back?

It doesn't. But that it doesn't *MAKES* *NO* *DIFFERENCE* because, in 
practice, that is *EXACTLY* what happens anyway.
<snip>
> > I don't call Linux "Free Software". I haven't called it that for close to
> > ten years! Because I think the term "Open Source" is a lot better.
>
> I can appreciate that you think it's better, but unfortunately it
> appears to be playing a significant role in confusing your
> interpretation of the GPL.  The GPL is not just about making the
> source code visible, or even modifyable by others.  It's about
> respecting others' freedoms.  No matter how badly you prefer Open
> Source over Free Software, how badly you'd rather disregard the
> freedoms in the spirit and in the legal terms of the GPL, you chose a
> license designed to protect those freedoms, not only the ability to
> see and modify source code.

Which is not in question here. The objections Linus (and others) have to the 
GPLv3 may share some specifics with my own objections, but my own are that 
GPLv2 respects my freedoms in their entirety. GPLv3 restricts my freedoms 
because one (or more) of the people behind it have a political agenda. (No, 
that term isn't entirely accurate, but its the best one I have found for the 
situation. Explanation: We don't like the way the law of one or more 
munincipalities/political divisions/countries is written and rather than 
trying to get it changed through other channels, we are going to enforce the 
way we think it should be by using a related set of laws and the text of a 
license)

> >> The only thing the GPL demands is respect for others' freedoms, as in,
> >> "I, the author, respect your freedoms, so you, the licensee, must
> >> respect others' freedoms as well".  Is this the "in kind" you're
> >> talking about?  Or are you mistaken about the actual meaning of even
> >> GPLv2?
> >
> > I just ask that you give the software back in a usable form. That's
> > all I ask for.
>
> I'm afraid that's not what the GPLv2 says.  There's no provision
> whatsoever about giving anything back.  Not in the spirit, not in the
> legal terms.

But it does. If you are going to distribute your own, modified version - in 
any way - you have to make the source available as well.

No, it doesn't require that you give back a version if you aren't distributing 
it, but in that case it hardly matters.

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Rob Landley]

On Thursday 14 June 2007 13:14:09 Sean wrote:
> On Thu, 14 Jun 2007 09:01:32 -0700 (PDT)
>
> Linus Torvalds <torvalds@linux-foundation.org> wrote:
> > In other words, we're just *much* better off with a friendly license and
> > not trying to force people to choose sides, than with the rabid idealism
> > that was - and still is - the FSF. The FSF always makes for this horrible
> > "you're with us, or you're against us" black-and-white mentality, where
> > there are "evil" companies (Tivo) and "good" companies (although I dunno
> > if the FSF really sees anybody as truly "good").
>
> Linus,
>
> If you really believe that then why didn't you choose a BSD license
> for Linux?

BSD licenses encourage forking.  Specifically, if a BSD-licensed project 
becomes significantly commercially valuable, there's an incentive for 
companies to hire your developers away to work on a proprietary fork.

When Sun Microsystems started up in 1982, they hired away Bill Joy to work on 
a closed version of BSD (SunOS).  When Berkeley shut down the CSRG, BSDi 
hired those developers to work on another closed source BSD variant.  More 
recently, Apple hired people like Jordan Hubbard away from FreeBSD to do yet 
another fork: MacOS X.  The loss of people hurts the original project.

With BSD licensed code, companies can say "work on the codebase you love as a 
day job, and you can still work on the open version at night".  Then work 
them 90 hours/week.  Or even "we'll release this code open source after we 
can't sell it anymore, a year or two from now".  And then the deadline never 
comes, or the codebase is irrelevant by then, or too far diverged to merge.  
You won't get all the developers, but you'll get enough to cost the open 
project momentum.  BSD is 30 years old and the free version is still a pale 
shadow of its proprietary forks like MacOS X or the bits of it Windows 
incorporated.

Now think about trying to do that to a GPL project.  If you hire the 
developers away, they have to work on a _different_ codebase.  Much less 
compelling, both for the hirer and the hiree.  If you think Linux is 
compelling enough to commercialize, you MUST do so within the terms of the 
GPL or not do it at all.  You can't do a closed fork and distribute the 
result.  Maybe this means companies aren't as quick to jump on the bandwagon 
trying to commercialize it, but the project can then grow larger without 
interference until commercial participation _is_ compelling, on its own 
merits, despite being unable to corner the market on it.

> Instead you chose a license which enforced the so called tit-for-tat
> policy you think is fair.  But people who prefer the BSD license may
> think you're a moron for forcing your political agenda (ie. tit-for-tat)
> on users of your code.

It's not political, it's pragmatic.  GPLv2 has tangible benefits for project 
maintainers.

> The point of all that being, you _do_ believe 
> in enforcing restrictions or you wouldn't like the GPL v2.
>
> So you draw the line of "fairness" and belief that people will
> do-the-right-thing somewhere short of the BSD license.  Why is it
> so hard then to accept that the FSF draws the line short of the
> GPLv2 after having gained practical experience with it
> since its release?

Nobody objects to the FSF putting out new licenses if it changes its mind 
about what it wants to do.  They object to it pestering those of us 
continuing to use the old license because we prefer it to the new one.

The FSF _does_ draw the line in a different place than the Linux developers.  
Hence the Linux developers don't want to use the new license, they prefer the 
old one from back before Stallman went insane.  They have that right, and 
Stallman trying to deny them that right in the name of "freedom" is deeply 
ironic.

> You can argue till the cows come home the belief that _your_
> restrictions are more fair, moral and reasonable than theirs.

It's not the FSF's code being licensed.  It's the Linux developers' code being 
licensed.  The people who write the code get to choose the license.

> But at the end of the day it's all just a matter of opinion about
> what constitutes fair and reasonable.

Why is Linus's opinion less valuable than Stallman's when it comes to the 
license under which the project Linus founded, and which Linus still 
maintains, is distributed?

> You think its a fair trade 
> that you get code back, the FSF think its fair that people can hack
> and run the code anywhere its used..  It all comes down to the
> author of the code getting to attach whatever restrictions they
> choose.

Exactly.  And Linus likes GPLv2.  As do I.

> Sean

Rob

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Adrian Bunk]

On Thu, Jun 14, 2007 at 09:01:32AM -0700, Linus Torvalds wrote:
> On Thu, 14 Jun 2007, Adrian Bunk wrote:
> > 
> > Here in Germany, the rules at court are roughly "the loser pays 
> > everything including the costs of the winner", so if a big company is 
> > sure they will win at court there's no reason not to go there.
> 
> Well, the thing is (and I've said this before), a lawsuit is (and _should_ 
> be) very much a last resort.
> 
> I think that the Open Source community (and the FSF too) is much better 
> off *not* concentrating so much on "legal rules" of what can and cannot be 
> done, and instead spend much more effort on showing people why the whole 
> "Open Source" thing actually works.

I'm wondering more and more why you choose the GPL and not the BSD 
licence for the Linux kernel...

Companies are violating the GPL and this only works as long as noone 
starts taking legal actions against them.

And taking legal actions easily takes all of your fortune at risk.

> And in fact, I think that's _exactly_ what Linux has been doing for the 
> last decade!
> 
> A lot of companies are actually doing the Right Thing (tm).
> 
> Not because of anybody "forcing" them, but because they have literally 
> bought into the whole "Open Source can do things better" mentality. 

And there are some companies for whom it's better if they can take the 
open source code and turn it into some closed thing. There's a reason 
why TiVo warned it's investors that the GPLv3 might hurt their business.

> In fact, the whole "coercive" approach is counter-productive. It makes 
> people dislike you. It makes companies _resist_ open source, rather than 
> see it as a potential ally. 

And what are the risk of your allies actions?

Consider e.g. that your ally AMD offers legally questionable non-GPL 
modules and Debian shipping binaries.

If one of the many copyright holders of the kernel wants to take legal 
actions against this suing AMD might simply be out of reach due to 
financial reasons.

But legal actions against the maintainers of ftp.<your_country>.debian.org 
distributing binaries of these legally questionable modules have a good 
chance of success.

The legal risks might not be a problem for a big company like AMD, but 
anyone seeking an easy legal win will _naturally_ attack mirrors.

And considering the lucratice "cease and desist letter" business in 
Germany, it really seems to be only a matter of time until mirrors will 
have to pay (the lawyer costs of the ones sending the cease and desist 
letters) for the actions of your "allies".

> And no, I'm not speaking out of my *ss. Anybody who goes back fifteen 
> years and looks at how the FSF was acting wrt the GPL (v2, back then), and 
> how many friends - and enemies - they were making, should see that as a 
> big clue. Linux really *did* change the landscape - for the better (*). By 
> being much less contrary.
> 
> So look at Intel in the open source space. They're doing well. Look at 
> Sun. They aren't _forced_ to open-source, they see others open-sourcing, 
> and they see that it works damn well.

Look at AMD.
Look at NVIDIA.

Will they ever switch from providing illegal modules to open source?

> In the "Tivo space", look at Neuros. 
> 
> In other words, we're just *much* better off with a friendly license and 
> not trying to force people to choose sides, than with the rabid idealism 
> that was - and still is - the FSF. The FSF always makes for this horrible 
> "you're with us, or you're against us" black-and-white mentality, where 
> there are "evil" companies (Tivo) and "good" companies (although I dunno 
> if the FSF really sees anybody as truly "good").
> 
> I'd much rather just see "individuals" and "companies". They're not evil 
> or good, they are all in it for their own reasons (and their reasons are 
> *NOT* the same reasons they are for me, you, or anybody else), and we 
> should show them that the whole "Open Source" approach really does work 
> for them.
>...

Until someone starts to attack ftp mirrors and other people distributing 
the questionable code.

There are so many copyright holders of the Linux kernel (including e.g. 
the former Caldera), and if someone wants to spread fear among Linux 
users legal actions against people maintaining ftp mirrors or selling 
their old Linux distribution at ebay will be the cheap and effective way 
to achieve this goal.

> 			Linus
>...

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Thu, 14 Jun 2007, Adrian Bunk wrote:
> 
> I'm wondering more and more why you choose the GPL and not the BSD 
> licence for the Linux kernel...

Why do people confuse "anti-GPLv3" with "pro-BSD"?

What's the logic?

The BSD license is not doing tit-for-tat. It doesn't give me anything 
back. I don't believe in that kind of model. So I'd not use it for my 
projects.

The GPLv2 has a good balance. It encourages tit-for-tat, and it makes sure 
that the software is kept free. And it doesn't try to force anything else, 
or play politics. The only thing you have to believe in is "tit-for-tat".

The GPLv3 goes too far. It's no longer "tit-for-tat", it's "our software 
is worth _soo_ much, that we want to force you to behave well, or you 
cannot use it".

I think one of the above licenses are good. The fact that I reject the 
GPLv3 in _no_ way implies that I should like the BSD license. Both the BSD 
license and the GPLv3 are flawed - they are just flawed in fundamentally 
different ways.

So the whole question of "why don't you use he BSD license then" is just 
fundamentally bogus. A license is about a *balance* of things. "Fairness" 
is not about laissez-faire (BSD) or about total-control (GPLv3). To me, 
It's about something in the middle, where people give back in kind.

And btw, that "to me" is important. 

Different people have different opinions. That's _fine_. Use the GPLv3 for 
your projects. Go wild. Use the BSD license. It's your choice. 

But by the same token, it was _my_ choice (and it was an informed choice) 
to use the GPLv2. 

And to then come in fifteen years later and call me "confused" about a 
license I've chosen is a damn affront to me. I'm not confused. Somebody 
else may be, but it's not me.

		Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Rene Herman]

On 06/14/2007 06:01 PM, Linus Torvalds wrote:

> It's totally pointless to try to "force" people to be good. That's like 
> "curing" gay people. Not going to happen. 

Tangent, but that could in fact quite easily be construed as saying that gay 
people aren't good which I hope is not the point you are making :-/

Rene.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Lennart Sorensen]

On Thu, Jun 14, 2007 at 07:48:03PM +0200, Rene Herman wrote:
> On 06/14/2007 06:01 PM, Linus Torvalds wrote:
> 
> >It's totally pointless to try to "force" people to be good. That's like 
> >"curing" gay people. Not going to happen. 
> 
> Tangent, but that could in fact quite easily be construed as saying that 
> gay people aren't good which I hope is not the point you are making :-/

I certainly read that as 'trying to force people to be good is just as
crazy as trying to force people to not be gay'.  Some people are good,
and some aren't (no idea why), and similarly some people are gay and
some aren't (again, no idea why).  Neither can be changed by declaring
that it must be changed.

I always love Linus' analogies. :)

--
Len Sorensen

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Rene Herman]

On 06/14/2007 09:29 PM, Lennart Sorensen wrote:

> On Thu, Jun 14, 2007 at 07:48:03PM +0200, Rene Herman wrote:
>> On 06/14/2007 06:01 PM, Linus Torvalds wrote:
>>
>>> It's totally pointless to try to "force" people to be good. That's like 
>>> "curing" gay people. Not going to happen.
>> 
>> Tangent, but that could in fact quite easily be construed as saying
>> that gay people aren't good which I hope is not the point you are
>> making :-/
> 
> I certainly read that as 'trying to force people to be good is just as
> crazy as trying to force people to not be gay'.  Some people are good,
> and some aren't (no idea why), and similarly some people are gay and
> some aren't (again, no idea why).  Neither can be changed by declaring
> that it must be changed.

Yes, just my sense of humour, I'm afraid... ;-)

Rene.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:

> In fact, the whole "coercive" approach is counter-productive.

Let me see if I got your position right: when TiVO imposes
restrictions, that's ok, but when others want to find ways to stop it,
then it's not.  *Now* I'm confused ;-)

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Thu, 14 Jun 2007, Alexandre Oliva wrote:
> 
> Let me see if I got your position right: when TiVO imposes
> restrictions, that's ok

Sure. I think it's ok that Microsoft imposes restrictions too on the 
software they create. It's *their* choice.

And I think it's ok for you to impose any restrictions (including the ones 
in GPLv3) on the software *you* produce. It's your choice.

> but when others want to find ways to stop it, then it's not.  *Now* I'm 
> confused ;-)

You are indeed totally confused.

It's *ok* to impose restrictions on the stuff you create. Everybody has a 
different world-view, and for some it's about making money, for some it's 
about something else, and some don't want any restrictions at all.

For me, the GPLv2 was the license I liked. I didn't like the BSD license, 
so I didn't choose it. I don't like a license that restricts hardware, so 
I didn't choose that.

And I *still* don't choose that.

See? I think the GPLv3 is a *much* inferior license to the GPLv2. It's 
better than its drafts were, but it's still doing things I disagree with. 

So tell me, why do you think I'm confused about the GPLv3? Why do you 
think I should have said "GPLv2 or anything else the FSF comes up with"? 

So the only thing I want you to say is:

 (a) Linus knows what he is doing, and isn't actually confused.

and

 (b) It was my right to use the license of my choice for a project that I 
     started.

and

 (c) I have the right to see the difference between the GPLv2 and v3, and 
     think that the GPLv3 is the inferior license.

Comprende? MY CHOICE. Not the FSF's. Not yours. Not anybody elses.

		Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:

> On Thu, 14 Jun 2007, Alexandre Oliva wrote:
>> 
>> Let me see if I got your position right: when TiVO imposes
>> restrictions, that's ok

> Sure. I think it's ok that Microsoft imposes restrictions too on the 
> software they create. It's *their* choice.

Last I looked, TiVO was not the author of Linux.  Did you sell out or
something? ;-P :-D

> So tell me, why do you think I'm confused about the GPLv3?

I think you're confused about the spirit of the GPL, that applies
equally to v1, v2 and v3.

I think you're confused because you claim the GPL is tit-for-tat, that
it encourages/requires (you haven't been consistent) contributions in
kind, but the only contribution in kind is respect for the freedoms of
others.  But then, when measures are introduced to ensure compliance
with this twisted tit-for-tat notion, you claim they're wrong, that
they escape the spirit of the license.

This is why I think you're confused.


That said, it is possible that you disregarded the spirit of the GPL
entirely, focused on some of the legal terms and decided that was
something you wanted for your project.  And that it models what you
want for your project better than GPLv3 does, because GPLv3 takes
the spirit that you disregarded even more seriously than GPLv2.


I still fail to see why what it is in GPLv2 that makes it better to
satisfy your intentions WRT Linux than GPLv3.  I must assume that,
when you say "tit-for-tat", you mean something else, and not respect
for others' freedoms.  If you take the time to explain what it is,
then perhaps it will become clear why you consider the GPLv2 a better
license to achieve your goals, or perhaps it will show that you're
indeed confused about what GPLv2 and GPLv3 mean.

>  (a) Linus knows what he is doing, and isn't actually confused.

I can't say that yet.  Maybe after the points above are sufficiently
explored I will be able to say that.

>  (b) It was my right to use the license of my choice for a project that I 
>      started.

No doubt about it.

>  (c) I have the right to see the difference between the GPLv2 and v3, and 
>      think that the GPLv3 is the inferior license.

You sure do have that right.

> Comprende? MY CHOICE. Not the FSF's. Not yours. Not anybody elses.

Until you started accepting contributions from others, yes.


BTW, in Portuguese the correct spelling would be "compreende", with a
double 'e'.  "Comprende" is Spanish, and in Brazil, where I live, we
speak Portuguese.  But thanks for trying, that's appreciated ;-)

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Thu, 14 Jun 2007, Alexandre Oliva wrote:
> 
> Last I looked, TiVO was not the author of Linux.  Did you sell out or
> something? ;-P :-D

You're a moron.

I'm the original author, and I selected the GPLv2 for Linux.

Tivo accepted that, and followed the GPLv2. Even the FSF lawyers agreed 
that they followed it.

But Tivo *is* the "author" of their own proprietary applications, and it 
*is* the designer of their hardware.

And exactly like I had the right to the choice of license when it comes to 
Linux, they have the right of choice to license and behaviour when it 
comes to *their* software and hardware (that is not a derived work of 
Linux).

But you cannot follow a coherent argument, because you dont' *want* to 
follow it. Because following the logical argument would take you to a 
place where you don't want to be.

I'm not going to bother discussing this any more. You don't seem to 
respect my right to choose the license for my own code.

So one final time:

 - I chose the GPLv2, fully understanding that the Tivo kind of situation 
   is ok.

 - the FSF lawyers too have acknowledged that what Tivo did was not a 
   license violation, so I obviously am not confused about the issue: YOU 
   are.

 - I think that what Tivo did was not only "technically valid" by the 
   license, it was what I *intended* by my choice of license!

And you are apparently totally unable to understand - or respect - that I 
actually made an informed decision that happens to be different from what 
you *wish* it were.

		Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:

>  - I chose the GPLv2, fully understanding that the Tivo kind of
>  situation is ok.

Wow, do you remember the date when you first thought of this business
model?

> And you are apparently totally unable to understand - or respect - that I 
> actually made an informed decision that happens to be different from what 
> you *wish* it were.

While you insist in the nonsensical tit-for-tat argument and "in kind"
retributions, I don't think I have much of a choice, because this is
not what the GPL is about, this is not what it requires of licensees.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Thu, 14 Jun 2007, Alexandre Oliva wrote:
>
> On Jun 14, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:
> 
> >  - I chose the GPLv2, fully understanding that the Tivo kind of
> >  situation is ok.
> 
> Wow, do you remember the date when you first thought of this business
> model?

You know what? I'm intelligent. That's what you call people who see th 
consequences of their actions. I didn't see the *details* of what all the 
GPLv2 could result in, but yes, I claim that I knew what I was setting 
myself up for (in a license way) pretty much from the beginning.

Did it take me by surprise how people actually ended up using Linux? It 
sure did. But has the GPLv2 itself ever surprised me? Not really. I read 
it back then, and yes, I understood what it meant.

>From the very beginning of Linux, even before I chose the GPLv2 as the 
license, the thing I cared about was that source code be freely available. 
That was the first license, but more importantly, it was why I started 
Linux in the first place - my frustrations with Minix, and my memories of 
how painful it was to find an OS that I wanted to use and work with.

(That, btw, was not Minix-only: I actually originally was thinking about 
literally buying a commercial Unix for my PC too. The price factor kept me 
away from the commercial unixes, and in retrospect I'm obviously very 
happy).

So my first goal was "source must be available and it must be free (as 
in beer)". Which my first copyright license reflects very directly.

What happened a few months into the thing was that some people actually 
wanted to make floppy images of Linux available to Linux users groups, but 
they didn't want to have to actually *fund* the floppies and their work 
themselves, so they wanted to sell them at cost (which the first license 
actually didn't allow!).

And I realized that the money angle really wasn't what I ever really cared 
about. I cared about availability, but people sure could get paid for 
their effort in distributing the thing, as long as the source code 
remained open. I didn't want money, I didn't want hardware, I just wanted 
the improvements back.

So given that background, which license do you _think_ I should have 
chosen?

And given that background, do you see why the GPLv2 is _still_ better than 
the GPLv3? I don't care about the hardware. I'll use it, but it's not what 
Linux is all about. Linux is about something much bigger than any 
individual device.

And yeah, maybe I'm just better at abstracting things. Maybe I prefer 
seeing the big picture, and that the individual devices don't matter. What 
matters is the improvement in the *software*, because while each physical 
device is a one-off thing, in the long term, it's the *development* that 
matters.

And the GPLv2 protects that. 

It's a bit like evolution: individual organisms matter to *themselves* and 
to their immediate neighborhood, but in the end, the individuals will be 
gone and forgotten, and what remains is the development.

In those terms, I care about the DNA, and the *process* or recombination 
and the bigger picture. Any individual organism? Not so much. It's all 
part of a much bigger tapestry, and closed hardware is more like an eunuch 
(or a worker bee): it won't pass on its legacy, but it might help the 
people who do.

So instead of thinking of Tivo as something "evil", I think of Tivo as the 
working bee who will never pass on its genes, but it actually ended up 
helping the people who *do* pass on their genes: the kernel (to a small 
degree - not so much because of the patches themselves, as the *mindshare* 
in the PVR space) and projects like MythTV (again, not so much because of 
any patches, but because it helped grow peoples understanding of the 
problem space!).

Let's take another example: BitKeeper. The FSF follower people seem to 
view BitKeeper as something "evil". To me, BitKeeper was not just a great 
tool, but it also ended up being something that showed others how things 
*could* be done. And the world - including the open source world - is a 
better place for it!

See? In the big picture, individual devices and even projects won't 
matter. In a hundred years, I'll be long dead, and nobody will care. But 
in a hundred years, I hope that the "live and let live" open source 
mentality will still flourish, and maybe "Linux" itself won't live on, but 
some of the memories and impact may. And *that* is what matters.

A Tivo? It's just a toy. Who cares? It's not important. But source code 
that evolves? THAT can change the world!

			Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:

> From the very beginning of Linux, even before I chose the GPLv2 as the 
> license, the thing I cared about was that source code be freely available. 

Ok, the MIT license could get you that.  Even public domain could.

> I didn't want money, I didn't want hardware, I just wanted the
> improvements back.

GPL won't get you that.  You want a non-Free Software license.

It will only as long as people play along nicely and perceive the
benefits of cooperation.  But some players don't.

> So given that background, which license do you _think_ I should have 
> chosen?

I can't morally recommend a non-Free Software license.

> And given that background, do you see why the GPLv2 is _still_ better than 
> the GPLv3?

No.  Honestly, I really don't.  Even when I try and look at it from
your perspective, that you described very beautifully in the rest of
the message that I snipped, it's still a mistery to me why you think
permitting Tivoization could possibly be advantageous to your project.

What is it in the anti-Tivoization provision that gets you any less
improvements back?

If anything, I'd think that, by not permitting TiVO to prohibit users
from running modified versions of your code that they don't authorize
themselves, these users would do *more* than TiVO alone ever could,
and if a fraction of them contributes something back, you're way
better off.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Lennart Sorensen]

On Thu, Jun 14, 2007 at 07:31:52PM -0300, Alexandre Oliva wrote:
> Ok, the MIT license could get you that.  Even public domain could.

Those would not ensure that the source code stays free.

> > I didn't want money, I didn't want hardware, I just wanted the
> > improvements back.
> 
> GPL won't get you that.  You want a non-Free Software license.
> 
> It will only as long as people play along nicely and perceive the
> benefits of cooperation.  But some players don't.

It seems to work very well in practice though.

> No.  Honestly, I really don't.  Even when I try and look at it from
> your perspective, that you described very beautifully in the rest of
> the message that I snipped, it's still a mistery to me why you think
> permitting Tivoization could possibly be advantageous to your project.

Perhaps there is no benefit in permitting "Tivoization".  But at the
same time, perhaps there are benefits in not preventing "Tivoization" in
ways that may or may not be foreseen at this time.

> What is it in the anti-Tivoization provision that gets you any less
> improvements back?

Tivo has provided some code changes and improvements to Linux.  If they
had been totally unable to use Linux due to the license, they would
probably have used vxworks or BSD or something else, and Linux would
have gotten nothing back.  So the Linux source code improved and other
systems using the linux code base got better as a result.

> If anything, I'd think that, by not permitting TiVO to prohibit users
> from running modified versions of your code that they don't authorize
> themselves, these users would do *more* than TiVO alone ever could,
> and if a fraction of them contributes something back, you're way
> better off.

Users of the Tivo hardware would be able to do more, sure, but then
again, actualyl, maybe not.  After all if it ran vxworks or bsd, the
user still wouldn't be able to do anything about it.  The end result is
the same.  The answer is also still the same: Don't buy a tivo if you
want to change what it does, because it doesn't let you do that.

--
Len Sorensen

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Thu, 14 Jun 2007, Alexandre Oliva wrote:

> On Jun 14, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:
> 
> > From the very beginning of Linux, even before I chose the GPLv2 as the 
> > license, the thing I cared about was that source code be freely available. 
> 
> Ok, the MIT license could get you that.  Even public domain could.

Why do you bother sending out emails that just show that you cannot read 
or understand?

I want not just the code *I* write to be freely available. I want the 
modifications that others release that are based on my code to be freely 
available too!

That's what the whole "tit-for-tat" thing was all about!

Doyou even understand what "tit-for-tat" means?

Should I use another phrase? Do you understand the phrase "Quid pro quo"? 
Which is another phrase I've used to explain this over the years.

> > I didn't want money, I didn't want hardware, I just wanted the
> > improvements back.
> 
> GPL won't get you that.  You want a non-Free Software license.
> 
> It will only as long as people play along nicely and perceive the
> benefits of cooperation.  But some players don't.

You are living in some alternate world. The GPLv2 gives me exactly what I 
looked for.

Yes, people can do improvements in private, and by keeping them private 
they'll never need to release them to anybody else. Big deal. I don't 
care. By keeping them private, I never see the end result anyway, so they 
"don't exist" as far as I'm concerned.

> > And given that background, do you see why the GPLv2 is _still_ better than 
> > the GPLv3?
> 
> No.  Honestly, I really don't. 

Yeah. So stop bothering me then. Go cry on somebody elses shoulder. Just 
accept the fact that I'm a grown person, in full control of my faculties, 
and that I'm perfectly able to make my own judgements, and that I don't 
need to follow the FSF blindly.

And it doesn't even matter if you don't understand me. That is, as I've 
said, _your_ problem.  I've done my best to explain to you, but if you are 
so limited that you cannot understand that other people have other 
opinions than yours, there really is only so much I can do for you.

Go away. 

			Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:

> On Thu, 14 Jun 2007, Alexandre Oliva wrote:

>> On Jun 14, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:
>> 
>> > From the very beginning of Linux, even before I chose the GPLv2 as the 
>> > license, the thing I cared about was that source code be freely available. 
>> 
>> Ok, the MIT license could get you that.  Even public domain could.

> Why do you bother sending out emails that just show that you cannot read 
> or understand?

Because I can't divine what's in your mind, and if you don't make the
points you want to make clear, I may very well fail to understand.

> I want not just the code *I* write to be freely available. I want the 
> modifications that others release that are based on my code to be freely 
> available too!

With the exception of those who choose not to distribute their
changes.  Or who choose to distribute their changes to people who are
not willing to share them with you.  Fair enough.

> That's what the whole "tit-for-tat" thing was all about!

> Doyou even understand what "tit-for-tat" means?

Yes.  I even wrote an article about that.

http://fsfla.org/svnwiki/blogs/lxo/draft/gplv3-snowwhite

> Should I use another phrase? Do you understand the phrase "Quid pro quo"? 

Yes.  It's there in the article as well.  The difference is basically
in attitude.  Tit-for-tat is adversarial (equivalent retaliation),
whereas Quid pro quo is cooperative (a favor for a favor).

> Which is another phrase I've used to explain this over the years.

Yup, I remember that.

>> > I didn't want money, I didn't want hardware, I just wanted the
>> > improvements back.

>> GPL won't get you that.  You want a non-Free Software license.

>> It will only as long as people play along nicely and perceive the
>> benefits of cooperation.  But some players don't.

> You are living in some alternate world. The GPLv2 gives me exactly
> what I looked for.

And in what sense the GPLv3 anti-Tivoization clause doesn't?

In what sense does it give you *less* of what you want?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

RE: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Schwartz]

> Do you even understand what "tit-for-tat" means?

I don't. Could someone please explain it. Specifically:

1) What is "tat"?

2) How can I get some?

3) Where do I go to trade it in?

DS

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Tomas Neme]

> 1) What is "tat"?
>
> 2) How can I get some?
>
> 3) Where do I go to trade it in?

4) is it legal to consume it in my country?

5) should I have a designed driver when I do?

-- 
|_|0|_|
|_|_|0|
|0|0|0|

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bron Gondwana]

On Fri, Jun 15, 2007 at 11:24:08AM -0300, Tomas Neme wrote:
>> 1) What is "tat"?
>>
>> 2) How can I get some?
>>
>> 3) Where do I go to trade it in?
>
> 4) is it legal to consume it in my country?
>
> 5) should I have a designed driver when I do?

6) Is that allowed to be a binary-only driver or
   does it have to come with source code?

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Carlo Wood]

On Thu, Jun 14, 2007 at 01:09:46PM -0700, Linus Torvalds wrote:
> I'm the original author, and I selected the GPLv2 for Linux.
[...]
> I'm not going to bother discussing this any more. You don't seem to 
> respect my right to choose the license for my own code.

This is the main reason I dislike GPLwhatever: there is no notion
of "orginal author". You might have written 99% of the code, that
doesn't matter. You have no rights whatsoever once you release
something under the GPL (no more than ANYOne else).

The GPL is nice for the community, and for the users - but very,
very bad towards it's authors (taking all and every right you might
have). If John Doe wants to re-release the whole kernel under
GPLv3, then all he needs is a website and some bandwidth.

-- 
Carlo Wood <carlo@alinoe.com>

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alan Cox]

> This is the main reason I dislike GPLwhatever: there is no notion
> of "orginal author". You might have written 99% of the code, that

Every literary work (including thus software) has an author, and that
author has certain rights which are implicit in them being author.

> doesn't matter. You have no rights whatsoever once you release
> something under the GPL (no more than ANYOne else).

Wrong. The author has a collection of rights which vary by jurisdiction
but which are primarily governed by the Berne Convention and its sequels
notably TRIPS.

> The GPL is nice for the community, and for the users - but very,
> very bad towards it's authors (taking all and every right you might
> have). If John Doe wants to re-release the whole kernel under

You must be using a different GPL to the rest of us.

> GPLv3, then all he needs is a website and some bandwidth.

And a very good lawyer (oh and a GPL3 as there isn't one yet...)

Alan

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Carlo Wood]

On Fri, Jun 15, 2007 at 12:39:19AM +0100, Alan Cox wrote:
> > This is the main reason I dislike GPLwhatever: there is no notion
> > of "orginal author". You might have written 99% of the code, that
> 
> Every literary work (including thus software) has an author, and that
> author has certain rights which are implicit in them being author.

Like, they can release/sell the whole thing under some arbitrary
other license at their choice. But once you license it with the GPLv2,
then you can't stop anyone else (who got it under that license) from
using the code under that license anymore, as such it doesn't matter that
you are the original author.

> > doesn't matter. You have no rights whatsoever once you release
> > something under the GPL (no more than ANYOne else).
> 
> Wrong. The author has a collection of rights which vary by jurisdiction
> but which are primarily governed by the Berne Convention and its sequels
> notably TRIPS.
> 
> > The GPL is nice for the community, and for the users - but very,
> > very bad towards it's authors (taking all and every right you might
> > have). If John Doe wants to re-release the whole kernel under
> 
> You must be using a different GPL to the rest of us.

You actually had me check the license of the linux kernel :p
But really - it has this paragraph that I was refering to in most of it's
source files:

* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.

I might be wrong, but I always thought that that meant that John Doe
is free to redistribute the software under version 3 of the License,
as published by the FSF.

There are source files in the kernel without this phrase, ie - they
just say: * This file is released under the GPL.

But then the paragraph from COPYING kicks in, reading:

Each version is given a distinguishing version number.  If the Program
specifies a version number of this License which applies to it and "any
later version", you have the option of following the terms and conditions
either of that version or of any later version published by the Free
Software Foundation.  If the Program does not specify a version number of
this License, you may choose any version ever published by the Free Software
Foundation.

Any, 'any version' probably includes version 3 as well.

Finally, there are file that don't mention the GPL at all, for example
kernel/sys.c just says:

 *  Copyright (C) 1991, 1992  Linus Torvalds

But - if it weren't GPL-ed then that would be a violation of the GPL-ed
of the rest (Nevertheless, I think the license header should be added
to those files).

> > GPLv3, then all he needs is a website and some bandwidth.
> 
> And a very good lawyer (oh and a GPL3 as there isn't one yet...)

I really don't like license discussions - and after reading in the
mailinglist FAQ that license posts are taboo here - I was partly annoyed,
partly amazed to see this HUGE flood of mails with as subject line
"Dual-Licensing Linux Kernel with GPL V2 and GPL V3"

I am sorry that I drew the apparently wrong conclusion that GPL V3
is a pressing reality.

> Alan

-- 
Carlo Wood <carlo@alinoe.com>

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alan Cox]

> Like, they can release/sell the whole thing under some arbitrary
> other license at their choice. But once you license it with the GPLv2,
> then you can't stop anyone else (who got it under that license) from
> using the code under that license anymore, as such it doesn't matter that
> you are the original author.

This is true of most licences. Ask musicians about trying to get their
music back from a record company. 

> * This program is free software; you can redistribute it and/or modify
> * it under the terms of the GNU General Public License as published by
> * the Free Software Foundation; either version 2 of the License, or
> * (at your option) any later version.
> 
> I might be wrong, but I always thought that that meant that John Doe
> is free to redistribute the software under version 3 of the License,
> as published by the FSF.

For those marked parts yes - thats the authors choice. Some of the kernel
is dual licensed BSD even so you can use that bit for all sorts of stuff.
Again authors choice, some authors wanted to share code between Linux and
other projects. I believe you can buy proprietary licenses to reiserfs
too.

Some authors like GPLv2 or later, some don't trust the FSF, some will
decide once GPLv3 is out, some couldn't care etc.. as the kernel doesn't
do copyright assignment all these wishes are respected and that is how it
should be.

Alan

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Fri, 15 Jun 2007, Carlo Wood wrote:
> 
> But then the paragraph from COPYING kicks in, reading:

Read the COPYING file more closely, and realize that "the Program" has 
always specified a version number of this license.

It used to include it just by virtue of having the COPYING file *itself* 
be included (and that's v2), but since some people felt that was unclear, 
the COPYING file has this language pretty visibly at the top:

 Also note that the only valid version of the GPL as far as the kernel
 is concerned is _this_ particular version of the license (ie v2, not
 v2.2 or v3.x or whatever), unless explicitly otherwise stated.

> Any, 'any version' probably includes version 3 as well.

Not for the kernel. Exactly because the kernel _does_ specify the version. 
So the sequence you quoted is a non-issue.

In other words: any file that does not *explicitly* say that it's "v2 or 
later" is v2 only.

		Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Jeremy Maitin-Shepard]

Carlo Wood <carlo@alinoe.com> writes:

> On Thu, Jun 14, 2007 at 01:09:46PM -0700, Linus Torvalds wrote:
>> I'm the original author, and I selected the GPLv2 for Linux.
> [...]
>> I'm not going to bother discussing this any more. You don't seem to 
>> respect my right to choose the license for my own code.

> This is the main reason I dislike GPLwhatever: there is no notion
> of "orginal author". You might have written 99% of the code, that
> doesn't matter. You have no rights whatsoever once you release
> something under the GPL (no more than ANYOne else).

You retain the copyright, and in particular the right to relicense.
Only if you make the mistake of including the "or any later version"
phrase do you allow others to redistribute the work under a different
version of the GPL.  Although this provision may seem slightly
convenient to authors, its effect is to grant a very large amount of
relicensing permission to the FSF.  It almost certainly doesn't make
sense to place that much trust in a single organization.

> The GPL is nice for the community, and for the users - but very,
> very bad towards it's authors (taking all and every right you might
> have). If John Doe wants to re-release the whole kernel under
> GPLv3, then all he needs is a website and some bandwidth.

Well, he also needs one tiny little extra thing: the permission of every
copyright holder in Linux.

-- 
Jeremy Maitin-Shepard

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Rob Landley]

On Thursday 14 June 2007 19:18:12 Carlo Wood wrote:
> On Thu, Jun 14, 2007 at 01:09:46PM -0700, Linus Torvalds wrote:
> > I'm the original author, and I selected the GPLv2 for Linux.
>
> [...]
>
> > I'm not going to bother discussing this any more. You don't seem to
> > respect my right to choose the license for my own code.
>
> This is the main reason I dislike GPLwhatever: there is no notion
> of "orginal author". You might have written 99% of the code, that
> doesn't matter. You have no rights whatsoever once you release
> something under the GPL (no more than ANYOne else).

You mean if the original author gets hit by a bus and their estate gets sold 
to SCO they can't revoke our rights to the code?  How is this a down side?

And you do have more rights than anyone else: as the copyright holder you can 
issue other licenses, and you have standing to sue to enforce the code.  (If 
nobody else has a copyright on the code, they don't have standing to sue to 
enforce the license terms.)

(Right now, nobody EXCEPT the FSF has the right to sue somebody to enforce the 
license terms on something like gcc.  Do you find that a comforting thought?)

Rob 
-- 
"One of my most productive days was throwing away 1000 lines of code."
  - Ken Thompson.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bernd Paysan]

[-- Attachment #1: Type: text/plain, Size: 2732 bytes --]

On Friday 15 June 2007 03:49, Rob Landley wrote:
> (Right now, nobody EXCEPT the FSF has the right to sue somebody to
> enforce the license terms on something like gcc.  Do you find that a
> comforting thought?)

Have you ever signed a copyright transfer agreement to the FSF? Obviously 
not, because then you wouldn't utter such nonsense. The agreement reads 
that you transfer a non-exclusive right to the FSF to distribute the code 
under GPL (versions of your choice, they have this right anyway, but making 
it explicit is always good), and the right to enforce the license. You 
still have the right to relicense the work as you like. You also have the 
right to enforce the license yourself, or to transfer that right to 
somebody else like gpl-violations.org. The FSF even doesn't require to 
transfer copyright if you make a GNU project, but if you don't, the FSF 
won't help you (because they can't).

They make very obvious promises about what they care ("four freedoms"), and 
that they will be very consistent in doing so. So far, all track records 
have proven that they indeed are very consistent in doing so - the main 
controversy here is not whether the FSF protects the "four freedoms", but 
whether these four freedoms are the right goal, and if they really should 
try so hard to protect these four freedoms. This part of the discussion is 
fully acceptable, what's not acceptable is that the Linus-fancurve claims 
things the GPL sais which it doesn't (like "tit-for-tat") or doesn't say 
which it does (like section 6 - direct license from the licensor, and in 
cases like Linux where no copyright transfer agreements whatsoever exist, 
these are the individual contributors). Or that Linux 0.something was 
already under GPLv2 only, when GPLv2 clearly says that there may be 
updates, and when you as author don't say something, you are allowing users 
to update if they like.

The last point IMHO makes clear that my interpretation of the comment is 
valid: This is a commend made by Linus Torvalds, as how he understands or 
misunderstands the license text. It's not even something you can take as 
legal advice, because Linus is not a lawyer (fortunately - think how the 
kernel would look like if it was programmed by a lawyer ;-).

Sure, if you as outsider strip the kernel of obvious GPLv2-only code to 
relicense it as a whole under GPLv3, you need a good asbestos suite, a good 
lawyer, and good arguments. But let's assume Microsoft really succeeds with 
its patent FUD against Linux, and the only way out is GPLv3, when will 
opinions here change?

-- 
Bernd Paysan
"If you want it done right, you have to do it yourself"
http://www.jwdt.com/~paysan/

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Rob Landley]

On Friday 15 June 2007 07:57:10 Bernd Paysan wrote:
> On Friday 15 June 2007 03:49, Rob Landley wrote:
> > (Right now, nobody EXCEPT the FSF has the right to sue somebody to
> > enforce the license terms on something like gcc.  Do you find that a
> > comforting thought?)
>
> Have you ever signed a copyright transfer agreement to the FSF?

Nope, and I probably never will, but I'm glad to be wrong about this one.

> But let's assume Microsoft really succeeds with
> its patent FUD against Linux, and the only way out is GPLv3, when will
> opinions here change?

Ever heard of the Open Invention Network or the Software Freedom Law Center?

And since when did the FSF become an ally of Microsoft?  The rest of us seem 
to think the Novell deal's a minefield under GPLv2, and would just LOVE 
Microsoft to reopen the antitrust can of worms by suing somebody over Linux 
(they're still a convicted predatory monopolist, only the sentence part was 
defused)...

And yet GPLv3 was DOA until the Novell deal happened and suddenly the FSF had 
this huge thing.  Reminds me of everybody rallying behind Bush after 
September 11 to support abominations like the Patriot Act.   It's sad how 
fear makes people stop thinking.

Rob
-- 
"One of my most productive days was throwing away 1000 lines of code."
  - Ken Thompson.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Lennart Sorensen]

On Fri, Jun 15, 2007 at 01:57:10PM +0200, Bernd Paysan wrote:
> Have you ever signed a copyright transfer agreement to the FSF? Obviously 
> not, because then you wouldn't utter such nonsense. The agreement reads 
> that you transfer a non-exclusive right to the FSF to distribute the code 
> under GPL (versions of your choice, they have this right anyway, but making 
> it explicit is always good), and the right to enforce the license. You 
> still have the right to relicense the work as you like. You also have the 
> right to enforce the license yourself, or to transfer that right to 
> somebody else like gpl-violations.org. The FSF even doesn't require to 
> transfer copyright if you make a GNU project, but if you don't, the FSF 
> won't help you (because they can't).
> 
> They make very obvious promises about what they care ("four freedoms"), and 
> that they will be very consistent in doing so. So far, all track records 
> have proven that they indeed are very consistent in doing so - the main 
> controversy here is not whether the FSF protects the "four freedoms", but 
> whether these four freedoms are the right goal, and if they really should 
> try so hard to protect these four freedoms. This part of the discussion is 
> fully acceptable, what's not acceptable is that the Linus-fancurve claims 
> things the GPL sais which it doesn't (like "tit-for-tat") or doesn't say 
> which it does (like section 6 - direct license from the licensor, and in 
> cases like Linux where no copyright transfer agreements whatsoever exist, 
> these are the individual contributors). Or that Linux 0.something was 
> already under GPLv2 only, when GPLv2 clearly says that there may be 
> updates, and when you as author don't say something, you are allowing users 
> to update if they like.

So if Linus wasn't entirely clear to begin with that he wanted GPLv2
only, then that is just too bad?  Well why not then say that if the
GPLv2 didn't say that what Tivo did was bad, then too bad, you can't
change your mind later.  The FSF can't have it both ways.

> The last point IMHO makes clear that my interpretation of the comment is 
> valid: This is a commend made by Linus Torvalds, as how he understands or 
> misunderstands the license text. It's not even something you can take as 
> legal advice, because Linus is not a lawyer (fortunately - think how the 
> kernel would look like if it was programmed by a lawyer ;-).
> 
> Sure, if you as outsider strip the kernel of obvious GPLv2-only code to 
> relicense it as a whole under GPLv3, you need a good asbestos suite, a good 
> lawyer, and good arguments. But let's assume Microsoft really succeeds with 
> its patent FUD against Linux, and the only way out is GPLv3, when will 
> opinions here change?

No the GPLv3 is not the only solution.  Perhaps a GPLv2.1 that is
actually in the spirit of the VPLv2 with just enough changes to fix the
real problems, that may in fact be a solution.  The GPLv3 is very much
not the only solution.  A totally different kernel specific license may
even be an option.  There are many options, most of them just happen to
be pretty hard.

--
Len Sorensen

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Al Viro]

On Fri, Jun 15, 2007 at 02:04:45PM -0400, Lennart Sorensen wrote:
> So if Linus wasn't entirely clear to begin with that he wanted GPLv2
> only, then that is just too bad?  Well why not then say that if the
> GPLv2 didn't say that what Tivo did was bad, then too bad, you can't
> change your mind later.  The FSF can't have it both ways.

Of course they can - as any politician.  What, do you really think that
Bernd wakes up, looks in a mirror and spits in disgust?  I very much
doubt that.

Here's how that kind of stuff works: to be a successful politician one
needs to have ideals.  The more, the better.  Then there always will
be an ideal advanced by action one wants to take.  Experienced politician
is one that will be able to pick those automatically.  And feel damn
righteous at that.

Of course they can have it both ways - just say something like "we work
for the greater good of users", "what advances our goals helps the
interests of users", "we'll read both cases in a way that helps the
interests of users", "we are consistent in helping the interests of users".
And at that point any talk of inconsistencies will be deflected by the
highly moral considerations.  End of story.  Sprinkle with generic
constructs making the critics easier to ignore ("they do it because
they are Bad"/"they do it because they had been seduced by Evil"/
"they just parrot somebody"/"they do it because of <random psychobubble>"/
"they do it because of wrong ideology"/"they do it because of ego") and enjoy.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Diego Calleja]

El Thu, 14 Jun 2007 14:49:19 -0300, Alexandre Oliva <aoliva@redhat.com> escribió:

> Let me see if I got your position right: when TiVO imposes
> restrictions, that's ok, but when others want to find ways to stop it,
> then it's not.  *Now* I'm confused ;-)


Me, I agree that hardware shouldn't lock users. And since I'm one of those
evil european socialdemocrats, I may go as far as to think that there should
be laws that *forbid* selling such hardware.

But I think that all this iss a *hardware* issue. It seems to me that lot of
people at the FSF wants to regulate the hardware industry using the influence
of free/open software in the computing industry and the "V2 or later" phrase
from the GPL. 

But the fact is that free/open source runs on _top_ of hardware. You don't
control hardware, you only control the things that are built on top of your
software, not the parts you use to build your software.

And the FSF is trying to control the design and licensing of hardware throught
the influence of their software. And I think it's wrong. I'm all to forbid hardware
that imposes restrictions on hardware, but software licenses are NOT the way
to make it. That's a task for a "Free Hardware Foundation", not the FSF.

What the FSF is trying to do is EVIL. It's not about free software, it's
not about freedom, it's about the FSF trying to have to much control over
things that they shouldn't even try to control. I think that the FSF can do 
a terrible damage to free/open source with such stupid ideas. I wouldn't
even be surprised that some jugde rules that a software license that tries
to 'control' hardware is invalid

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Thu, 14 Jun 2007, Diego Calleja wrote:
> 
> And the FSF is trying to control the design and licensing of hardware throught
> the influence of their software. And I think it's wrong. I'm all to forbid hardware
> that imposes restrictions on hardware, but software licenses are NOT the way
> to make it. That's a task for a "Free Hardware Foundation", not the FSF.

Amen. And btw, opencores.org does actually exist.

I don't even think open hardware is a big issue: the worry-warts about 
hardware are likely wrong, and hardware today is a lot more open than it 
used to be even just a decade ago. You can much more easily design your 
own (FPGA's are cheap and powerful), and yes, it's more complex today, but 
that's actually an argument _for_ openness rather than against it (open 
processes work better in complex environments!).

The real issue is "open content", and we do actually have various 
organizations that support that in particular. I would heartily encourage 
people to get involved with the Creative Commons, and the EFF, and I think 
Larry Lessig is a really smart and articulate person, who you should 
listen to.

> What the FSF is trying to do is EVIL.

I wouldn't go that far (although, in the heat of the moment I probably 
_have_ gone that far. Oops ;).

I don't think the FSF is evil. They're just too single-minded, and look 
too much at one issue, and only care about the one thing they care about, 
and in the process, they tend to have a really hard time seeing the other 
side of the coin.

They define "freedom" one way, and by defining it in a very particular 
way, they miss the fact that what is "freedom" to them is not "freedom" to 
somebody else.

They have a very particular agenda, and in having that agenda and a very 
strict view of how the world should look (according to the FSF), they 
dismiss the fact that other people have _other_ agenda's, and see the same 
world totally differently.

And I think that kind of single-mindedness is silly and 
counter-productive. 

I literally think that the GPLv2 has worked so well exactly because you 
can strip it of its high-falutin' morality and the FSF Kool-Aid, and just 
see it as a "tit-for-tat" license. It allows everybody to see that the 
work they put in (into the _software_) is protected, and people cannot 
make improved versions of that software and distribute those improved 
versions without giving you the right back to use those improvements (to 
the _software_).

So the GPLv2 may have come out of a very single-minded endeavor, but I 
think it ended up being capable of so much more than rms really even 
envisioned, exactly because you don't have to _view_ it in that manner.

		Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:

> On Thu, 14 Jun 2007, Diego Calleja wrote:

>> And the FSF is trying to control the design and licensing of
>> hardware throught the influence of their software.

It's not.  It's only working to ensure recipients of the Free Software
can modify and share the software.

>> What the FSF is trying to do is EVIL.

> I wouldn't go that far (although, in the heat of the moment I probably 
> _have_ gone that far. Oops ;).

:-)

> I literally think that the GPLv2 has worked so well exactly because you 
> can strip it of its high-falutin' morality and the FSF Kool-Aid, and just 
> see it as a "tit-for-tat" license. It allows everybody to see that the 
> work they put in (into the _software_) is protected, and people cannot 
> make improved versions of that software and distribute those improved 
> versions without giving you the right back to use those improvements (to 
> the _software_).

Can you explain to me how it is that the Tivoization provisions (the
only objection you have to GPLv3) conflict with this?

(nevermind our disagreement as to whether "tit-for-tat" applies to
 either GPLv2 or GPLv3)

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

RE: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Schwartz]

> Can you explain to me how it is that the Tivoization provisions (the
> only objection you have to GPLv3) conflict with this?

Is it really that hard to understand? GPLv2 applied only to works people
chose to place under that license or to works that contain so much code that
someone chose to place under that license that they are legally considered a
derivative work. GPLv3, on the other hand, attempts to extend control over
works that don't contain any code that anyone ever chose to place under the
GPL.

This is a night and day difference.

The GPLv2 stands within the legal scope of copyright. If I create a work, I
have some rights to control that work. If you create a work *based* *on*
*my* *work* I can retain some rights over how this new work is used because
it actually *contains* parts of my work in it.

The GPLv2 makes no attempt to exercise any control over anything else. The
GPLv3, however, attempts to leverage copyright control to restrict what can
be done with things completely outside the covered works.

DS

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bongani Hlope]

On Thursday 14 June 2007 21:55:09 Alexandre Oliva wrote:
> On Jun 14, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:
> > On Thu, 14 Jun 2007, Diego Calleja wrote:
> >> And the FSF is trying to control the design and licensing of
> >> hardware throught the influence of their software.
>
> It's not.  It's only working to ensure recipients of the Free Software
> can modify and share the software.
                                       ^^^^^
Exactly what has been said to you the whole time, but you still refuse to 
accept that. If Linus develops and runs his code on a PowerPC and I struggle 
to install the code that he has released for me to modify and share on a 
PowerPC (maybe because I'm an idiot). Should I create a license with a 
Linusation term, because he is evil he runs his code on a PowerPC and I 
can't?

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Bongani Hlope <bonganilinux@mweb.co.za> wrote:

> On Thursday 14 June 2007 21:55:09 Alexandre Oliva wrote:
>> On Jun 14, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:
>> > On Thu, 14 Jun 2007, Diego Calleja wrote:
>> >> And the FSF is trying to control the design and licensing of
>> >> hardware throught the influence of their software.
>> 
>> It's not.  It's only working to ensure recipients of the Free Software
>> can modify and share the software.
>                                        ^^^^^
> Exactly what has been said to you the whole time, but you still refuse to 
> accept that. If Linus develops and runs his code on a PowerPC and I struggle 
> to install the code that he has released for me to modify and share on a 
> PowerPC (maybe because I'm an idiot). Should I create a license with a 
> Linusation term, because he is evil he runs his code on a PowerPC and I 
> can't?

Depends.  In this hypothetical scenario, what did he do to stop you
from installing and running the modified version (or even the pristine
version, merely recompiled) in your PowerPC?  And why did he do it?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Diego Calleja]

El Thu, 14 Jun 2007 16:55:09 -0300, Alexandre Oliva <aoliva@redhat.com> escribió:

> > On Thu, 14 Jun 2007, Diego Calleja wrote:
> 
> >> And the FSF is trying to control the design and licensing of
> >> hardware throught the influence of their software.
> 
> It's not.  It's only working to ensure recipients of the Free Software
> can modify and share the software.

Those may be the intentions, but I claim that your statement is false. The
anti-tivoisation FSF movement is not "working to ensure recipients of the
Free Software can modify and share the software". 

They can't, because the fact is that hardware vendors can NOT stop you
from "modifing and sharing the software". They only can stop you from
running your modifications, which is very different. So this is a 
_hardware_ limitation. It's pointless to try to address this problem
with software licenses.

What the anti-tivoisation movement is trying to do: "If you are a vendor
of tivoized hardware you must give your users whatever information is
needed to run modifications of their software"

How it works in the real world: "You can't run this software in hardware
that doesn't allow to run code modifications of this software"

So while the anti-tivoisation movement is trying to limit hardware
design/licensing, the fact is that what you are restricting is not the
hardware, but the _software_, in a way very different from the 'restrictions'
that the GPL has when compared with the BSD ie: in a way that doesn't
benefit freedom or contribution of code. Because your users already
can modify and share their code regardless of what hardware they're
using (even if they can't run their modifications), you're just adding
pointless prohibitions.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Thursday 14 June 2007 11:20:34 Adrian Bunk wrote:
> On Thu, Jun 14, 2007 at 12:00:17AM -0400, Valdis.Kletnieks@vt.edu wrote:
> > On Thu, 14 Jun 2007 04:56:40 +0200, Adrian Bunk said:
> > > Reality check:
> > >
> > > Harald convinced companies that they have to provide the private keys
> > > required to run the Linux kernel they ship on their hardware.
> >
> > No, the *real* reality check:
> >
> > The operative words here are "convinced companies" - as opposed to
> > "convinced a judge to rule that private keys are required to be
> > disclosed". (I just checked around on gpl-violations.org, and I don't see
> > any news items that say they actually generated citable case law on the
> > topic of keys...)
> >
> > Harald convinced companies that it was easier/cheaper/faster to provide
> > the private keys than to continue in a long legal battle with an
> > uncertain outcome. If the company estimates the total loss due to keys
> > being released is US$100K, but the costs of taking it to court are
> > estimated at US$200K, it's obviously a win (lesser loss, actually) for
> > the company to just fold.
> >...
>
> Here in Germany, the rules at court are roughly "the loser pays
> everything including the costs of the winner", so if a big company is
> sure they will win at court there's no reason not to go there.
>
> And if they did the effort of using private keys to only allow running
> an official firmware, they must have seen an advantage from doing so.
>
> I'm not saying it legally clear the other way round, my statement was
> an answer to Daniel's emails claiming it was clear what such companies
> do was legal.

I'm sorry if I gave anyone that impression. My point was that it would be 
pointless to argue the case in the US because here it really is, 
usually , "buy the best justice for the money".

DRH

> cu
> Adrian



-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Rob Landley]

On Thursday 14 June 2007 15:49:13 Daniel Hazelton wrote:
> > I'm not saying it legally clear the other way round, my statement was
> > an answer to Daniel's emails claiming it was clear what such companies
> > do was legal.
>
> I'm sorry if I gave anyone that impression. My point was that it would be
> pointless to argue the case in the US because here it really is,
> usually , "buy the best justice for the money".

Or do what BusyBox and uClibc did (on the advice of Pamela Jones of Groklaw) 
and sign up with the the Software Freedom Law Center so they can enforce your 
copyrights for you.

Didn't cost us a dime, and they were ok with GPLv2 without the "or later" 
clause...

Rob
-- 
"One of my most productive days was throwing away 1000 lines of code."
  - Ken Thompson.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 13, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> Either a device is distributed, like the common PC, that is designed
> for the user to change and update the software on, or, like the PS2
> it isn't designed for that.

Have you ever installed GNU/Linux on a PC "Designed for Microsoft Windows"?

How dare you? ;-)

> If, OTOH, the hardware was never meant for the end-user to install custom 
> versions of the software on, then while the signing keys are still 
> *technically* part of the source, in practice they are not. Why? Because in 
> most of those cases the end-user isn't granted the right to install and run 
> custom binaries on the hardware.

And distributing the GPLed software under this restriction is quite
likely copyright infringement.

> I know this. As I said, I doubt that anyone who tried this in
> America would have the success he has had.

On Wednesday 13 June 2007 21:24:01 Adrian Bunk wrote:

>> Are you an idiot, or do you just choose to ignore all proof that
>> doesn't fit your preconceived beliefs?

;-) :-P :-D

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Krzysztof Halasa]

Daniel Hazelton <dhazelton@enter.net> writes:

> Nope. Merely stating a distinction. Either a device is distributed, like the 
> common PC, that is designed for the user to change and update the software 
> on, or, like the PS2 it isn't designed for that. If I find a way to update my 
> PS2 to run Linux and find that it doesn't want to start the "Linux Firmware" 
> because I'm lacking a signing key...
>
> In the case of a device that internally runs Linux (or any other GPL'd 
> software) and wasn't designed for the end-user to change the software running 
> on it then the signing keys aren't part of the source. OTOH, if I sell a PC 
> running Linux that requires the kernel be signed then the signing keys *are* 
> part of the source, since a PC is designed for the end-user to change the 
> software running on it.

Come on, GPL is software licence, the hardware isn't part of the
equation. One can argue that keys are or aren't part of the source
(= that digital signature is or isn't part of the executable) but
it's totally independent of any hardware and its purpose.

For example, it doesn't matter if the signature is merely for checking
file integrity (and any signature would do) or if it's for restricting
users from running something.
-- 
Krzysztof Halasa

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bernd Paysan]

[-- Attachment #1: Type: text/plain, Size: 762 bytes --]

On Thursday 14 June 2007 03:24, Adrian Bunk wrote:
> Harald is in Germany, and he therefore takes legal action against people
> distributing products violating his copyright on the Linux kernel
> in Germany at German courts based on German laws.

And if Tivo did sell their crap in Germany, I bet, Harald had brought them 
down years ago (as he did in the "tivoized" Siemens router case). But Tivo 
doesn't (they started in the UK, and stopped doing so right after Harald 
unlocked that Siemens router ;-), and in the US, courts may think 
different. Or they rely that there simply is no Harald Welte in the US, who 
goes after the violators.

-- 
Bernd Paysan
"If you want it done right, you have to do it yourself"
http://www.jwdt.com/~paysan/

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Rob Landley]

On Thursday 14 June 2007 05:32:47 Bernd Paysan wrote:
> On Thursday 14 June 2007 03:24, Adrian Bunk wrote:
> > Harald is in Germany, and he therefore takes legal action against people
> > distributing products violating his copyright on the Linux kernel
> > in Germany at German courts based on German laws.
>
> And if Tivo did sell their crap in Germany, I bet, Harald had brought them
> down years ago (as he did in the "tivoized" Siemens router case). But Tivo
> doesn't (they started in the UK, and stopped doing so right after Harald
> unlocked that Siemens router ;-), and in the US, courts may think
> different. Or they rely that there simply is no Harald Welte in the US, who
> goes after the violators.

On http://www.busybox.net the March 26, 2006 entry reads:

> 27 March 2006 -- Software Freedom Law Center representing BusyBox and
> uClibc One issue Erik Andersen wanted to resolve when handing off BusyBox
> maintainership to Rob Landley was license enforcement. BusyBox and uClibc's
> existing license enforcement efforts (pro-bono representation by Erik's
> father's law firm, and the Hall of Shame), haven't scaled to match the
> popularity of the projects. So we put our heads together and did the
> obvious thing: ask Pamela Jones of Groklaw for suggestions. She referred us
> to the fine folks at softwarefreedom.org.
>
> As a result, we're pleased to 
> announce that the Software Freedom Law Center has agreed to represent
> BusyBox and uClibc. We join a number of other free and open source software
> projects (such as X.org, Wine, and Plone in being represented by a fairly
> cool bunch of lawyers, which is not a phrase you get to use every day.

See also the September 29, 2006 entry where we set up an email address to 
forward license violation reports directly to them so we wouldn't have to 
deal with any of it.

I'd say this "hasn't cost me a dime", but I believe I'm on my third stamp.

(I also note that I'm not busybox maintainer anymore and Erik isn't uclibc 
maintainer anymore either, but since it's our copyrights they're basing the 
enforcement actions on, they still bounce an email off us every few months.)

Rob
-- 
"One of my most productive days was throwing away 1000 lines of code."
  - Ken Thompson.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alan Cox]

> I haven't looked into what Harald Welte did, but I'd be surprised if someone 
> tried following suit in America and had as much success.

One of the distinct advantages of German law over the US system (and to a
large extent the UK system its based upon) is that German law favours
justice over money.

> > AFAIK there haven't been any court rulings on this issue, and it could
> > even be that courts in different countries will decide differently.
> 
> Agreed.

That in theory shouldn't happen as the conventions on copyright are
supposed to stop that mess occuring.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Adrian Bunk]

On Thu, Jun 14, 2007 at 02:45:40AM +0100, Alan Cox wrote:
>...
> > > AFAIK there haven't been any court rulings on this issue, and it could
> > > even be that courts in different countries will decide differently.
> > 
> > Agreed.
> 
> That in theory shouldn't happen as the conventions on copyright are
> supposed to stop that mess occuring.

Is there any way how this would be resolved?

I can easily imagine that two courts, no matter whether they are in the 
same or different countries, would decide differently in grey areas like 
non-GPL modules or the GPLv2 and private keys.

If the two courts are in the same country there's usually a higher court 
above both that can resolve this. But what if let's say the highest 
court in the USA and the highest court in Germany would disagree on such 
a matter?

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 13, 2007, Adrian Bunk <bunk@stusta.de> wrote:

> If the two courts are in the same country there's usually a higher court 
> above both that can resolve this. But what if let's say the highest 
> court in the USA and the highest court in Germany would disagree on such 
> a matter?

Upgrade the license so as to provide guidance as to the intent of the
authors, such that the disagreement doesn't happen again.

If there's room in each country's laws to fix the problem, that is.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Adrian Bunk]

On Thu, Jun 14, 2007 at 03:07:17AM -0300, Alexandre Oliva wrote:
> On Jun 13, 2007, Adrian Bunk <bunk@stusta.de> wrote:
> 
> > If the two courts are in the same country there's usually a higher court 
> > above both that can resolve this. But what if let's say the highest 
> > court in the USA and the highest court in Germany would disagree on such 
> > a matter?
> 
> Upgrade the license so as to provide guidance as to the intent of the
> authors, such that the disagreement doesn't happen again.
> 
> If there's room in each country's laws to fix the problem, that is.

I don't think that's an option.

Consider the question of whether non-GPL kernel modules are legal at 
all and the number and different opinions of Linux kernel authors.

Plus the general question whether any "upgrade the license" would be 
valid in all jurisdictions ("GPL version 2 or any later version"  
as licence might make it possible, but in all other cases I'd have 
serious doubts).

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alan Cox]

> If the two courts are in the same country there's usually a higher court 
> above both that can resolve this. But what if let's say the highest 
> court in the USA and the highest court in Germany would disagree on such 
> a matter?

It would be very unusual for both cases to proceed in parallel. The one
court will then strongly consider the decision of the other.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Thu, 14 Jun 2007, Adrian Bunk wrote:
> 
> "For an executable work, complete source code means all the source code 
> for all modules it contains, plus any associated interface definition 
> files, plus the scripts used to control compilation and installation of 
> the executable."
> 
> The question is whether this includes private keys.

No. That's the question as the FSF would like to frame it.

But the real fact is that it *not* the right question.

You can install Linux on a Tivo all you like. Take out the harddisk, 
install your own version of Linux on it, and put it back in. That's pretty 
much how Tivo installs Linux on the things too, afaik, although they don't 
need to take the disk out (since they just assemble it).

No magic needed. In fact, no keys needed.

Now, maybe the hardware/firmware knows to expect a certain SHA1 on that 
disk, that's a different issue. Tivo could even tell you exactly what the 
SHA1 they are checking is. Maybe they have a list of SHA1's, and maybe 
they have a way to upgrade THEIR OWN FIRMWARE with new SHA1's, and they 
could still tell you all of them, and be very open.

And you could actually replace their copy of Linux with another one. It 
would have to have the same SHA1 to actually start _running_, but that's 
the hardware's choice. 

See? No private keys needed. No magic install scripts. It really _is_ that 
easy.

Of course, using private keys, and signing the image with them is possibly 
a technically more flexible/easier/more obvious way to do it, but in the 
end, do you really want to argue technical details?

But I think the whole thing is totally misguided, because the fact is, the 
GPLv2 doesn't talk about "in place" or "on the same hardware". 

So take another example: I obviously distribute code that is copyrighted 
by others under the GPLv2. Do I follow the GPLv2? I sure as hell do! But 
do I give you the same rights as I have to modify the copy on 
master.kernel.org as I have? I sure as hell DO NOT!

So by the idiotic logic of "modifying in place", I'm violating the GPLv2 
every time I'm makign a release - because I make Linux available, but I 
don't actually give people the "same rights" to that particular copy that 
I have! Oh horrors of horrors! You need to make a _copy_ of the thing I 
distribute, and then you have the same rights I have to that _copy_, but 
you never had the same rights to the thing I actually distributed!

And here's a big clue for people: anybody who thinks that I'm violating 
the GPLv2 by not giving out my private SSH key to master.kernel.org is a 
f*cking moron! You have the right to modify *copies* of the kernel I 
distribute, but you cannot actually modify the _actual_ entity that I made 
available!

See any parallels here? Any parallel to a CD-ROM distribution, or a Tivo 
distribution? The rights that the GPLv2 gives to "the software", is to 
something much bigger than "the particular copy of the software". 

Can people really not see the difference between "the software" and "a 
particular encoded copy of the software"? 

I'm sorry, but people who cannot see that difference are just stupid.

			Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:

> On Thu, 14 Jun 2007, Adrian Bunk wrote:
>> 
>> "For an executable work, complete source code means all the source code 
>> for all modules it contains, plus any associated interface definition 
>> files, plus the scripts used to control compilation and installation of 
>> the executable."
>> 
>> The question is whether this includes private keys.

> No. That's the question as the FSF would like to frame it.

No.  The FSF actually does *not* want to take this position.  That's
why it chose the formulation of Installation Instructions.  It doesn't
share my view that the keys needed to sign a binary in order for it to
work are part of the source code.

> And you could actually replace their copy of Linux with another one. It 
> would have to have the same SHA1 to actually start _running_, but that's 
> the hardware's choice. 

That's the hardware imposing a restriction on modification of the
software.  It doesn't matter how elaborate the excuse is to justify
denying users' freedoms: it's against the spirit of the GPL, and the
GPL will be amended as needed to plug such holes.

> So take another example: I obviously distribute code that is copyrighted 
> by others under the GPLv2. Do I follow the GPLv2? I sure as hell do! But 
> do I give you the same rights as I have to modify the copy on 
> master.kernel.org as I have? I sure as hell DO NOT!

That's an interesting argument.

People don't get your copy, so they're not entitled to anything about
it.

When they download the software, they get another copy, and they have
a right to modify that copy.

> And here's a big clue for people: anybody who thinks that I'm violating 
> the GPLv2 by not giving out my private SSH key to master.kernel.org is a 
> f*cking moron!

Agreed, except I'd probably use a lighter term.

> See any parallels here? Any parallel to a CD-ROM distribution, or a Tivo 
> distribution?

Yes.  You see how TiVO is different?  It is modifyable, and I actually
receive the copy that TiVO can still modify, but I can't.

> The rights that the GPLv2 gives to "the software", is to something
> much bigger than "the particular copy of the software".

Indeed, it's something bigger.  But this doesn't exclude the smaller
things, does it?

> Can people really not see the difference between "the software" and "a 
> particular encoded copy of the software"? 

There is a difference.  But the GPL doesn't limit itself to the
former.  It explicitly talks about "copies".

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Thursday 14 June 2007 02:36:12 Alexandre Oliva wrote:
> On Jun 14, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:
> > On Thu, 14 Jun 2007, Adrian Bunk wrote:
> >> "For an executable work, complete source code means all the source code
> >> for all modules it contains, plus any associated interface definition
> >> files, plus the scripts used to control compilation and installation of
> >> the executable."
> >>
> >> The question is whether this includes private keys.
> >
> > No. That's the question as the FSF would like to frame it.
>
> No.  The FSF actually does *not* want to take this position.  That's
> why it chose the formulation of Installation Instructions.  It doesn't
> share my view that the keys needed to sign a binary in order for it to
> work are part of the source code.
>
> > And you could actually replace their copy of Linux with another one. It
> > would have to have the same SHA1 to actually start _running_, but that's
> > the hardware's choice.
>
> That's the hardware imposing a restriction on modification of the
> software.  It doesn't matter how elaborate the excuse is to justify
> denying users' freedoms: it's against the spirit of the GPL, and the
> GPL will be amended as needed to plug such holes.

And? There is *absolutely* *nothing* in any version of the GPL *prior* to 3 
that says that hardware cannot impose restrictions. What the GPL *does* say 
is that you can't "add additional restrictions to the license" - (IMHO) a 
piece of hardware having a restriction isn't an "additional restriction added 
to the license". As well, as Linus stated, there is nothing *anywhere* - 
AFAICT, not even in GPLv3 - that says that you have to be able to run the 
software "in place" or "on the same hardware".

If a hardware manufacturer - like TiVO - uses GPL'd code in their product - 
and complies with the terms of the license - they aren't required to allow 
you to run modified code on that hardware. Without it mentioned anywhere in 
the GPL *OR* the assorted writings of RMS (who founded the FSF and wrote the 
original GPL) that "modified software must be able to run on the same 
hardware" then it cannot be in the "spirit" of the license to allow this.

> > So take another example: I obviously distribute code that is copyrighted
> > by others under the GPLv2. Do I follow the GPLv2? I sure as hell do! But
> > do I give you the same rights as I have to modify the copy on
> > master.kernel.org as I have? I sure as hell DO NOT!
>
> That's an interesting argument.
>
> People don't get your copy, so they're not entitled to anything about
> it.
>
> When they download the software, they get another copy, and they have
> a right to modify that copy.

But you get the TiVO corporations copy of the software? I smell a logical 
fallacy here, but can't remember the name for it.

> > And here's a big clue for people: anybody who thinks that I'm violating
> > the GPLv2 by not giving out my private SSH key to master.kernel.org is a
> > f*cking moron!
>
> Agreed, except I'd probably use a lighter term.
>
> > See any parallels here? Any parallel to a CD-ROM distribution, or a Tivo
> > distribution?
>
> Yes.  You see how TiVO is different?  It is modifyable, and I actually
> receive the copy that TiVO can still modify, but I can't.

I don't. You don't get the TiVO corporations copy of the software. You get 
your own copy, with all the rights that TiVO had when receiving the software. 
The right to install and run the kernel in the TiVO device is independent of 
the rights to copy, modify, distribute and run the software. (because the GPL 
never guarantees you the right to run the software on a particular piece of 
hardware.)

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> And? There is *absolutely* *nothing* in any version of the GPL *prior* to 3 
> that says that hardware cannot impose restrictions.

It's not that the hardware is deciding to impose restrictions on its
own.  It's the hardware distributor that is deciding to use the
hardware to impose restrictions on the user.  Seems like a violation
of section 6 of GPLv2 to me.

> What the GPL *does* say is that you can't "add additional
> restrictions to the license"

Not quite.  It's more general than that:

  You may not impose any further restrictions on the recipients'
  exercise of the rights granted herein.

>> > So take another example: I obviously distribute code that is copyrighted
>> > by others under the GPLv2. Do I follow the GPLv2? I sure as hell do! But
>> > do I give you the same rights as I have to modify the copy on
>> > master.kernel.org as I have? I sure as hell DO NOT!

>> That's an interesting argument.

>> People don't get your copy, so they're not entitled to anything about
>> it.

>> When they download the software, they get another copy, and they have
>> a right to modify that copy.

> But you get the TiVO corporations copy of the software?

Yes.  The customer gets the copy that TiVO stored in the hard disk in
the device it sells.  And it's that copy that the customer is entitled
to modify because TiVO is still able to modify it.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Thu, 14 Jun 2007, Alexandre Oliva wrote:
> 
> It's not that the hardware is deciding to impose restrictions on its
> own.  It's the hardware distributor that is deciding to use the
> hardware to impose restrictions on the user.  Seems like a violation
> of section 6 of GPLv2 to me.

You *still* haven't figured out the difference between "the software" and 
"a particular copy of the software", have you?

What's your problem?

I doubt you're really stupid, so I think your problem is that if you admit 
that "the software" is something *different* from "a particular copy of 
the software", you realize (perhaps subconsciously) that your arguments do 
not make any sense. So you do not allow yourself to think clearly about 
the matter.

So let's look at that "section 6" that you talk about, and quote the 
relevant parts, will  we:

	You may not impose any further restrictions on the recipients' 
	exercise of the rights granted herein.

and then let's look at Red Hat sending me a CD-ROM or a DVD.

Now, Red Hat clearly *did* "further restrict" my rights as it pertains TO 
THAT COPY ON THE CD-ROM! I cannot change it! Waa waa waa! I'll sue your 
sorry ass off!

See the issue? You are continually making the mistake of thinking that the 
GPLv2 talks about individual copies of software. And I'm continually 
having to point out that that is a MISTAKE.

And it's OBVIOUSLY a mistake, because your reading is nonsensical. If you 
think that Tivo does somethign bad, then hat Red Hat does is the same 
badness, thousads times over! I strongly suspect Red Hat has shipped a lot 
more CD-ROM's than Tivo has shipped boxes!

So let me iterate AGAIN:

 - the rights that the GPLv2 gives *cannot* be about "the particular copy" 
   that you send, since that would be INSANE. Red Hat sends lots of copies 
   of software that are NOT MODIFIABLE!

 - ergo, the rights about "the software" in the GPLv2 must be about 
   something else. 

See? Your argument about "individual copies" simply DOES NOT MAKE SENSE!

Just admit it.

> > What the GPL *does* say is that you can't "add additional
> > restrictions to the license"
> 
> Not quite.  It's more general than that:
> 
>   You may not impose any further restrictions on the recipients'
>   exercise of the rights granted herein.

And your point is? Nothing.

The rights granted are the rights to "distribute and modify the software". 
But by "the software", the license is not talking about a particular 
*copy* of the software, it's talking about the software IN THE ABSTRACT.

In other words, the reason that Red Hat is not violating the GPLv2 is that 
no, I cannot change the copy on the software on the particular CD-ROM or 
DVD, but I can get a copy of the sources other ways, and make my own 
modifications *SOMEWHERE*ELSE*!

The fact that Red Hat made a "restricted copy" is totally irrelevant.

In fact, it's exactly as irrelevant as the fact that Tivo makes a 
"restricted copy". The *software* is still free!

		Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:

> On Thu, 14 Jun 2007, Alexandre Oliva wrote:
>> 
>> It's not that the hardware is deciding to impose restrictions on its
>> own.  It's the hardware distributor that is deciding to use the
>> hardware to impose restrictions on the user.  Seems like a violation
>> of section 6 of GPLv2 to me.

> You *still* haven't figured out the difference between "the software" and 
> "a particular copy of the software", have you?

I have.  And so has GPLv2, look:

  2. You may modify your copy or copies of the Program or any portion
                         ^^^^

> What's your problem?

Trying to get you to see what is so obvious to me.

> So let's look at that "section 6" that you talk about, and quote the 
> relevant parts, will  we:

> 	You may not impose any further restrictions on the recipients' 
> 	exercise of the rights granted herein.

> and then let's look at Red Hat sending me a CD-ROM or a DVD.

> Now, Red Hat clearly *did* "further restrict" my rights as it pertains TO 
> THAT COPY ON THE CD-ROM! I cannot change it! Waa waa waa! I'll sue your 
> sorry ass off!

Red Hat is not stopping you from making changes.  The media is, and
that's not something Red Hat can control.

Compare this with the TiVO.  TiVO *designs* the thing such that it can
still make changes, but customers can't.

That's the difference.

TiVO is using hardware to "impose further restrictions on the
recipients' exercise of the rights granted herein", and this violates
section 6 of GPLv2.

> See the issue? You are continually making the mistake of thinking that the 
> GPLv2 talks about individual copies of software.

It does.  You're making the mistake of thinking that it doens't.  And
even in the legal terms that you claimed to have understood so
thoroughly.

> The rights granted are the rights to "distribute and modify the software". 

More specifically, some of the rights are:

  copy and distribute verbatim copies of the Program's source code as
  you receive it

  modify your copy or copies of the Program or any portion of it, thus
  forming a work based on the Program, and copy and distribute such
  modifications or work

> But by "the software", the license is not talking about a particular 
> *copy* of the software, it's talking about the software IN THE ABSTRACT.

Please read it again.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Thursday 14 June 2007 14:35:29 Alexandre Oliva wrote:
<snip>
> > So let's look at that "section 6" that you talk about, and quote the
> > relevant parts, will  we:
> >
> > 	You may not impose any further restrictions on the recipients'
> > 	exercise of the rights granted herein.
> >
> > and then let's look at Red Hat sending me a CD-ROM or a DVD.
> >
> > Now, Red Hat clearly *did* "further restrict" my rights as it pertains TO
> > THAT COPY ON THE CD-ROM! I cannot change it! Waa waa waa! I'll sue your
> > sorry ass off!
>
> Red Hat is not stopping you from making changes.  The media is, and
> that's not something Red Hat can control.

TiVO isn't stopping you from making changes - the *media* is. (in this case 
the "Media" isn't even doing as much as a CD-ROM does. The only thing a TiVO 
box restricts is which binaries it will execute as the operating system)

>
> Compare this with the TiVO.  TiVO *designs* the thing such that it can
> still make changes, but customers can't.
>
> That's the difference.

No, it isn't. Look at any motherboard. The Bios on the last three or four 
motherboards I've purchased check for a digital signature on the Bios 
updates. The motherboard manufacturer can make changes, but the customer 
can't. Is there any difference? Nope.

> TiVO is using hardware to "impose further restrictions on the
> recipients' exercise of the rights granted herein", and this violates
> section 6 of GPLv2.

No, they don't. The GPLv2 makes no provisions for you being able to execute a 
modified copy of the code on the same media or hardware that you received it 
on. The fact is that claiming it was "the spirit" doesn't matter at all - 
this isn't philosophy you're arguing, its *LAW*, and in law, if it isn't 
clearly spelled out, it doesn't exist.

> > See the issue? You are continually making the mistake of thinking that
> > the GPLv2 talks about individual copies of software.
>
> It does.  You're making the mistake of thinking that it doens't.  And
> even in the legal terms that you claimed to have understood so
> thoroughly.
>
> > The rights granted are the rights to "distribute and modify the
> > software".
>
> More specifically, some of the rights are:
>
>   copy and distribute verbatim copies of the Program's source code as
>   you receive it
>
>   modify your copy or copies of the Program or any portion of it, thus
>   forming a work based on the Program, and copy and distribute such
>   modifications or work

And where does it say that you even have the right to run the "work based on 
the Program", or even a self-compiled copy of the "verbatim copy of the code" 
on any given piece of hardware?

> > But by "the software", the license is not talking about a particular
> > *copy* of the software, it's talking about the software IN THE ABSTRACT.
>
> Please read it again.

Done. Section 3 of GPLv2 covers the right to distribute "object code" forms of 
a licensed work. At no point does it even *mention* that, if the object code 
form comes on a device capable of executing it, you have to give the right to 
execute a modified form of the work on the same platform. If this has been 
the "intent and spirit" of the license from the beginning, it should be there 
somewhere.

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> On Thursday 14 June 2007 14:35:29 Alexandre Oliva wrote:
> <snip>
>> > So let's look at that "section 6" that you talk about, and quote the
>> > relevant parts, will  we:
>> >
>> > 	You may not impose any further restrictions on the recipients'
>> > 	exercise of the rights granted herein.
>> >
>> > and then let's look at Red Hat sending me a CD-ROM or a DVD.
>> >
>> > Now, Red Hat clearly *did* "further restrict" my rights as it pertains TO
>> > THAT COPY ON THE CD-ROM! I cannot change it! Waa waa waa! I'll sue your
>> > sorry ass off!
>> 
>> Red Hat is not stopping you from making changes.  The media is, and
>> that's not something Red Hat can control.

> TiVO isn't stopping you from making changes - the *media* is.

TiVO made it so, that's the difference.

I'll give you that it's not so much about making changes per se, or
even installing them, as it is about running the modified versions for
any purpose.

>> Compare this with the TiVO.  TiVO *designs* the thing such that it can
>> still make changes, but customers can't.

>> That's the difference.

> No, it isn't. Look at any motherboard. The Bios on the last three or four 
> motherboards I've purchased check for a digital signature on the Bios 
> updates. The motherboard manufacturer can make changes, but the customer 
> can't. Is there any difference? Nope.

Is the BIOS code under the GPL?

> The fact is that claiming it was "the spirit" doesn't matter at all
> - this isn't philosophy you're arguing, its *LAW*, and in law, if it
> isn't clearly spelled out, it doesn't exist.

That's exactly what makes for the difference between the spirit and
the precise legal terms, and why GPLv3 is fixing these divergences.

> And where does it say that you even have the right to run the "work based on 
> the Program", or even a self-compiled copy of the "verbatim copy of the code" 
> on any given piece of hardware?

It doesn't.  The license can't demand the software, or modified
versions thereof, to run.  The only thing it can demand is that
licensees don't impose restrictions on others' abilities to do so.

>> > But by "the software", the license is not talking about a particular
>> > *copy* of the software, it's talking about the software IN THE ABSTRACT.
>> 
>> Please read it again.

> Done.

  2. You may modify your copy or copies of the Program or any portion
  of it                  ^^^^

> If this has been the "intent and spirit" of the license from the
> beginning, it should be there somewhere.

I think you're missing what 'spirit' means.  It's guidance, it's not
the legal terms.  And it's precisely because the implementation (the
legal terms) failed to meet that design (the spirit, encoded in the
preamble) that the license needs patching.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Thursday 14 June 2007 21:43:07 Alexandre Oliva wrote:
> On Jun 14, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> > On Thursday 14 June 2007 14:35:29 Alexandre Oliva wrote:
> > <snip>
> >
> >> > So let's look at that "section 6" that you talk about, and quote the
> >> > relevant parts, will  we:
> >> >
> >> > 	You may not impose any further restrictions on the recipients'
> >> > 	exercise of the rights granted herein.
> >> >
> >> > and then let's look at Red Hat sending me a CD-ROM or a DVD.
> >> >
> >> > Now, Red Hat clearly *did* "further restrict" my rights as it pertains
> >> > TO THAT COPY ON THE CD-ROM! I cannot change it! Waa waa waa! I'll sue
> >> > your sorry ass off!
> >>
> >> Red Hat is not stopping you from making changes.  The media is, and
> >> that's not something Red Hat can control.
> >
> > TiVO isn't stopping you from making changes - the *media* is.
>
> TiVO made it so, that's the difference.
>
> I'll give you that it's not so much about making changes per se, or
> even installing them, as it is about running the modified versions for
> any purpose.

Artificial distinction on your part.

> >> Compare this with the TiVO.  TiVO *designs* the thing such that it can
> >> still make changes, but customers can't.
> >>
> >> That's the difference.
> >
> > No, it isn't. Look at any motherboard. The Bios on the last three or four
> > motherboards I've purchased check for a digital signature on the Bios
> > updates. The motherboard manufacturer can make changes, but the customer
> > can't. Is there any difference? Nope.
>
> Is the BIOS code under the GPL?

By your reasoning it doesn't even matter. I own the hardware, I should be able 
to change the BIOS to *any* chunk of code I want.

Do you see the fallacy here? You're making an artificial distinction based on 
whether the *SOFTWARE* has a certain license or not.

> > The fact is that claiming it was "the spirit" doesn't matter at all
> > - this isn't philosophy you're arguing, its *LAW*, and in law, if it
> > isn't clearly spelled out, it doesn't exist.
>
> That's exactly what makes for the difference between the spirit and
> the precise legal terms, and why GPLv3 is fixing these divergences.

And the reason behind this is all "ethics and morals". In other words, you are 
forcing those "ethics and morals" on others and hiding it by giving it a 
different name.

Wasn't it Shakespear who said: "What is in a name? A Rose by any other name 
would smell as sweet"

> > And where does it say that you even have the right to run the "work based
> > on the Program", or even a self-compiled copy of the "verbatim copy of
> > the code" on any given piece of hardware?
>
> It doesn't.  The license can't demand the software, or modified
> versions thereof, to run.  The only thing it can demand is that
> licensees don't impose restrictions on others' abilities to do so.

No, it doesn't. There is no requirement in the license in question that makes 
a persons ability to run the program on any given piece of hardware. What it 
does say is you can't stop someone from *TRYING* to do that.

> >> > But by "the software", the license is not talking about a particular
> >> > *copy* of the software, it's talking about the software IN THE
> >> > ABSTRACT.
> >>
> >> Please read it again.
> >
> > Done.
>
>   2. You may modify your copy or copies of the Program or any portion
>   of it                  ^^^^
>
> > If this has been the "intent and spirit" of the license from the
> > beginning, it should be there somewhere.
>
> I think you're missing what 'spirit' means.  It's guidance, it's not
> the legal terms.  And it's precisely because the implementation (the
> legal terms) failed to meet that design (the spirit, encoded in the
> preamble) that the license needs patching.

If the intent of a law (or license) is to do A but it doesn't say that, then 
how is the intent to be known? Your answer: Ask the author. Question: how can 
we be absolutely certain that the authors intent *hasn't* changed since the 
law (or license) was written? *ONLY* answer: It is impossible. Conclusion: 
Unless the intent is clearly spelled out at the time the law (or license) is 
written, or is available in other writings by the author of the law/license 
from the same time period as the law/license then it is impossible.

Question: How do you know what the "spirit" of a license is?
Your Answer: Ask the author.
Question: How do we know that the Author hasn't changed their mind about 
the "spirit" of the license since it was written?
*ONLY* Answer: See the answer to the parallel question about "intent".

Now that I've knocked down your "Intent" and "Spirit" straw-men you have no 
way to argue that the GPLv3 is written with the same "spirit" as the GPLv2. 

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> You're making an artificial distinction based on whether the
> *SOFTWARE* has a certain license or not.

What matters to me is that, when the GPL says you can't impose further
restrictions, then you can't, no matter how convoluted your argument
is

>> That's exactly what makes for the difference between the spirit and
>> the precise legal terms, and why GPLv3 is fixing these divergences.

> And the reason behind this is all "ethics and morals".

There was never any attempt to hide that this was what the Free
Software movement was about, and that the GPL was about defending
these freedoms.

Sure, it has other advantages.  But the goal has always been the same,
and it's not going to change.

> If the intent of a law (or license) is to do A but it doesn't say
> that, then how is the intent to be known?  Your answer: Ask the
> author.

No, you interpret based on what the author wrote then.

You read the preamble, and any other rationales associated with the
license or law.  I don't know how it's elsewhere, but in Brazil every
law has a rationale, and that's often used to guide its interpretation
in courts, even though the rationale is not part of the law.


If the author realizes what he wrote was not enough, or it got
misinterpreted, author his text, and then whoever feels like it and is
entitled to adopts the revised version.


In the GPLv2=>v3 case, all that needed revision was the legalese.  The
preamble has barely changed.  This is a strong indication that the
spirit remains the same, is it not?

> Unless the intent is clearly spelled out at the time the law (or
> license) is written, or is available in other writings by the author
> of the law/license from the same time period as the law/license then
> it is impossible.

Is there anything not clear about freedom #0, in the free software
definition, alluded to by the preamble that talks about free software
in very similar terms?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Thursday 14 June 2007 23:39:50 Alexandre Oliva wrote:
> On Jun 14, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> > You're making an artificial distinction based on whether the
> > *SOFTWARE* has a certain license or not.
>
> What matters to me is that, when the GPL says you can't impose further
> restrictions, then you can't, no matter how convoluted your argument
> is

Convoluted? Not in the least. Every example I have given has been an example 
of the application of your logic. If my examples are convoluted, then, QED, 
so is your logic.

> >> That's exactly what makes for the difference between the spirit and
> >> the precise legal terms, and why GPLv3 is fixing these divergences.
> >
> > And the reason behind this is all "ethics and morals".
>
> There was never any attempt to hide that this was what the Free
> Software movement was about, and that the GPL was about defending
> these freedoms.
>
> Sure, it has other advantages.  But the goal has always been the same,
> and it's not going to change.

I'm not trying to change that. My point in making that statement is to prove 
that the FSF is doing exactly what the Spanish Inquisition did, what 
every "Communist Revolution" has done and what Hitler did. Saying "My ethics 
and morals are better than everyone else, so I'm going to force everyone else 
to have my morals and ethics". That the FSF isn't doing this through force of 
arms or threat of violence just shows how sophisticated people have really 
become in the sixty years that have passed since Hitler - they now use threat 
of legal action.

> > If the intent of a law (or license) is to do A but it doesn't say
> > that, then how is the intent to be known?  Your answer: Ask the
> > author.
>
> No, you interpret based on what the author wrote then.

Really? Well I must say I'm surprised at the sudden change of heart. I have 
several mails here in which you have either said "You ask the author" or that 
line has been quoted.

> You read the preamble, and any other rationales associated with the
> license or law.  I don't know how it's elsewhere, but in Brazil every
> law has a rationale, and that's often used to guide its interpretation
> in courts, even though the rationale is not part of the law.

Show me where in the preamble that this issue of "it must run on any given 
piece of hardware" or even less generally, "it must run on the hardware it 
came on" is even *hinted* at. You wont find it. Nor will you find any mention 
of anything of the sort in the publicly available writings of RMS.

But let me go re-read the GPLv2 preamble again and see if it even hints at 
this issue... oh, wait, I read it earlier and didn't see anything that hinted 
at this. So I can safely conclude that no lawyer or judge would find it when 
interpreting the license. QED: The Tivo clause of GPLv3 causes it to break 
spirit with the GPLv2.

(And, by the way, if the FSF decided to release a GPLv4 that had an active 
section that said "You must turn over all copyright rights to a work released 
under this license to the FSF" it wouldn't "break spirit" with the GPL (v2 or 
v3). Why? Because *both* contain the following paragraph:

"We protect your rights with two steps: (1) copyright the software, and
(2) offer you this license which gives you legal permission to copy,
distribute and/or modify the software."

By your logic it is the *intent* of the FSF to hold copyright on all software 
released under the GPL *and* only give the rights detailed in the license to 
other people - including the person who has placed the work under the GPL.

Can you see the problem with your logic ?

>
> If the author realizes what he wrote was not enough, or it got
> misinterpreted, author his text, and then whoever feels like it and is
> entitled to adopts the revised version.
>
>
> In the GPLv2=>v3 case, all that needed revision was the legalese.  The
> preamble has barely changed.  This is a strong indication that the
> spirit remains the same, is it not?

If "tivoization" was against the spirit, then all that would have been needed 
was one extra clause clearly explaining that. Instead there are more than 6 
extra sections in the GPLv3.

If "DRM" was against the license then an extra section clearly explaining that 
could have been added. (in the DRM case I  actually understand the reasoning 
and agree with it.)

> > Unless the intent is clearly spelled out at the time the law (or
> > license) is written, or is available in other writings by the author
> > of the law/license from the same time period as the law/license then
> > it is impossible.
>
> Is there anything not clear about freedom #0, in the free software
> definition, alluded to by the preamble that talks about free software
> in very similar terms?

0. ... Activities other than copying, distribution and modification are not
covered by this License; they are outside its scope.  The act of
running the Program is not restricted, and the output from the Program
is covered only if its contents constitute a work based on the
Program (independent of having been made by running the Program).
Whether that is true depends on what the Program does.

Note that it says, in very clear and simple English, that the scope of the 
license is *ONLY* the activities of "copying, distribution and modification". 
I must be an idiot, because I completely fail to see how an activity besides 
one of the three that are mentioned as being the only ones in the scope of 
the license is in the "spirit" of the license.

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 15, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> On Thursday 14 June 2007 23:39:50 Alexandre Oliva wrote:
>> On Jun 14, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
>> > You're making an artificial distinction based on whether the
>> > *SOFTWARE* has a certain license or not.
>> 
>> What matters to me is that, when the GPL says you can't impose further
>> restrictions, then you can't, no matter how convoluted your argument
>> is

> Convoluted? Not in the least.

I didn't say your arguments were convoluted, and I know I didn't mean
to say that.  But I've heard enough arguments about excuses to escape
the obligations of the GPL (and other licenses and obligations, FWIW)
to know that such arguments can get very convoluted.

That said, I was actually trying to quote Eben Moglen, who once spoke
about this, but the word he used was "elaborate", not "convoluted".
Unfortunately, the right word escaped me ATM.

>> > If the intent of a law (or license) is to do A but it doesn't say
>> > that, then how is the intent to be known?  Your answer: Ask the
>> > author.
>> 
>> No, you interpret based on what the author wrote then.

> Really? Well I must say I'm surprised at the sudden change of heart. I have 
> several mails here in which you have either said "You ask the author" or that 
> line has been quoted.

It's no change.  You interpret what's there.  If it's clear, good.  If
there's a dispute, you have to ask the author, only s/he knows what
s/he meant.  It's really that simple.

> Show me where in the preamble that this issue of "it must run on any given 
> piece of hardware"

Why is the burden of the proof on me?

You show me where it says "one may impose restrictions on what
particupar pieces of hardware the program can run", to override the
general spirit of "passing on all the rights one has".

> (And, by the way, if the FSF decided to release a GPLv4 that had an active 
> section that said "You must turn over all copyright rights to a work released 
> under this license to the FSF" it wouldn't "break spirit" with the GPL (v2 or 
> v3).

Can't.  These terms wouldn't apply to the copyright holder (the only
person who could make the transfer), only to licensees.

> If "tivoization" was against the spirit, then all that would have been needed 
> was one extra clause clearly explaining that. Instead there are more than 6 
> extra sections in the GPLv3.

Erhm...  How did you get the (completely flawed, BTW) impression that
tivoization was all GPLv3 was about?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Friday 15 June 2007 16:04:15 Alexandre Oliva wrote:
> On Jun 15, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> > On Thursday 14 June 2007 23:39:50 Alexandre Oliva wrote:
> >> On Jun 14, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> >> > You're making an artificial distinction based on whether the
> >> > *SOFTWARE* has a certain license or not.
> >>
> >> What matters to me is that, when the GPL says you can't impose further
> >> restrictions, then you can't, no matter how convoluted your argument
> >> is
> >
> > Convoluted? Not in the least.
>
> I didn't say your arguments were convoluted, and I know I didn't mean
> to say that.  But I've heard enough arguments about excuses to escape
> the obligations of the GPL (and other licenses and obligations, FWIW)
> to know that such arguments can get very convoluted.
>
> That said, I was actually trying to quote Eben Moglen, who once spoke
> about this, but the word he used was "elaborate", not "convoluted".
> Unfortunately, the right word escaped me ATM.

Thanks for the clarification.

> >> > If the intent of a law (or license) is to do A but it doesn't say
> >> > that, then how is the intent to be known?  Your answer: Ask the
> >> > author.
> >>
> >> No, you interpret based on what the author wrote then.
> >
> > Really? Well I must say I'm surprised at the sudden change of heart. I
> > have several mails here in which you have either said "You ask the
> > author" or that line has been quoted.
>
> It's no change.  You interpret what's there.  If it's clear, good.  If
> there's a dispute, you have to ask the author, only s/he knows what
> s/he meant.  It's really that simple.

And as I have hopefully given good proof for, asking the author is not a good 
solution. The author can change their mind about their intent at any point in 
time - even during the process of writing the license.

> > Show me where in the preamble that this issue of "it must run on any
> > given piece of hardware"
>
> Why is the burden of the proof on me?
>
> You show me where it says "one may impose restrictions on what
> particupar pieces of hardware the program can run", to override the
> general spirit of "passing on all the rights one has".

It's "pass on all rights granted under this license". If I had to pass on "all 
rights I have" I'd have to pass on my right to change the license on my code. 
Since that isn't a right I'm obligated to pass on - and you could never 
convice me it is - I'm not "passing on all the rights I have" at *all*.

> > (And, by the way, if the FSF decided to release a GPLv4 that had an
> > active section that said "You must turn over all copyright rights to a
> > work released under this license to the FSF" it wouldn't "break spirit"
> > with the GPL (v2 or v3).
>
> Can't.  These terms wouldn't apply to the copyright holder (the only
> person who could make the transfer), only to licensees.

It's part of the preamble, in which the "We" refers to the FSF. If the 
preamble determines the "intent" and "spirit" of the license, then part of 
the "intent" and "spirit" of the license is collective aggregation of all 
copyright rights to all software released under the GPL by the FSF.

> > If "tivoization" was against the spirit, then all that would have been
> > needed was one extra clause clearly explaining that. Instead there are
> > more than 6 extra sections in the GPLv3.
>
> Erhm...  How did you get the (completely flawed, BTW) impression that
> tivoization was all GPLv3 was about?

I've looked through the GPLv3 and "tivoization" and DRM are the only things 
that are functionally different. In reading the GPLv3 *again* today I got the 
impression that there are more restrictions than grants of rights.

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 15, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> I've looked through the GPLv3 and "tivoization" and DRM are the only things 
> that are functionally different. In reading the GPLv3 *again* today I got the 
> impression that there are more restrictions than grants of rights.

http://fsfla.org/svnwiki/blogs/lxo/draft/gplv3-snowwhite discusses
each one of the significant changes (and some of the insignificant
ones) and shows why each one of them is more "tit-for-tat" than v2.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Florin Malita]

Alexandre Oliva wrote:
> On Jun 14, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
>   
>>> When they download the software, they get another copy, and they have
>>> a right to modify that copy.
>>>       
>> But you get the TiVO corporations copy of the software?
>>     
>
> Yes.  The customer gets the copy that TiVO stored in the hard disk in
> the device it sells.  And it's that copy that the customer is entitled
> to modify because TiVO is still able to modify it.
>   

No, by this twisted logic Tivo *cannot* modify that particular copy any 
more than you can. They can modify *another* copy (just like you) and 
they can *replace* the copy in your device with the new version (unlike 
you).

So your entire logical construct does not stand because this is not 
(cannot be) about modifying a particular copy (how would you do that 
anyway? hexedit the binary blob in place?) but about the ability to 
deploy the software on a particular platform.

---
fm

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Florin Malita <fmalita@gmail.com> wrote:

> Alexandre Oliva wrote:
>> On Jun 14, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
>> 
>>>> When they download the software, they get another copy, and they have
>>>> a right to modify that copy.

>>> But you get the TiVO corporations copy of the software?

>> Yes.  The customer gets the copy that TiVO stored in the hard disk in
>> the device it sells.  And it's that copy that the customer is entitled
>> to modify because TiVO is still able to modify it.

> No, by this twisted logic Tivo *cannot* modify that particular copy
> any more than you can. They can modify *another* copy (just like you)
> and they can *replace* the copy in your device with the new version
> (unlike you).

Again, replacing is one form of modification.

What do you think you do when you save a modified source file in your
editor?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Florin Malita]

On 06/14/2007 02:27 PM, Alexandre Oliva wrote:
>> No, by this twisted logic Tivo *cannot* modify that particular copy
>> any more than you can. They can modify *another* copy (just like you)
>> and they can *replace* the copy in your device with the new version
>> (unlike you).
>>     
>
> Again, replacing is one form of modification.
>   

No, it's not: replacing does not create derivative work. Modification does.

You've chosen to attach a physical dimension to "program copy" and I'm 
arguing that even under this distorted line of reasoning you can't 
support your position:

> The customer gets the copy that TiVO stored in the hard disk in
> the device it sells.  And it's that copy that the customer is entitled
> to modify because TiVO is still able to modify it.

* Tivo takes public sources, modifies them and builds a brand new blob
* Tivo installs this new copy on the device, most likely side-by-side 
with the old one - notice how the new copy is derived from public 
sources and has absolutely nothing to do with the old version (heck, it 
can be a totally different kernel for what it's worth)
* Tivo deletes the old copy from the device

It seems pretty obvious that the only right Tivo is withholding is the 
right to install new versions on the device - they never do (and really 
never could) "modify" the physical copy on your device (which is your 
main argument).


> What do you think you do when you save a modified source file in your
> editor?

Don't skip the part where the in-memory version started as an exact copy 
of the original being replaced. Notice the difference? ;)

---
fm

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Florin Malita <fmalita@gmail.com> wrote:

> On 06/14/2007 02:27 PM, Alexandre Oliva wrote:
>>> No, by this twisted logic Tivo *cannot* modify that particular copy
>>> any more than you can. They can modify *another* copy (just like you)
>>> and they can *replace* the copy in your device with the new version
>>> (unlike you).

>> Again, replacing is one form of modification.

> No, it's not: replacing does not create derivative
> work. Modification does.

Thanks.  Good point.  This convinces me that this doesn't work as a
legal argument under copyright.

I still stand by my understanding that this restriction violates the
spirit of the license.

And since the specific implementation involves creating a derived work
of the GPLed kernel (the signature, or the signed image, or what have
you) and refraining from providing the corresponding sources to that
derived work (the key and the signature "build scripts"), I still
think this specific case is a violation of the letter of the GPLv2,
even if the FSF doesn't take this position.

> It seems pretty obvious that the only right Tivo is withholding is the
> right to install new versions on the device

Actually, no.  They withhold the right to run versions that they don't
authorize themselves.

Back when GPLv2 was written, the right to run was never considered an
issue.  It was taken for granted, because copyright didn't control
that in the US (it does in Brazil), and nobody had thought of
technical measures to stop people from running modified copies of
software.  At least nobody involved in GPLv2, AFAIK.

The landscape has changed, and GPLv3 is meant to defend this freedom
that was taken for granted.

> they never do (and really never could) "modify" the physical copy on
> your device (which is your main argument).

Qualifying it as the main argument is a bit of an exaggeration.  I
have a number of different arguments.  The one about incomplete
sources is the most solid IMHO.

>> What do you think you do when you save a modified source file in your
>> editor?

> Don't skip the part where the in-memory version started as an exact
> copy of the original being replaced. Notice the difference? ;)

Sorry, I really don't follow.  Both versions of the kernel binary also
started from a common source ancestor.  Were you trying to make a
distinction on these grounds?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bill Nottingham]

Alexandre Oliva (aoliva@redhat.com) said: 
> And since the specific implementation involves creating a derived work
> of the GPLed kernel (the signature, or the signed image, or what have
> you)

Wait, a signed filesystem image that happens to contain GPL code
is now a derived work? Under what sort of interpretation does *that*
occur?

(This pretty much throws the 'aggregation' premise in GPLv2 completely
out.)

Bill

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alan Cox]

On Thu, 14 Jun 2007 19:45:08 -0400
Bill Nottingham <notting@redhat.com> wrote:

> Alexandre Oliva (aoliva@redhat.com) said: 
> > And since the specific implementation involves creating a derived work
> > of the GPLed kernel (the signature, or the signed image, or what have
> > you)
> 
> Wait, a signed filesystem image that happens to contain GPL code
> is now a derived work? Under what sort of interpretation does *that*
> occur?
> 
> (This pretty much throws the 'aggregation' premise in GPLv2 completely
> out.)

Perhaps the FSF will in future remember to pack a copy of the GPL in each
of its md5sum files on the mirror if this is a derivative work, and
modify the bittorrent protocol to include a copy of the GPL in the seed
files 8)

Alan

RE: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Schwartz]

> Perhaps the FSF will in future remember to pack a copy of the GPL in each
> of its md5sum files on the mirror if this is a derivative work, and
> modify the bittorrent protocol to include a copy of the GPL in the seed
> files 8)
>
> Alan

I realize you're joking, but for the benefit of anyone who might not
understand how this works:

A derivative work must, first and foremost, be a work. If it's not a work,
it's not a derivative work because a derivative work is a type of work.
Aggregations of multiple works, when that aggregation is performed in an
automated way, are not works. Even if I compile and link a bunch of source
code, provided there is no creative input in the compile and link process,
the result is not a work for copyright purposes. It is simply an aggregate
of the bits of source code. The gist of a compilation or derivative work is
the creative selection and modification process.

If someone argues that a program is a derivative work of a header file it
was compiled with, he is probably just being sloppy. The resulting
executable contains the header file combines with other works. Of course, a
source code file that is designed based on the contents of a header file may
be a derivative work of that header file, but that would be because the
human who wrote the source code file used bits of the header file in the
source code itself. It would not be because the compiler combined them --
such an automated combination has no creative input and so cannot produce a
work, and so cannot produce a derivative work.

This is grossly oversimplified, but should give you the idea.

DS

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Bill Nottingham <notting@redhat.com> wrote:

> Alexandre Oliva (aoliva@redhat.com) said: 
>> And since the specific implementation involves creating a derived work
>> of the GPLed kernel (the signature, or the signed image, or what have
>> you)

> Wait, a signed filesystem image that happens to contain GPL code
> is now a derived work? Under what sort of interpretation does *that*
> occur?

Is the signature not derived from the bits in the GPLed component, as
much as it is derived from the key?

Isn't the signature is a functional portion of the image, i.e., if I
take it out from the system, it won't work any more?

> (This pretty much throws the 'aggregation' premise in GPLv2 completely
> out.)

Not really.  It could take some explicit distinguishing between
functional and non-functional signatures, but that's about it.

GPLv3 chose a different path to make this clarification.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bill Nottingham]

Alexandre Oliva (aoliva@redhat.com) said: 
> > Wait, a signed filesystem image that happens to contain GPL code
> > is now a derived work? Under what sort of interpretation does *that*
> > occur?
> 
> Is the signature not derived from the bits in the GPLed component, as
> much as it is derived from the key?
> 
> Isn't the signature is a functional portion of the image, i.e., if I
> take it out from the system, it won't work any more?
> 
> > (This pretty much throws the 'aggregation' premise in GPLv2 completely
> > out.)
> 
> Not really.  It could take some explicit distinguishing between
> functional and non-functional signatures, but that's about it.

OK. Let's take this to the simple and logical conclusion. A signed
filesystem image containing both GPL and non-GPL code. From your
point A, this is a derived work. 

Let's read the license...

2. b) You must cause any work that ... is derived from the Program or any
   part thereof, to be licensed as a whole at no charge to all third
   parties under the terms of this License.

...
 But when you
 distribute the same sections as part of a whole which is a work based
 on the Program, the distribution of the whole must be on the terms of
 this License, whose permissions for other licensees extend to the
 entire whole, and thus to each and every part regardless of who wrote it.
 
and yet later:

In addition, mere aggregation of another work not based on the Program
with the Program (or with a work based on the Program) on a volume of
a storage or distribution medium does not bring the other work under
the scope of this License.

Pick one. They can't both be valid.

Moreover, this interpretation means that Red Hat (and pretty much
any other Linux distributor) should close up shop, as that's what
we've been doing for years.

Bill

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Bill Nottingham <notting@redhat.com> wrote:

> OK. Let's take this to the simple and logical conclusion. A signed
> filesystem image containing both GPL and non-GPL code. From your
> point A, this is a derived work. 

I claim the signature is derived from the GPLed bits, yes.  Whether
that's a derived work, in the legal sense, I'm not qualified to say.

And I claim that, in the case of TiVO, it is not only a functional
piece of the system that's derived from GPLed code and missing the
corresponding sources, but also it's being used to impose restrictions
on the exercise of the freedoms that the GPL is designed to protect.
And these conditions are what make it a bad thing, and that deviate,
if not from the legal conditions, at least from the spirit of the
license.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[david]

On Fri, 15 Jun 2007, Alexandre Oliva wrote:

> Subject: Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3
> 
> On Jun 14, 2007, Bill Nottingham <notting@redhat.com> wrote:
>
>> OK. Let's take this to the simple and logical conclusion. A signed
>> filesystem image containing both GPL and non-GPL code. From your
>> point A, this is a derived work.
>
> I claim the signature is derived from the GPLed bits, yes.  Whether
> that's a derived work, in the legal sense, I'm not qualified to say.

it's also derived from the non-GPLed bits as well.

so if it were a derived work in a legal sense (nessasary for your 
argument to have any legal meaning) then it's now illegal to make and 
distribute a checksum of a CD that contains software with incompatible 
licenses.

> And I claim that, in the case of TiVO, it is not only a functional
> piece of the system that's derived from GPLed code and missing the
> corresponding sources, but also it's being used to impose restrictions
> on the exercise of the freedoms that the GPL is designed to protect.
> And these conditions are what make it a bad thing, and that deviate,
> if not from the legal conditions, at least from the spirit of the
> license.

you keep claiming this, but other people claim you are wrong. what good do 
your claims do (why are your claims about what's in the spirit of the 
license and what's not any more valid than anyone else's?)

David Lang

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Michael Poole]

david@lang.hm writes:

> On Fri, 15 Jun 2007, Alexandre Oliva wrote:
>
>> Subject: Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3
>>
>> On Jun 14, 2007, Bill Nottingham <notting@redhat.com> wrote:
>>
>>> OK. Let's take this to the simple and logical conclusion. A signed
>>> filesystem image containing both GPL and non-GPL code. From your
>>> point A, this is a derived work.
>>
>> I claim the signature is derived from the GPLed bits, yes.  Whether
>> that's a derived work, in the legal sense, I'm not qualified to say.
>
> it's also derived from the non-GPLed bits as well.
>
> so if it were a derived work in a legal sense (nessasary for your
> argument to have any legal meaning) then it's now illegal to make and
> distribute a checksum of a CD that contains software with incompatible
> licenses.

It is not necessary for the end item to be a derived work in order for
the GPL to apply.  A literal copy is not a derived work; a translation
is not a derived work; an executable version of a program is not a
derived work of its source code; and so forth.

What is necessary is that the "work based on the [GPLed] Program" be
more than a mere aggregation of the GPLed component(s) with non-GPLed
components.  The fact that part of the work-as-a-whole is a descriptor
of the GPLed part does not mean all descriptions the GPLed part are
governed by the GPL.  The critical factor is that the GPLed part will
not function properly without the DRM signature.

Michael Poole

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Rob Landley]

On Thursday 14 June 2007 22:25:57 Alexandre Oliva wrote:
> On Jun 14, 2007, Bill Nottingham <notting@redhat.com> wrote:
> > Alexandre Oliva (aoliva@redhat.com) said:
> >> And since the specific implementation involves creating a derived work
> >> of the GPLed kernel (the signature, or the signed image, or what have
> >> you)
> >
> > Wait, a signed filesystem image that happens to contain GPL code
> > is now a derived work? Under what sort of interpretation does *that*
> > occur?
>
> Is the signature not derived from the bits in the GPLed component, as
> much as it is derived from the key?

Actually, you can't copyright, trademark, or patent a number.  In order to 
copyright something it has to have some creative element.  You also can't 
copyright (or trademark) book titles.  So no, last I checked you can't 
copyright an MD5sum or SHA1sum.

I vaguely recall somebody dredging around for the smallest thing there was a 
legal precedent explicitly affirming you could copyright it, and it was a 
haiku.  So they put an uncompressed ascii haiku in their protocol...

Now if you sign the executable binary, then the binary (as a whole) is a 
derivative work of your copyrighted code etc. ad nauseum pluribus unum and so 
on.  And THAT is due to Apple vs Franklin in 1983:
  http://www.internetlegal.com/impactof.htm
Before which copyright was only guaranteed to apply to source code, not to 
binaries (which are basically big numbers).  That's why everybody distributed 
source code before then: it was the only thing they knew you could enforce a 
copyright on...

IBM's "Object code only" initiative happened around the same time...
  http://landley.net/history/mirror/ibm/oco.html
Along with the AT&T breakup commercializing Unix, the launch of the GNU 
project, and the general rise of "shrinkwrap" software.

(There's this marvelous book called "Legal battles that shaped the computer 
industry" by Lawrence D. Graham, devotes a few pages to Apple vs Franklin.  
Franklin honestly didn't think Apple's binary ROMs were copyrightable.  Just 
as in a 1980 interview with Bill Gates, he couldn't stop somebody from 
printing a book with an annotated printout of the TRS-80 ROMs Microsoft had a 
copyright to.  He sounded kind of pissed about it, actually.  Also young and 
whiny:
Transcript:
  http://slashdot.org/features/00/01/20/1316236.shtml
Audio:
  http://landley.net/history/mirror/ms/gates.mp3
)

Rob
-- 
"One of my most productive days was throwing away 1000 lines of code."
  - Ken Thompson.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 15, 2007, Rob Landley <rob@landley.net> wrote:

> On Thursday 14 June 2007 22:25:57 Alexandre Oliva wrote:
>> On Jun 14, 2007, Bill Nottingham <notting@redhat.com> wrote:
>> > Alexandre Oliva (aoliva@redhat.com) said:
>> >> And since the specific implementation involves creating a derived work
>> >> of the GPLed kernel (the signature, or the signed image, or what have
>> >> you)
>> >
>> > Wait, a signed filesystem image that happens to contain GPL code
>> > is now a derived work? Under what sort of interpretation does *that*
>> > occur?
>> 
>> Is the signature not derived from the bits in the GPLed component, as
>> much as it is derived from the key?

> Actually, you can't copyright, trademark, or patent a number.

Agreed.  And this counter-argument of yours is a distraction.

I was careful to not talk about "derived work".  Please read it again
under this clarification (that I'm pretty sure I'd already made
before, but it's getting hard to keep track of everything in this
thread ;-)

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Rob Landley]

On Friday 15 June 2007 15:28:29 Alexandre Oliva wrote:
> On Jun 15, 2007, Rob Landley <rob@landley.net> wrote:
> > On Thursday 14 June 2007 22:25:57 Alexandre Oliva wrote:
> >> Is the signature not derived from the bits in the GPLed component, as
> >> much as it is derived from the key?
> >
> > Actually, you can't copyright, trademark, or patent a number.
>
> Agreed.  And this counter-argument of yours is a distraction.
>
> I was careful to not talk about "derived work". 

"Is the signature not derived from X as much as it is derived from Y."

"I was careful to not talk about "derived work"."

Which personality of yours am I currently addressing?

> Please read it again 
> under this clarification (that I'm pretty sure I'd already made
> before, but it's getting hard to keep track of everything in this
> thread ;-)

I'm going to stop feeding the troll now...

Rob
-- 
"One of my most productive days was throwing away 1000 lines of code."
  - Ken Thompson.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 15, 2007, Rob Landley <rob@landley.net> wrote:

> On Friday 15 June 2007 15:28:29 Alexandre Oliva wrote:
>> On Jun 15, 2007, Rob Landley <rob@landley.net> wrote:
>> > On Thursday 14 June 2007 22:25:57 Alexandre Oliva wrote:
>> >> Is the signature not derived from the bits in the GPLed component, as
>> >> much as it is derived from the key?

>> > Actually, you can't copyright, trademark, or patent a number.

>> Agreed.  And this counter-argument of yours is a distraction.

>> I was careful to not talk about "derived work". 

> "Is the signature not derived from X as much as it is derived from Y."

> "I was careful to not talk about "derived work"."

> Which personality of yours am I currently addressing?

The one that speaks English, not Legalese.  IANAL.

Last I looked it up, "derived" was a plain-English word.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Thursday 14 June 2007 17:39:32 Alexandre Oliva wrote:
<snip>
>
> And since the specific implementation involves creating a derived work
> of the GPLed kernel (the signature, or the signed image, or what have
> you) and refraining from providing the corresponding sources to that
> derived work (the key and the signature "build scripts"), I still
> think this specific case is a violation of the letter of the GPLv2,
> even if the FSF doesn't take this position.

Not entirely correct. If TiVO is making a change to the binary to include the 
signature, then it *could* be considered a derivative work. If the signature 
is stored in another place - say a bit of Flash or a separate file on the 
disc - then there is no way for it to be considered a derivative work. (Under 
US law, IIRC and I I've interpreted it (and the related cases) properly then 
the change would have to be to the source of the program for it to be 
considered a "derivative". But, as you say often and I should make clear 
myself, IANAL) 

>
> > It seems pretty obvious that the only right Tivo is withholding is the
> > right to install new versions on the device
>
> Actually, no.  They withhold the right to run versions that they don't
> authorize themselves.

And this is relevant to a software license in which way? In particular how is 
this relevant to the GPL, which has always *only* guaranteed access to the 
source if you have access to the binary, the right to distribute your own 
versions and the right to modify the code.

Since the "right to run code" was never guaranteed it *cannot* be a violation. 
It might be in conflict with what RMS intended when he wrote the first 
version of the GPL and in conflict with the intent of the people that 
contributed to GPLv2 but that doesn't matter. However, I will not use (or 
recommend) the GPLv3 in its current form because I feel it makes unnecessary 
restrictions. The fact that you have to "allow additional rights" to make it 
equal to the GPLv2 makes a functional (and spiritual) difference to me.

(Why? Because I'm opposed to "In order to protect freedom X we have to 
restrict freedom Y. Its happening in the US *RIGHT* *NOW* and I have been 
doing what I can to fight that. Now the same faulty logic is being applied by 
the FSF with GPLv3)

> Back when GPLv2 was written, the right to run was never considered an
> issue.  It was taken for granted, because copyright didn't control
> that in the US (it does in Brazil), and nobody had thought of
> technical measures to stop people from running modified copies of
> software.  At least nobody involved in GPLv2, AFAIK.

Why isn't it in the US? Because the binary form of a program does not and 
cannot have a separate copyright than the source code. Since it is the 
*SOURCE* that is actually copyright (mechanical translation cannot create a 
new work, only a new form of an already copyrighted work) guaranteeing 
the "right to run" is pointless.

And you are wrong about that "Nobody thought of it" thing - what you mean is 
that "Nobody that had a hand in drafting and ratifying the GPLv2 thought of 
it". The "NSA Guidelines for Trusted Systems" (aka "The Orange Book") talks 
about methods of preventing the execution of code.

What you and the rest of the FSF is doing in response to "tivoization" is 
saying "we don't care if it wasn't designed to do X, we want to be able to 
make it do that anyway *AND* the manufacturer has to help us do it". There is 
no legal way for you to make that demand of a hardware manufacturer. Instead 
you've gone with the only legal recourse - saying "If you want to use my 
copyrighted work under license X, you have to do Y".  I have no problem with 
that, and if the FSF wants that, it's fine by me. But, as I said, I could 
care less where and/or if something I release under the GPL is used. This 
makes the GPLv2 perfect for me.

> The landscape has changed, and GPLv3 is meant to defend this freedom
> that was taken for granted.
>
> > they never do (and really never could) "modify" the physical copy on
> > your device (which is your main argument).
>
> Qualifying it as the main argument is a bit of an exaggeration.  I
> have a number of different arguments.  The one about incomplete
> sources is the most solid IMHO.

Yes, it is. But your argument about the TiVO is "they can modify the copy on 
it but I can't". Hence it is your main argument. And remember, replace != 
modify.

> >> What do you think you do when you save a modified source file in your
> >> editor?
> >
> > Don't skip the part where the in-memory version started as an exact
> > copy of the original being replaced. Notice the difference? ;)
>
> Sorry, I really don't follow.  Both versions of the kernel binary also
> started from a common source ancestor.  Were you trying to make a
> distinction on these grounds?

No. He was making a distinction that I have seen made a number of times. That 
is, a copy of a copyright work in a computers RAM is a *distinct* copy, 
separate from the file on disc it started as. (And that has actually been 
argued in a court case. I don't recall the specifics, but I do recall that 
the argument was held as valid)

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Florin Malita]

On 06/14/2007 05:39 PM, Alexandre Oliva wrote:
> On Jun 14, 2007, Florin Malita <fmalita@gmail.com> wrote:
>   
>> No, it's not: replacing does not create derivative
>> work. Modification does.
>>     
>
> Thanks.  Good point.  This convinces me that this doesn't work as a
> legal argument under copyright.
>
> I still stand by my understanding that this restriction violates the
> spirit of the license.
>   

But since this elusive "spirit" is subject to everybody's interpretation 
of the preamble, you must surely admit that it remains just a matter of 
opinion ;)

>> It seems pretty obvious that the only right Tivo is withholding is the
>> right to install new versions on the device
>>     
>
> Actually, no.  They withhold the right to run versions that they don't
> authorize themselves.
>   

On that particular piece of hw, yes. But who's granted you the right to 
*run* your modified copy *there* in the first place? GPLv2 explicitly 
steers clear of anything "other than copying, distribution and 
modification".

> Back when GPLv2 was written, the right to run was never considered an
> issue.  It was taken for granted, because copyright didn't control
> that in the US (it does in Brazil), and nobody had thought of
> technical measures to stop people from running modified copies of
> software.  At least nobody involved in GPLv2, AFAIK.
>   
> The landscape has changed, and GPLv3 is meant to defend this freedom
> that was taken for granted.
>   

Then you agree that GPLv2 does not protect your freedom (taken for 
granted) to run a modified copy on any particular device, or am I 
misreading?

>>> What do you think you do when you save a modified source file in your
>>> editor?
>>>       
>
>   
>> Don't skip the part where the in-memory version started as an exact
>> copy of the original being replaced. Notice the difference? ;)
>>     
>
> Sorry, I really don't follow.  Both versions of the kernel binary also
> started from a common source ancestor.  Were you trying to make a
> distinction on these grounds?
>   

Exactly: they have a common ancestor, they are both derived from it. But 
there's no ancestry relationship *between* them (unlike your edited file 
example) so you cannot argue that one is a modification of the other. 
Hence, Tivo is not really *modifying* the copies it distributes with the 
device - they're *installing* brand new copies instead. They also choose 
not to offer everybody the same privilege :-|

Does this go against the intent of the GPLv2 authors? Probably. Does it 
go against the letter of GPLv2? Apparently not. Does it go against 
your/some people's interpretation of the GPL "spirit"? Obviously. Does 
it go against everybody's interpretation? Obviously not.

---
fm

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Florin Malita <fmalita@gmail.com> wrote:

> On 06/14/2007 05:39 PM, Alexandre Oliva wrote:
>> Back when GPLv2 was written, the right to run was never considered an
>> issue.  It was taken for granted, because copyright didn't control
>> that in the US (it does in Brazil), and nobody had thought of
>> technical measures to stop people from running modified copies of
>> software.  At least nobody involved in GPLv2, AFAIK.

>> The landscape has changed, and GPLv3 is meant to defend this
>> freedom that was taken for granted.

> Then you agree that GPLv2 does not protect your freedom (taken for
> granted) to run a modified copy on any particular device, or am I
> misreading?

IANAL, but AFAICT it doesn't.  Still, encoded in the spirit (that
refers to free software, bringing in the free software definition), is
the notion of protecting users' freedoms, among them the freeom #0, to
run the software for any purpose.

That's why I believe it's in the spirit of the license to defend this
freedom.

And that's why lawyers in Brazil believe that, even though the GPL
does not affirm the right to run the software, it fits the bill,
because, under the light of the preamble, the free software
definition, and the US copyright law, it should be interpreted as an
intent to grant permission to run the software.

> Hence, Tivo is not really *modifying* the copies it distributes with
> the device - they're *installing* brand new copies instead. They
> also choose not to offer everybody the same privilege :-|

Got it.  That's bad.  :-(

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Thursday 14 June 2007 23:19:24 Alexandre Oliva wrote:
> On Jun 14, 2007, Florin Malita <fmalita@gmail.com> wrote:
> > On 06/14/2007 05:39 PM, Alexandre Oliva wrote:
> >> Back when GPLv2 was written, the right to run was never considered an
> >> issue.  It was taken for granted, because copyright didn't control
> >> that in the US (it does in Brazil), and nobody had thought of
> >> technical measures to stop people from running modified copies of
> >> software.  At least nobody involved in GPLv2, AFAIK.
> >>
> >> The landscape has changed, and GPLv3 is meant to defend this
> >> freedom that was taken for granted.
> >
> > Then you agree that GPLv2 does not protect your freedom (taken for
> > granted) to run a modified copy on any particular device, or am I
> > misreading?
>
> IANAL, but AFAICT it doesn't.  Still, encoded in the spirit (that
> refers to free software, bringing in the free software definition), is
> the notion of protecting users' freedoms, among them the freeom #0, to
> run the software for any purpose.

And where in GPLv2 is "Freedom #0"?

As a simple matter of fact, the *only* activities covered by the GPLv2 
are "copying, distributing and modifying". It says so in the license itself.

> That's why I believe it's in the spirit of the license to defend this
> freedom.
>
> And that's why lawyers in Brazil believe that, even though the GPL
> does not affirm the right to run the software, it fits the bill,
> because, under the light of the preamble, the free software
> definition, and the US copyright law, it should be interpreted as an
> intent to grant permission to run the software.

Then they have made a bad decision. While it can be argued that "the right to 
run the software" is guaranteed, the truth is that the license is very clear 
about what it covers. That's *DIRECTLY* in section 0 of the license. If 
someone has interpreted it to cover something besides what it explicitly 
states then it has been badly interpreted.

In case you don't remember, GPLv2, section 2, paragraph 2:

"Activities other than copying, distribution and modification are not
covered by this License; they are outside its scope.  The act of
running the Program is not restricted, and the output from the Program
is covered only if its contents constitute a work based on the
Program (independent of having been made by running the Program).
Whether that is true depends on what the Program does."

In other words, the license cannot be sanely interpreted to cover *execution* 
of the program. Yes, it says that the *license* doesn't restrict you from 
running the program, but that *DOESN'T* matter, because the opening sentence 
says: "Activities other than copying, distribution and modification are not 
covered by this License; they are outside its scope." QED: The intent of the 
license is clear and it is to guarantee those three stated rights.

DRH

> > Hence, Tivo is not really *modifying* the copies it distributes with
> > the device - they're *installing* brand new copies instead. They
> > also choose not to offer everybody the same privilege :-|
>
> Got it.  That's bad.  :-(



-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Glauber de Oliveira Costa]

> As a simple matter of fact, the *only* activities covered by the GPLv2
> are "copying, distributing and modifying". It says so in the license itself.
Unless I have explicitly installed linux myself in the box, I have
received the binary from them, so it can fall in the distribution
case.

-- 
Glauber de Oliveira Costa.
"Free as in Freedom"
http://glommer.net

"The less confident you are, the more serious you have to act."

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Friday 15 June 2007 02:29:32 Glauber de Oliveira Costa wrote:
> > As a simple matter of fact, the *only* activities covered by the GPLv2
> > are "copying, distributing and modifying". It says so in the license
> > itself.
>
> Unless I have explicitly installed linux myself in the box, I have
> received the binary from them, so it can fall in the distribution
> case.


Sorry if you missed the rest of the discussion, but the above statement was a 
rebuttal of the "The GPLv2 intended to guarantee me the right to run the 
software on any given piece of hardware" argument that has been used as the 
justification for the addition of the "tivoization" language to the GPLv3. As 
I stated, I fail to see how "running" the program is, in any way, intended by 
the license, since it *explicitly* states that it only covers "copying, 
distribution and modification". The exact place where it does that 
is "Section 0, paragraph 2, first sentence". I'll quote it here again:
"Activities other than copying, distribution and modification are not
covered by this License; they are outside its scope."

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 15, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> On Thursday 14 June 2007 23:19:24 Alexandre Oliva wrote:

>> IANAL, but AFAICT it doesn't.  Still, encoded in the spirit (that
>> refers to free software, bringing in the free software definition), is
>> the notion of protecting users' freedoms, among them the freeom #0, to
>> run the software for any purpose.

> And where in GPLv2 is "Freedom #0"?

It may sound like thin evidence for someone arriving from Venus today,
but the preamble talks about "free software", some passages clearly
imply that software under this license is "free software", the license
is published by the Free Software Foundation, and the Free Software
Foundation has a published definition of Free Software that
establishes the 4 freedoms.

The freedoms defined there resonate very strongly with the
freedoms/rights that the license talks about.  I hope this is enough
evidence to convince you that this is the intent.

The only of the freedoms that's not explicitly mentioned in the
preamble, the freedom to run the software for any purpose, is
mentioned in the legal terms as unrestricted, which is very much in
line with freedom #0, but is outside the scope of a copyright license
because running the program does not require copyright permission.

I'll give you that the preamble doesn't make it clear that the license
is purported to defend freedom #0 too.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Friday 15 June 2007 15:49:00 Alexandre Oliva wrote:
> On Jun 15, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> > On Thursday 14 June 2007 23:19:24 Alexandre Oliva wrote:
> >> IANAL, but AFAICT it doesn't.  Still, encoded in the spirit (that
> >> refers to free software, bringing in the free software definition), is
> >> the notion of protecting users' freedoms, among them the freeom #0, to
> >> run the software for any purpose.
> >
> > And where in GPLv2 is "Freedom #0"?
>
> It may sound like thin evidence for someone arriving from Venus today,
> but the preamble talks about "free software", some passages clearly
> imply that software under this license is "free software", the license
> is published by the Free Software Foundation, and the Free Software
> Foundation has a published definition of Free Software that
> establishes the 4 freedoms.

And that doesn't matter. In the context of the GPLv2 the only legally active 
parts *ARE* in the GPLv2, under this heading:

                    GNU GENERAL PUBLIC LICENSE
   TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION


No other text published by the FSF has any legal bearing on the GPLv2 *except* 
when it is the FSF that holds copyright to the work and has placed it under 
the GPLv2. When I place a program under the GPL, it becomes *my* 
interpretation of the license and those "published texts" I might have (that 
are relevant to the situation) which have bearing on the license.

> The freedoms defined there resonate very strongly with the
> freedoms/rights that the license talks about.  I hope this is enough
> evidence to convince you that this is the intent.

Nope. Because the intent of the author of the license is worth nothing. The 
intent of the person who has placed the code under the GPL is, however, worth 
quite a bit.

> The only of the freedoms that's not explicitly mentioned in the
> preamble, the freedom to run the software for any purpose, is
> mentioned in the legal terms as unrestricted, which is very much in
> line with freedom #0, but is outside the scope of a copyright license
> because running the program does not require copyright permission.

In the context of the license it means that the copyright holder is placing no 
restrictions. And that is because they legally *cannot* - as you agreed. But 
the copyright holder *can*, OTOH, make it clear that they wish for there to 
be no limits, artificial or otherwise, placed on the code. 

> I'll give you that the preamble doesn't make it clear that the license
> is purported to defend freedom #0 too.

And the preamble carries no legal weight.

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 15, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> On Friday 15 June 2007 15:49:00 Alexandre Oliva wrote:
>> On Jun 15, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
>> > On Thursday 14 June 2007 23:19:24 Alexandre Oliva wrote:
>> >> IANAL, but AFAICT it doesn't.  Still, encoded in the spirit (that
>> >> refers to free software, bringing in the free software definition), is
>> >> the notion of protecting users' freedoms, among them the freeom #0, to
>> >> run the software for any purpose.
>> >
>> > And where in GPLv2 is "Freedom #0"?
>> 
>> It may sound like thin evidence for someone arriving from Venus today,
>> but the preamble talks about "free software", some passages clearly
>> imply that software under this license is "free software", the license
>> is published by the Free Software Foundation, and the Free Software
>> Foundation has a published definition of Free Software that
>> establishes the 4 freedoms.

> And that doesn't matter.

Doens't matter for what?

To indicate what the Linux copyright holders meant?  Sure it doesn't.
I never claimed it did.

To indicate what the authors of the GPL meant?  To indicate the spirit
of the license they wrote?  Yes, it matters a lot.

And the latter is what my participation here is all about: to show
that the spirit didn't change at all.

Until you acknowledge and understand this, I should refrain from
answering your other postings.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Chris Adams]

Once upon a time, Alexandre Oliva  <aoliva@redhat.com> said:
>> What the GPL *does* say is that you can't "add additional
>> restrictions to the license"
>
>Not quite.  It's more general than that:
>
>  You may not impose any further restrictions on the recipients'
>  exercise of the rights granted herein.

GPLv2 section 0 says:

  Activities other than copying, distribution and modification are not
  covered by this License; they are outside its scope.  The act of
  running the Program is not restricted, ...

The license does not cover running of the program.  It doesn't restrict
it, but it doesn't cover it.  Claiming otherwise is turning the GPL into
yet another dreaded EULA.

Nowhere does the GPLv2 define modification as "modify and run in place".
The Preamble emphasizes sharing; hardware is a fixed object and can't be
shared in the same fashion as software.

Also, GPLv2 section 2 includes:

  In addition, mere aggregation of another work not based on the Program
  with the Program (or with a work based on the Program) on a volume of
  a storage or distribution medium does not bring the other work under
  the scope of this License.

TiVo's firmware (and any restrictions it may carry) is not affected by
the GPLv2.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alan Cox]

>   Activities other than copying, distribution and modification are not
>   covered by this License; they are outside its scope.  The act of
>   running the Program is not restricted, ...
> 
> The license does not cover running of the program.  It doesn't restrict
> it, but it doesn't cover it.  Claiming otherwise is turning the GPL into
> yet another dreaded EULA.

For many juridisctions loading from disk into memory is copying and in
some from memory to CPU cache a second copy. This is one reason as I
understand it GPLv3 talks about "conveying" - to avoid that mess and
confusion.

>   In addition, mere aggregation of another work not based on the Program
>   with the Program (or with a work based on the Program) on a volume of
>   a storage or distribution medium does not bring the other work under
>   the scope of this License.
> 
> TiVo's firmware (and any restrictions it may carry) is not affected by
> the GPLv2.

Really irrelevant to the discussion. Tivo's firmware is up to them.
Whether the resulting system permits them to include GPLv2 software with
it is what matters.

Alan

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:

>> Activities other than copying, distribution and modification are not
>> covered by this License; they are outside its scope.  The act of
>> running the Program is not restricted, ...
>> 
>> The license does not cover running of the program.  It doesn't restrict
>> it, but it doesn't cover it.  Claiming otherwise is turning the GPL into
>> yet another dreaded EULA.

> For many juridisctions loading from disk into memory is copying and in
> some from memory to CPU cache a second copy. This is one reason as I
> understand it GPLv3 talks about "conveying" - to avoid that mess and
> confusion.

Hmm...  This is interesting.  Let me sidetrack a little bit.

Who would be held liable should the copy not be authorized by the
copyright holder?  The designer of the hardware?  The seller?  The
person who powerer the computer on?  The author of the boot loader (if
I'm talking about the kernel about to be loaded).

> Really irrelevant to the discussion. Tivo's firmware is up to them.
> Whether the resulting system permits them to include GPLv2 software with
> it is what matters.

+1

It all boils down to whether they're keeping their promise to not
impose further restrictions on the exercise of the rights granted in
the license.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bernd Petrovitsch]

On Thu, 2007-06-14 at 19:37 -0300, Alexandre Oliva wrote:
> On Jun 14, 2007, Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:
[...]
> > For many juridisctions loading from disk into memory is copying and in
> > some from memory to CPU cache a second copy. This is one reason as I
> > understand it GPLv3 talks about "conveying" - to avoid that mess and
> > confusion.
> 
> Hmm...  This is interesting.  Let me sidetrack a little bit.
> 
> Who would be held liable should the copy not be authorized by the
> copyright holder?  The designer of the hardware?  The seller?  The
> person who powerer the computer on?  The author of the boot loader (if

The person copying the software (read: the poor user) - of course with
no doubt.
Is the manufacturer of a knife or rifle or car or brick responsible if I
kill someone with it?

> I'm talking about the kernel about to be loaded).

And to solve the above "legal" problem, (at least in .at) it is
explicitly legal (and stated in the local law) to "copy" software for
personal use as long as you don't pass it on to others. And this
includes (of course) e.g. backup copies on DVDs (and remember, you are
not allowed to give them away, just store them in your desk).

	Bernd
-- 
Firmix Software GmbH                   http://www.firmix.at/
mobil: +43 664 4416156                 fax: +43 1 7890849-55
          Embedded Linux Development and Services

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Thu, 14 Jun 2007, Alexandre Oliva wrote:
> 
> People don't get your copy, so they're not entitled to anything about
> it.
> 
> When they download the software, they get another copy, and they have
> a right to modify that copy.

Umm. I notice how you must have known how *idiotic* your response was, 
because you snipped away the part where I talked about Red Haty 
distributing CD-ROM's.

In other words, Red Hat distributes copies (and yes, you *get* that copy), 
and you cannot modify that copy that you got.

So the "right to make changes" _must_ be separate from the actual copy of 
the image.

And don't get fooled by the "all the rights that you have". That 
_obviously_ and clearly talks about "the program", which in turn 
equally obviously and clearly has to be about something bigger than the 
"one copy", since the GPLv2 requires you have the right to change it.

So you edited out the part where I talked about CD's. That's the proof 
that your reading is untenable, because obviously you cannot change the 
program on the CD: you got a copy, but the right to make modifications 
wasn't ON THAT HARDWARE.

> > And here's a big clue for people: anybody who thinks that I'm violating 
> > the GPLv2 by not giving out my private SSH key to master.kernel.org is a 
> > f*cking moron!
> 
> Agreed, except I'd probably use a lighter term.

Hey, I'm not exactly known for being polite. I tell it how I see it, and I 
tend to be pretty damn blunt about  it. 

> > See any parallels here? Any parallel to a CD-ROM distribution, or a Tivo 
> > distribution?
> 
> Yes.  You see how TiVO is different?  It is modifyable, and I actually
> receive the copy that TiVO can still modify, but I can't.

You keep on harping on that "modifyable", but no-where in the GPLv2 is 
that an issue. I claim that it *cannot* be an issue, since CD's are 
obviously ok.

So the "modifyable" part is a totally new thing to the GPLv3.

You cannot use that as an argument that the GPLv3 didn't change things, 
that's a circular agument: "the GPLv3 says so, so thus the GPLv3 is in the 
same spirit as the GPLv2". Doesn't make sense.

The fact is, the GPLv3 does fundamentally new things. Things I didn't sign 
up for (and things that nobody _else_ signed up for either) when I chose 
the GPLv2 for the kernel.

The fact that some people would like to change the kernel license to GPLv3 
is no different from the fact that some other people would like to cgange 
the kernel license to the BSD license.

Those people who have argued for using the BSD license, btw, argued so in 
the name of "freedom". No different from you. Do you think they were 
right? If so, why the hell do you think _you_ are right?

So here's what it fundamentally boils down to:

 - do you admit that the GPLv3 is a new license that does new and 
   different things?

 - do you admit that I chose the GPLv2, and have argued that I chose it 
   because I understood what it said?

 - do you admit that authors have the right to choose their own licenses?

And here's the fundamental answer:

 - if you answered "no" to any of the above questions, you're either 
   stupid (the first two questions) or a douche-bag (the third one)

 - if you answered "yes" to all the above questions, HOW THE HELL can you 
   call me confused, and argue against me when I say that the GPLv2 is a 
   better license? It wasn't your choice. 

It really is that easy. 

			Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:

> On Thu, 14 Jun 2007, Alexandre Oliva wrote:
>> 
>> People don't get your copy, so they're not entitled to anything about
>> it.
>> 
>> When they download the software, they get another copy, and they have
>> a right to modify that copy.

> Umm. I notice how you must have known how *idiotic* your response was, 
> because you snipped away the part where I talked about Red Haty 
> distributing CD-ROM's.

> In other words, Red Hat distributes copies (and yes, you *get* that copy), 
> and you cannot modify that copy that you got.

And Red Hat can't either.  I thought that was quite obvious.

>> > See any parallels here? Any parallel to a CD-ROM distribution, or a Tivo 
>> > distribution?

>> Yes.  You see how TiVO is different?  It is modifyable, and I actually
>> receive the copy that TiVO can still modify, but I can't.

> You keep on harping on that "modifyable", but no-where in the GPLv2 is 
> that an issue. I claim that it *cannot* be an issue, since CD's are 
> obviously ok.

The 'passing on the rights you have' makes it an issue.

> You cannot use that as an argument that the GPLv3 didn't change things, 

Compare the preambles of v2 and v3 and you'll understand why the
argument is sound, and not circular.

> Those people who have argued for using the BSD license, btw, argued so in 
> the name of "freedom".

But individual freedom, rather than community freedom.

Think local vs global optimization.

> If so, why the hell do you think _you_ are right?

Because, like you, I'm always right, even though not everyone agrees
with that assessment? ;-P :-D

>  - do you admit that the GPLv3 is a new license that does new and 
>    different things?

Yes, of course.  The new legal terms are answers to new threats to the
freedoms depicted in the preamble, that didn't exist or hadn't been
thought of by the time GPLv2 was published.

>  - do you admit that I chose the GPLv2, and have argued that I chose it 
>    because I understood what it said?

Yes.

>  - do you admit that authors have the right to choose their own licenses?

Within the boundaries of ethics and morals, yes.

>  - if you answered "yes" to all the above questions, HOW THE HELL can you 
>    call me confused, and argue against me when I say that the GPLv2 is a 
>    better license? It wasn't your choice. 

The thing is I'm not arguing that point.

I'm disputing that there was a change in the spirit of the license
between v2 and v3.  Heck, a mere 48 hours ago you couldn't even tell
the spirit from the legal terms.


I still think v3 will serve better any Free Software community,
because it will push away the abusers that contribute little, or turn
them into cooperative or at least harmless participants, that further
enable the active participation of their downstream users.  This would
enable wider participation under the same 'tit-for-tat' conditions
that you attribute to GPLv2.


It appears to me that the only significant point of contention
remaining is the issue of Tivoization.  If you feel so strongly about
permitting Tivoization, even though it denies the freedoms that the
original spirit of the license you chose says they are entitled to
have, you can make this provision by means of an additional permission
for that, on top of GPLv3, and be done with it.

> It really is that easy. 

Yes, indeed.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Thu, 14 Jun 2007, Alexandre Oliva wrote:
> 
> > In other words, Red Hat distributes copies (and yes, you *get* that copy), 
> > and you cannot modify that copy that you got.
> 
> And Red Hat can't either.  I thought that was quite obvious.

That's TOTALLY IRRELEVANT!

There is no language in the GPLv2 (only in the GPLv3 drafts) about "same 
upgradability as third parties".

You're arguing a point that DOES NOT EXIST in the GPLv2.

The GPLv2 talks about specific rights, like the ability to make changes 
and distribute things, and says that you have to give downstream all those 
same rights.

And I've pointed out to you (now about five times) that those rights 
CANNOT be able "in-place", since even Red Hat does not actually give you 
the right to do in-place modification of the software they sell.

> The 'passing on the rights you have' makes it an issue.

No. It does not.

I have extra rights as a copyright holder, and that "the rights you have" 
are as they pertain to the software under the GPLv2, not as it pertains to 
the physical device, or outside the GPLv2.

For example, for any code that I have full copyright over, I have rights 
that you DO NOT HAVE! I have the right to re-license it under some other 
license. The fact that I pass on a copy of the software to you under the 
GPLv2 does *not* give you those rights, but that's not even what the GPLv2 
asks for!

The GPLv2, when it talks about "passing on the rights", talks about the 
rights you got *per*the*GPLv2*. 

Any other reading is nonsensical, since the copyrigth owner *always* has 
more rights than a licensee! I legally literally *couldn't* pass over all 
the rights I have to my software! If you read the GPLv2 as meaning that I 
have to, you are mis-reading it. It's that simple.

Anyway, I'm not interested in continuing this flame war.

The fact is, the license for the kernel is the GPLv2. And I think it's a 
superior license. As such, I'd be a total moron to relicense the kernel 
under what I believe is a worse license.

So if you want to argue that I should re-license, you should argue that 
the GPLv3 is better. And quite frankly, you haven't.

		Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:

> On Thu, 14 Jun 2007, Alexandre Oliva wrote:
>> 
>> > In other words, Red Hat distributes copies (and yes, you *get* that copy), 
>> > and you cannot modify that copy that you got.
>> 
>> And Red Hat can't either.  I thought that was quite obvious.

> The GPLv2 talks about specific rights, like the ability to make changes 
> and distribute things, and says that you have to give downstream all those 
> same rights.

The spirit gives the intuition of "passing on all the rights".  The
legal terms have to be more careful about that, to avoid the very
situation you're debating, so they state "you can't impose further
restrictions on the exercise of the rights".

Do you understand the difference?

> For example, for any code that I have full copyright over, I have rights 
> that you DO NOT HAVE!

No dispute about that, and this is irrelevant to this point.  I've
already responded and clarified this point 2 or 3 times in this
thread.  Do you need me to find a URL for you?  It was in respose to
Dmitri Torokhov.

> So if you want to argue that I should re-license, you should argue that 
> the GPLv3 is better. And quite frankly, you haven't.

In fact, I haven't even tried.  So far, I've merely been trying to
show that it still follows the same spirit, dispelling the muth that
it doesn't, and trying to understand why you think GPLv2 is so much
better, which I think is related with tit-for-tat and retribution in
kind.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 13, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> Now stop parroting the FSF's worn and tired tripe.

Are you playing Linus' sheeple and parroting his lines just to make a
point, or are you like that all the time? ;-)

> PS: Looking at your .sig I guess maybe you can't do that without getting 
> kicked out of the FSF-LA

Don't worry, I'm not speaking even for FSFLA, and I'm entitled to my
own opinion.

I haven't consulted other FSFLA members about this.  This is all my
own personal opinion.

It just so happens that I'm very closely involved in the process, I've
spent a lot of time thinking about it, and I happen to share a similar
moral and ethical background with others involved in the process, so I
arrive at similar conclusions.

And then, I influence the process myself, so it's not like some of the
arguments I brought up here weren't taken into account while creating
the GPLv3, and adopted by its other proponents.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Wednesday 13 June 2007 21:04:42 Alexandre Oliva wrote:
> On Jun 13, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> > Now stop parroting the FSF's worn and tired tripe.
>
> Are you playing Linus' sheeple and parroting his lines just to make a
> point, or are you like that all the time? ;-)

Nope. I'm just tired of giving proof after proof that you're wrong and having 
you restate the same tripe.

> > PS: Looking at your .sig I guess maybe you can't do that without getting
> > kicked out of the FSF-LA
>
> Don't worry, I'm not speaking even for FSFLA, and I'm entitled to my
> own opinion.

Certainly. I never said otherwise. What I stated and then *implied* was that 
you are repeating the same false logic over and over again trying to make 
people believe that it isn't borked and that that false logic is exactly the 
same crap I've seen from the FSF numerous times.

> I haven't consulted other FSFLA members about this.  This is all my
> own personal opinion.

Where I am examining the facts and drawing a logical conclusion. That it 
happens to form an opinion is secondary to the truth.

> It just so happens that I'm very closely involved in the process, I've
> spent a lot of time thinking about it, and I happen to share a similar
> moral and ethical background with others involved in the process, so I
> arrive at similar conclusions.

Okay. Still doesn't explain why you have argued that the GPLv3 doesn't attempt 
to cover hardware and then provide proof that it does. 

> And then, I influence the process myself, so it's not like some of the
> arguments I brought up here weren't taken into account while creating
> the GPLv3, and adopted by its other proponents.

This is no surprise - I had a feeling this was the truth. Not that it changes 
my opinion at all. As I've said, I have never liked the GPL at all, but v2 is 
the best that exists - even though I've put together custom licenses myself, 
none of them have had the number of lawyers look at them that the GPLv2 has 
had.

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 13, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> Still doesn't explain why you have argued that the GPLv3 doesn't
> attempt to cover hardware and then provide proof that it does.

It doesn't cover hardware, in the same way that it doesn't cover
patents, and it doesn't cover pro-DRM laws.  It merely arranges, as
best as we've managed a copyright license to do, that they can't be
used as excuses (or tools) to disrespect the freedoms that the GPL
demands all licensees to respect for other users.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Wednesday 13 June 2007 22:04:04 Alexandre Oliva wrote:
> On Jun 13, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> > Still doesn't explain why you have argued that the GPLv3 doesn't
> > attempt to cover hardware and then provide proof that it does.
>
> It doesn't cover hardware, in the same way that it doesn't cover
> patents, and it doesn't cover pro-DRM laws.  It merely arranges, as
> best as we've managed a copyright license to do, that they can't be
> used as excuses (or tools) to disrespect the freedoms that the GPL
> demands all licensees to respect for other users.

Consider this scenario:
Small company A is manufacturing a new WiFi router.
They decide to have it run HURD as the OS.
In complying with the GPLv3 they supply the signing keys and everything else 
needed to install a new kernel on the hardware.
User B buys the router and modifies the kernel so it drives the WiFi to an 
output power twice that which it is licensed to carry.
FCC finds out and prosecutes User B for violating the regulations.
FCC then pulls the small companies license until they change their hardware so 
the driver can't push it to transmit at a higher power level and levies a 
fine.
Small company A loses the money paid on the fine, has to recall all the 
devices that can be modified (through software) to break the law at a massive 
cost *AND* has to redesign their hardware. The total cost drives the company 
into bankruptcy.

Small companies C,D and E, in order to avoid the fate of small company A, 
purchases a license for proprietary OS "F" to drive their new hardware.

Net loss: A lot of the users and publicity that "Free Software" used to get, 
because GPLv3 contains language that opens the companies to lawsuits that 
they wouldn't otherwise face.

Which is better: Growing the base of installed GPL covered software, 
or "ethics and morals" that demand the language that has been added to the 
GPLv3 ? Personally I'd like to see proprietary software driven into a very 
small "niche" market or entirely out of existence. However much I want this 
to happen, I cannot be anything *BUT* scared of the GPLv3 simply because I 
see it creating massive problems - and all because of a *small* portion of 
the new language it contains. It has taken almost 15 years for "Free 
Software" to make a dent in the market, and, IMHO, a lot of that is both 
Linux and the "holes" in GPLv2.

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> User B buys the router and modifies the kernel so it drives the WiFi to an 
> output power twice that which it is licensed to carry.
> FCC finds out and prosecutes User B for violating the regulations.

Ok so far.

> FCC then pulls the small companies license until they change their
> hardware so the driver can't push it to transmit at a higher power
> level and levies a fine.

I'd say this is unfair, but if it can happen, then maybe the small
company could have been more careful about the regulations.  There are
various ways to prevent these changes that don't involve imposing
restrictions of modification on any software in the device, all the
way from hardware-constrained output power to hardware-verified
authorized configuration parameters.

> Growing the base of installed GPL covered software,

When this doesn't bring freedom to people, when people can't actually
enjoy the freedoms that the software is supposed to provide, I don't
see why this would be a good thing.  What's the merit in being able to
claim "vendor X chose my Free Software and locked it down such that
users don't get the freedoms I meant for them, and I'm happy about it?"

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Robin Getz]

On Thu 14 Jun 2007 01:07, Alexandre Oliva pondered:
> On Jun 14, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> 
> > User B buys the router and modifies the kernel so it drives the WiFi
> to an 
> > output power twice that which it is licensed to carry.
> > FCC finds out and prosecutes User B for violating the regulations.
> 
> Ok so far.
> 
> > FCC then pulls the small companies license until they change their
> > hardware so the driver can't push it to transmit at a higher power
> > level and levies a fine.
> 
> I'd say this is unfair, but if it can happen, 

it can (at least hypothetically in the US - I do not know of any actual 
cases - anyone?).

> then maybe the small 
> company could have been more careful about the regulations.  There are
> various ways to prevent these changes that don't involve imposing
> restrictions of modification on any software in the device, all the
> way from hardware-constrained output power to hardware-verified
> authorized configuration parameters.

As a person pretty familiar with the hardware in these types of devices - this 
just isn't practical. 

The power output goes to the air, has some to do with any external power amp, 
the antenna design/connectors, the layout of the printed circuit board (PCB), 
the materials of the PCB, and numerous other factors that are outside the 
chip where the register sits that controls the "output power". This is why 
the register is there - so many things exist in the real world that require 
it to be changed to actually get desired power output in the air.

What you are asking is possible - that people who make the standard CMOS 
radios add EEPROM or FLASH to their processes - or they could add an 
interface to a serial EEPROM where these hardware limits could sit - and 
during manufacturing could program these to add a hardware limit.

But this seems - just silly - adding extra cost/complexity to the hardware - 
because you limit things in the zero-cost space?

Or run the driver in user space, and forget about it. (Like everyone is 
starting to do on the desktop).

> When this doesn't bring freedom to people, when people can't actually
> enjoy the freedoms that the software is supposed to provide, I don't
> see why this would be a good thing.  What's the merit in being able to
> claim "vendor X chose my Free Software and locked it down such that
> users don't get the freedoms I meant for them, and I'm happy about it?"

Sometimes joy comes from knowing that your contribution was valuable.

I would ask about two other examples:

 - medical devices - which must have their software certified by the FDA (at 
least in the US) - and the manufacture can not allow non-certified software 
to be loaded on it - and would be classified "designed or sold for 
incorporation into a dwelling" (per the GPL3 dd1v3). Think devices like :

http://www.powerheart.com/products/ [1]

Are Linux developers "happy" knowing their contributions will never help save 
a life? And this joy goes to Microsoft or VxWorks? Today it can go to Linux 
developers, since the GPL2 license doesn't restrict this like the GPL3 does.


 - gambling devices - which must have their software certified by various 
government agencies - to make sure that the odds are known, and there are no 
backdoors, and consumers don't get screwed - the manufacture can not allow 
non-certified software to be loaded on it. If these are in a hotel - where 
various people live - is that considered "incorporation into a dwelling"?

Not wanting to start a debate about the morality of being involved in the 
gambling industry - (if the statically challenged are giving the government 
money to keep my taxes down, I am mostly OK with it) - but I'm not "happy" 
thinking that someone can ledgistate restrictions on embedded OS choice, just 
because it must be verified by a third party.


Hypothetically - tragedy strikes - use of a modified Nao robot [2] contributes 
to someone's death (remember - modifications of the OS removed the three 
laws). Someone (maybe Microsoft or VxWorks, or GreenHills) steps up, and 
lobbies the government - that embedded OSes need tighter control to ensure 
safety - and need to be validated by the government. (like the FCC and FDA is 
already doing in the US).

Since the GPL3 states" - 
"If you cannot convey the Program, or other covered work, so as to satisfy 
simultaneously your obligations under this License and any other pertinent 
obligations, then as a consequence you may not convey it at all."

Forget DRM - forget Tivo - anyone who develops a product which must be 
certified can not use anything that is under the GPL3.

On Sun, 10 Jun 2007 10:29:04 Linus Torvalds pondered:
> I still think GPLv2 is simply the better license.
>
> I consider dual-licensing unlikely (and technically quite hard), but at 
> least _possible_ in theory. I have yet to see any actual *reasons* for 
> licensing under the GPLv3, though.

I think there are actual reasons _not_ to.

What I have told people who have asked me (who are thinking about putting 
Linux in a medical device) is kernel version 2.6.xx (current version at the 
time), is released under GPL2.  No one can go back and change the license on 
a previously released package. 

It might be easier to say that the kernel licence will only be reviewed during 
a major bump in the kernel version number (3.x.x) and that all 2.x kernels 
will be under the same license as what exists today.

-Robin

[1] - I don't know if this device has Linux in it or not - but this just 
represents classes of devices that may never be able to use Linux.

[2]  http://linuxdevices.com/news/NS6263763539.html

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Robin Getz <rgetz@blackfin.uclinux.org> wrote:

> On Thu 14 Jun 2007 01:07, Alexandre Oliva pondered:

>> then maybe the small 
>> company could have been more careful about the regulations.  There are
>> various ways to prevent these changes that don't involve imposing
>> restrictions of modification on any software in the device, all the
>> way from hardware-constrained output power to hardware-verified
>> authorized configuration parameters.

> As a person pretty familiar with the hardware in these types of
> devices - this just isn't practical.

I actually left out the most obvious one: store the program in ROM.
Is that not practical?

You're claiming that adding hardware locks and chains and bolts,
implemented with help from the loader software, is simpler than just
using ROM?

Well, then, ok: do all that loader and hardware signature-checking
dancing, sign the image, store it in the machine, and throw the
signing key away.  This should be good for the highly-regulated areas
you're talking about.  And then, since you can no longer modify the
program, you don't have to let the user do that any more.  Problem
solved.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Rob Landley]

On Thursday 14 June 2007 13:46:40 Alexandre Oliva wrote:
> On Jun 14, 2007, Robin Getz <rgetz@blackfin.uclinux.org> wrote:
> > On Thu 14 Jun 2007 01:07, Alexandre Oliva pondered:
> >> then maybe the small
> >> company could have been more careful about the regulations.  There are
> >> various ways to prevent these changes that don't involve imposing
> >> restrictions of modification on any software in the device, all the
> >> way from hardware-constrained output power to hardware-verified
> >> authorized configuration parameters.
> >
> > As a person pretty familiar with the hardware in these types of
> > devices - this just isn't practical.
>
> I actually left out the most obvious one: store the program in ROM.
> Is that not practical?
>
> You're claiming that adding hardware locks and chains and bolts,
> implemented with help from the loader software, is simpler than just
> using ROM?

As far as I know, I'm the first one who brought up the "the current GPLv3 
draft forbids burning your code into ROM, you idiots" argument back before 
Bruce Perens cost the BusyBox project my services over this very issue.  (Not 
that I didn't lock the license of that down to v2 and chase him away before I 
left, I was just too disgusted to ever again contribute to a project he'd 
named.  Yeah, I hold a grudge.)

Although it's kind of amusing to watch you attempt to dictate terms to 
hardware manufacturers, the answer to your question is "yes".  Having flash 
is sometimes simpler and cheaper than having ROM.  It means you don't have to 
burn a new mask to bump the firmware revision (especially on a low-volume 
production run, where "low volume" here is tens or hundreds of thousands 
instead of millions).  It makes the thing a lot more field serviceable (you 
can upgrade the firmware without a chip puller).  It means one physical chip 
can give you both read-only and persistent writeable storage.   And flash 
chips produced in high enough unit volumes honestly can be cheaper than a 
custom ROM, pricing in semiconductors is all about unit volume.

> Well, then, ok: do all that loader and hardware signature-checking
> dancing, sign the image, store it in the machine, and throw the
> signing key away.  This should be good for the highly-regulated areas
> you're talking about.  And then, since you can no longer modify the
> program, you don't have to let the user do that any more.  Problem
> solved.

A) Does that actually satisfy the terms of GPLv3?  If so, can't they just wait 
until they get sued and destroy the keys then?

B) There are actually manufacturers who would be happy with your straw man.  
Lots of companies in the far east produce products that infringe on patents 
from 30 different competitors, and rather than try to license everything 
(which isn't even always possible) they spin off a shell company (or nested 
series thereof), design and manufacture a product, sell a production run of 
them into the distribution channel, and then dissolve the shell company 
before the inventory hits retailers.  But the time anybody is in a position 
to take an enforcement action, the company to take the action against is 
gone.  (Who are you going to sue, customers who bought the devices?  The 
distributors who bought the inventory in good faith, and will then refuse to 
distribute any of YOUR product if you attack 'em?)  There's always the 
possibility that somebody will sit down and follow the paper trail back to 
the parent company (through the multiple legal jurisdictions where nobody 
speaks english), but since they're likely as not to destroy this kind of info 
_anyway_ while burying their trail...

Rob

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alan Cox]

> B) There are actually manufacturers who would be happy with your straw man.  
> Lots of companies in the far east produce products that infringe on patents 
> from 30 different competitors, and rather than try to license everything 
> (which isn't even always possible) they spin off a shell company (or nested 
> series thereof), design and manufacture a product, sell a production run of 
> them into the distribution channel, and then dissolve the shell company 
> before the inventory hits retailers.  But the time anybody is in a position 

This isn't just done for IPR, in fact in many fields IPR is a non-issue.
The primary reason for this practice is to render US health and safety
regulation irrelevant and to prevent class action suits if/when your
device kills someone.

Alan

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Rob Landley <rob@landley.net> wrote:

> On Thursday 14 June 2007 13:46:40 Alexandre Oliva wrote:

>> Well, then, ok: do all that loader and hardware signature-checking
>> dancing, sign the image, store it in the machine, and throw the
>> signing key away.  This should be good for the highly-regulated areas
>> you're talking about.  And then, since you can no longer modify the
>> program, you don't have to let the user do that any more.  Problem
>> solved.

> A) Does that actually satisfy the terms of GPLv3?

I think so:

  this requirement does not apply if neither you nor any third party
  retains the ability to install modified object code on the User
  Product

> If so, can't they just wait until they get sued and destroy the keys
> then?

I don't think this woulnd't satisfy the above.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Robin Getz]

On Thu 14 Jun 2007 13:46, Alexandre Oliva pondered:
> On Jun 14, 2007, Robin Getz <rgetz@blackfin.uclinux.org> wrote:
> > As a person pretty familiar with the hardware in these types of
> > devices - this just isn't practical.
> 
> Well, then, ok: do all that loader and hardware signature-checking
> dancing, sign the image, store it in the machine, and throw the
> signing key away.  This should be good for the highly-regulated areas
> you're talking about.  And then, since you can no longer modify the
> program, you don't have to let the user do that any more.  Problem
> solved.

I don't think so - the GPL3 doesn't state that you must convey the same rights 
to end users that you have, it says you must provide installation 
information, including your keys, or you can not ship the product.

That is the way I read the following sections (let me know if I mis-read 
anything):

======================
"Installation Information" for a User Product means any methods, procedures, 
authorization keys, or other information required to install and execute 
modified versions of a covered work in that User Product from a modified 
version of its Corresponding Source. 

The information must suffice to ensure that the continued functioning of the 
modified object code is in no case prevented or interfered with solely 
because modification has been made.

If conditions are imposed on you (whether by court order, agreement or 
otherwise) that contradict the conditions of this License, they do not excuse 
you from the conditions of this License. 

If you cannot convey the Program, or other covered work, so as to satisfy 
simultaneously your obligations under this License and any other pertinent 
obligations, then as a consequence you may not convey it at all.

===================

I read "are imposed on you (or otherwise)" to mean "by you" as well. If so, 
you throwing away the private keys are not an option.

I need to think a bit more of Rob's opinion of ROM's are illegal - but 
providing the installation information of "send $1M NRE and object code to 
xxx ROM vendor, and wait 16 weeks for 500k units, take one to a board shop, 
pay $1k for them to re-work your BGA - if the xray says it is screwed up, you 
have 499,999 other units to try." - may meet the language, but doesn't meet 
the spirit of the GPL either...

-Robin

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 15, 2007, Robin Getz <rgetz@blackfin.uclinux.org> wrote:

> On Thu 14 Jun 2007 13:46, Alexandre Oliva pondered:
>> On Jun 14, 2007, Robin Getz <rgetz@blackfin.uclinux.org> wrote:
>> > As a person pretty familiar with the hardware in these types of
>> > devices - this just isn't practical.
>> 
>> Well, then, ok: do all that loader and hardware signature-checking
>> dancing, sign the image, store it in the machine, and throw the
>> signing key away.  This should be good for the highly-regulated areas
>> you're talking about.  And then, since you can no longer modify the
>> program, you don't have to let the user do that any more.  Problem
>> solved.

> I don't think so - the GPL3 doesn't state that you must convey the same rights 
> to end users that you have,

Right, this is only in the preamble.

> it says you must provide installation information, including your
> keys, or you can not ship the product.

Unless you throw the keys away:

  this requirement does not apply if neither you nor any third party
  retains the ability to install modified object code on the User
  Product (for example, the work has been installed in ROM).

> I need to think a bit more of Rob's opinion of ROM's are illegal

See above ;-)

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bron Gondwana]

On Thu, Jun 14, 2007 at 10:14:21AM -0400, Robin Getz wrote:
>  - gambling devices - which must have their software certified by various 
> government agencies - to make sure that the odds are known, and there are no 
> backdoors, and consumers don't get screwed - the manufacture can not allow 
> non-certified software to be loaded on it. If these are in a hotel - where 
> various people live - is that considered "incorporation into a dwelling"?
> 
> Not wanting to start a debate about the morality of being involved in the 
> gambling industry - (if the statically challenged are giving the government 
> money to keep my taxes down, I am mostly OK with it) - but I'm not "happy" 
> thinking that someone can ledgistate restrictions on embedded OS choice, just 
> because it must be verified by a third party.

Why not go really controversial and dive straight in with "voting
machines".  There's a whole 'nother can of worms.

Bron.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alan Cox]

> the new language it contains. It has taken almost 15 years for "Free 
> Software" to make a dent in the market, and, IMHO, a lot of that is both 
> Linux and the "holes" in GPLv2.


You appear terminally confused. The purpose of the GPL as defined by its
authors is not commercial success, world domination or making zillions of
dollars - it is keeping the software protected by that license "free" in
terms of liberty as measured against the set of freedoms to
run/modify/etc they discuss in the licence document.

The fact this is a good license for making zillions of dollars, producing
good software and the like is either incidental or a logical result of the
protection of freedoms depending upon which views you believe.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Ingo Molnar]

* Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:

> > the new language it contains. It has taken almost 15 years for "Free 
> > Software" to make a dent in the market, and, IMHO, a lot of that is both 
> > Linux and the "holes" in GPLv2.
> 
> You appear terminally confused. The purpose of the GPL as defined by 
> its authors is not commercial success, world domination or making 
> zillions of dollars - it is keeping the software protected by that 
> license "free" in terms of liberty as measured against the set of 
> freedoms to run/modify/etc they discuss in the licence document.
> 
> The fact this is a good license for making zillions of dollars, 
> producing good software and the like is either incidental or a logical 
> result of the protection of freedoms depending upon which views you 
> believe.

that's fine, but the fundamental question is: where is the moral 
boundary of the power that the copyright license gives? The FSF seems to 
believe "nowhere, anything that copyright law allows us to achieve our 
goals is a fair game" - and the GPLv3 shows that belief. I dont 
subscribe to that view. I think the proper limit is the boundary where 
the limit of the software is - because that's the only sane and globally 
workable way to stop the power-hungry. I.e. the information we produce 
is covered by the rules of the GPL. It might be used in ways 
inconvenient to us, it might be put on hardware we dont like (be that a 
Tivo, a landmine or an abortion instrument) but that does not change the 
fundamental fact: it's outside the _moral scope_ of our power. Whether 
some jurisdictions allow the control of _other_ information via our 
information is immaterial. If a jurisdiction allows the control of 
hardware that is associated with our software, so what? If a 
jurisdiction allows the controlling of various aspects of movie theaters 
that happen to play copyrighted movies, does it make it morally right?

	Ingo

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alan Cox]

> that's fine, but the fundamental question is: where is the moral 
> boundary of the power that the copyright license gives? The FSF seems to 

Assuming a democratic state then the laws of the land ought to reflect the
'general will' (if you believe Rousseau anyway). They should thus define
the boundary ['derivative work' generally ] according to the general good
and with the consent of the people.

> hardware that is associated with our software, so what? If a 
> jurisdiction allows the controlling of various aspects of movie theaters 
> that happen to play copyrighted movies, does it make it morally right?

Does that question not suppose some positivist absolute morality ? I
suspect many would argue that it is moral to do so if the end goal of the
controls is moral. You might also want to apply the tests in Fuller's
Internal Morality of Law ?

I'm not sure Philosophy is on topic for l/k however 8)

Alan

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Ingo Molnar]

* Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:

> > that's fine, but the fundamental question is: where is the moral 
> > boundary of the power that the copyright license gives? The FSF 
> > seems to
> 
> Assuming a democratic state then the laws of the land ought to reflect 
> the 'general will' (if you believe Rousseau anyway). They should thus 
> define the boundary ['derivative work' generally ] according to the 
> general good and with the consent of the people.

uhm, so if the MPAA and the RIAA pays for another nice piece of 
legislation that extends the power of copyright owners, do you find it 
morally justified to use those powers, as long as it's argued to be in 
favor of some long-term goal that you judge to be moral, even if it 
results in some "temporary injustice"?

i think that could be the main difference in thinking. I argue that the 
only way to be moral is to be moral _now_, not "later, once this very 
important fight for the common good is over". I think the moral approach 
to this is to say _no_ to attempts to extend the license to beyond the 
"moral scope" of the software we wrote - regardless of what new powers 
are legislated into the hands of copyright owners. It's naturally hard 
to do, because giving up power is always hard to do.

In other words: we need to apply our concepts of freedom and fairness 
not only to the end result, but to the means and methods of achieving 
those end results as well. The end goals are often forgotten, it's the 
process that matters to the end result.

Or in yet another set of words: this concept of morality also happens to 
be expressed fairly accurately in the thousands of years of 'quid pro 
quo' concept. (shared amongst many, many cultures on this planet, shared 
amongst far more cultures than the western 'freedom' concept.) (which 
concept of quid-pro-quo fairness is likely coded into our brains and 
into our thinking genetically - because it's a simple and very efficient 
group survival method.)

	Ingo

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Rob Landley]

On Thursday 14 June 2007 08:25:46 Ingo Molnar wrote:
> * Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:
> > > that's fine, but the fundamental question is: where is the moral
> > > boundary of the power that the copyright license gives? The FSF
> > > seems to
> >
> > Assuming a democratic state then the laws of the land ought to reflect
> > the 'general will' (if you believe Rousseau anyway). They should thus
> > define the boundary ['derivative work' generally ] according to the
> > general good and with the consent of the people.
>
> uhm, so if the MPAA and the RIAA pays for another nice piece of
> legislation that extends the power of copyright owners, do you find it
> morally justified to use those powers, as long as it's argued to be in
> favor of some long-term goal that you judge to be moral, even if it
> results in some "temporary injustice"?

Turnabout is fair play, and unilateral disarmament is a bad strategy in a 
mexican standoff?

Finding it morally justified to _have_ powers is not the same as finding it 
morally justified to _use_ powers you have anyway.  Lots of companies (like 
Red Hat) amass defensive software patent portfolios because the patent system 
is so screwed up.

But then, I'm a pragmatist, not an idealist.  You can be one or the other and 
make it work.  Mixing the two tends to suck.  Being ruthlessly pragmatic in 
the pursuit of an ideal (as the FSF seems to be doing) has often been a 
recipe for disaster...

Rob

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Ingo Molnar]

* Rob Landley <rob@landley.net> wrote:

> > uhm, so if the MPAA and the RIAA pays for another nice piece of 
> > legislation that extends the power of copyright owners, do you find 
> > it morally justified to use those powers, as long as it's argued to 
> > be in favor of some long-term goal that you judge to be moral, even 
> > if it results in some "temporary injustice"?
> 
> Turnabout is fair play, and unilateral disarmament is a bad strategy 
> in a mexican standoff?
> 
> Finding it morally justified to _have_ powers is not the same as 
> finding it morally justified to _use_ powers you have anyway.  Lots of 
> companies (like Red Hat) amass defensive software patent portfolios 
> because the patent system is so screwed up.

but the GPLv3 definitely takes action against Tivo. It's not "defensive" 
in any way. It is outright hostile, it irreversibly cuts off certain 
people from being to distribute GPLv3-ed software alongside with certain 
types of hardware that the FSF's president does not like. (who, 
incidentally, is a mathematician who last wrote significant free 
software perhaps a decade ago, and who thus must have a great and 
thorough understanding of how hardware and software works today and who 
must also have a deep knowledge about what makes the free software 
community tick.)

The GPLv2 never did this kind of restriction _of other works_. Yes, you 
can use copyright law to control other works and thus (if the affected 
work is a hardware device for example) to control the _use_ of the free 
software, but it is _wrong_. The GPLv2 specifically said, in section 0:

  Activities other than copying, distribution and modification are not
  covered by this License; they are outside its scope.

guess why this section has been completely removed from the GPLv3, 
without a replacement?

	Ingo

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 15, 2007, Ingo Molnar <mingo@elte.hu> wrote:

> it irreversibly cuts off certain people from being to distribute
> GPLv3-ed software alongside with certain types of hardware that the
> FSF's president does not like.

That's not true.  They can just as well throw the key away and refrain
from modifying the installed software behind the users' back.

> The GPLv2 never did this kind of restriction _of other works_.

How about other works in which GPLed software is distributed?

I think your interpretation is mistaken or at least incomplete.

>   Activities other than copying, distribution and modification are not
>   covered by this License; they are outside its scope.

> guess why this section has been completely removed from the GPLv3, 
> without a replacement?

My guess:

First, because it was redundant, given that the license didn't quite
discuss other activities.  Unless you count say "imposing restrictions
on the exercise of others' freedoms" as other activities, even though
these are associated with modification and distribution.

Second, because GPLv3 does indeed talk about other activities, such as
starting lawsuits on patent and pro-DRM grounds, or entering
agreements for distribution of software along with limited patent
licenses.  All of these are still associated, at least to some extent,
with modification and distribution, but I guess it was worth
clarifying that claiming that such harmful activities are outside the
scope of the license isn't a valid excuse to escape the conditions
determined by the license.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Ingo Molnar]

* Alexandre Oliva <aoliva@redhat.com> wrote:

> On Jun 15, 2007, Ingo Molnar <mingo@elte.hu> wrote:
> 
> > it irreversibly cuts off certain people from being to distribute
> > GPLv3-ed software alongside with certain types of hardware that the
> > FSF's president does not like.
> 
> That's not true.  They can just as well throw the key away and refrain 
> from modifying the installed software behind the users' back.

uhm, so you claim that my argument is false, and your proof for that is 
a "non-upgradeable Tivo"?? <sarcasm> That is a _great_ idea. Not being 
able to patch security holes. Not being able to fix bugs. Not being able 
to add new features. Makes complete sense. Will be a hit on the market! 
Every PVR maker will flock from Windows to Linux i'm sure. </sarcasm>

really, do you even _read_ what you write? All your arguments so far 
were instantly debunkable. This is one of the lowest quality GPL 
discussions i was ever involved in ...

furthermore, the fact that the GPLv3 had to add carved out exceptions 
for the anti-Tivo languge is further _proof_ that the whole idea is 
absurd to begin with! It's like writing a nice new function to implement 
something, and then when it shows many design flaws, you'd not just 
admit that it's flawed and would get rid of it and redesign it, you'd 
instead pretend that it's fine and you'd carve out a few of the more 
common failure modes and would hack it around in that case.

> >   Activities other than copying, distribution and modification are 
> >   not covered by this License; they are outside its scope.
> 
> > guess why this section has been completely removed from the GPLv3, 
> > without a replacement?
> 
> My guess:
> 
> First, because it was redundant, given that the license didn't quite 
> discuss other activities.  Unless you count say "imposing restrictions 
> on the exercise of others' freedoms" as other activities, even though 
> these are associated with modification and distribution.

here you prove that you cannot even read what i wrote. I wrote that this 
section has been removed from the GPLv3. What relevance does it have 
that in your opinion this section was redundant in the GPLv2?? It would 
clearly not be redundant in the GPLv3: it would contradict and 
_completely neutralize_ most of the crap from the GPLv3 that we are 
talking about here ...
 
> Second, because GPLv3 does indeed talk about other activities, such as 
> starting lawsuits on patent and pro-DRM grounds, or entering 
> agreements for distribution of software along with limited patent 
> licenses.  All of these are still associated, at least to some extent, 
> with modification and distribution, but I guess it was worth 
> clarifying that claiming that such harmful activities are outside the 
> scope of the license isn't a valid excuse to escape the conditions 
> determined by the license.

dont you realize that declaring certain types of activities by hardware 
makers as being "against freedom" is _exactly_ such an activity that the 
GPLv2 did not attempt to control? I could tell you offhand a dozen more 
examples of human activities that restrict the 4 GNU freedoms of users 
_much more_ than the Tivo ever did: for example censure, opression of 
free speech, out of control climate, dictatorship, campaign financing 
laws, the WIN32 API and human stupidity. By your argument we'd have to 
add prohibition against those restrictions of freedom to the license 
too, right? Your argument still leads to absurd results, even now that 
you've modified it a few times already ...

	Ingo

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 15, 2007, Ingo Molnar <mingo@elte.hu> wrote:

> * Alexandre Oliva <aoliva@redhat.com> wrote:

>> On Jun 15, 2007, Ingo Molnar <mingo@elte.hu> wrote:
>> 
>> > it irreversibly cuts off certain people from being to distribute
>> > GPLv3-ed software alongside with certain types of hardware that the
>> > FSF's president does not like.

>> That's not true.  They can just as well throw the key away and refrain 
>> from modifying the installed software behind the users' back.

> uhm, so you claim that my argument is false, and your proof for that is 
> a "non-upgradeable Tivo"?? <sarcasm> That is a _great_ idea. Not being 
> able to patch security holes. Not being able to fix bugs. Not being able 
> to add new features. Makes complete sense.

Oh, so you think patching security holes, fixing bugs and adding new
features are good ideas?  What if you can't do it in your TiVo?

>> > guess why this section has been completely removed from the GPLv3, 
>> > without a replacement?

>> My guess:

>> First, because it was redundant, given that the license didn't quite 
>> discuss other activities.  Unless you count say "imposing restrictions 
>> on the exercise of others' freedoms" as other activities, even though 
>> these are associated with modification and distribution.

> here you prove that you cannot even read what i wrote. I wrote that this 
> section has been removed from the GPLv3. What relevance does it have 
> that in your opinion this section was redundant in the GPLv2??

If you didn't mean "removed from the GPLv3 as compared with v2", I
misunderstood what you wrote.

The fact that it's redundant is v2 means it is reasonable to take it
out.  That's the relevance.

> It would clearly not be redundant in the GPLv3: it would contradict
> and _completely neutralize_ most of the crap from the GPLv3 that we
> are talking about here ...
 
And, per the same reasoning, some of the v2 provisions as well.

> dont you realize that declaring certain types of activities by hardware 
> makers as being "against freedom" is _exactly_ such an activity that the 
> GPLv2 did not attempt to control?

No.  And some Linux hackers disagree with your assessment too.

> censure, opression of free speech, out of control climate,
> dictatorship, campaign financing laws, the WIN32 API and human
> stupidity. By your argument we'd have to add prohibition against
> those restrictions of freedom to the license too, right?

-ENONSEQUITUR

How do these stop a user's exercise of the four freedoms of a piece of
software licensed under the GPL?

> Your argument still leads to absurd results, even now that you've
> modified it a few times already ...

I hope you're not saying that my listening to you, recognizing
mistakes in my arguments and fixing them up is a bad thing.

But hey, at least I'm not modifying my arguments as much as you are!
;-)

It's pretty easy to shoot a straw man and claim the original argument
was broken.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Scott Preece]

On 6/15/07, Alexandre Oliva <aoliva@redhat.com> wrote:
> On Jun 15, 2007, Ingo Molnar <mingo@elte.hu> wrote:
>

> How do these stop a user's exercise of the four freedoms of a piece of
> software licensed under the GPL?
---

I know you don't see it that way, but I still find it bizarre that
"the right to modify the software" should be construed as "implies the
right to modify the device that the software was shipped in".

I do agree that it's not a change in "spirit" - I'm sure the GPL
authors would have disliked TiVoization 15 years ago as much as they
do today, if they had thought about it (regardless of the Stallman
interview where he said he didn't care very much about devices).

However, whether it is a change "in spirit" or not, it clearly is a
qualitative change that substantially changes the rights granted under
the license and it's perfectly reasonable for some authors who liked
the GPLv2 to dislike and reject GPLv3.

scott

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Friday 15 June 2007 20:22:50 Alexandre Oliva wrote:
> On Jun 15, 2007, Ingo Molnar <mingo@elte.hu> wrote:
> > * Alexandre Oliva <aoliva@redhat.com> wrote:
> >> On Jun 15, 2007, Ingo Molnar <mingo@elte.hu> wrote:
> >> > it irreversibly cuts off certain people from being to distribute
> >> > GPLv3-ed software alongside with certain types of hardware that the
> >> > FSF's president does not like.
> >>
> >> That's not true.  They can just as well throw the key away and refrain
> >> from modifying the installed software behind the users' back.
> >
> > uhm, so you claim that my argument is false, and your proof for that is
> > a "non-upgradeable Tivo"?? <sarcasm> That is a _great_ idea. Not being
> > able to patch security holes. Not being able to fix bugs. Not being able
> > to add new features. Makes complete sense.
>
> Oh, so you think patching security holes, fixing bugs and adding new
> features are good ideas?  What if you can't do it in your TiVo?
>
> >> > guess why this section has been completely removed from the GPLv3,
> >> > without a replacement?
> >>
> >> My guess:
> >>
> >> First, because it was redundant, given that the license didn't quite
> >> discuss other activities.  Unless you count say "imposing restrictions
> >> on the exercise of others' freedoms" as other activities, even though
> >> these are associated with modification and distribution.
> >
> > here you prove that you cannot even read what i wrote. I wrote that this
> > section has been removed from the GPLv3. What relevance does it have
> > that in your opinion this section was redundant in the GPLv2??
>
> If you didn't mean "removed from the GPLv3 as compared with v2", I
> misunderstood what you wrote.
>
> The fact that it's redundant is v2 means it is reasonable to take it
> out.  That's the relevance.

It isn't redundant at all. I specifies the definitions of several terms used 
in the GPLv2 and also defines the exact scope of the license. If you feel 
that the definition of the terms and the limitation of scope were redundant 
then you are sadly mistaken.

> > It would clearly not be redundant in the GPLv3: it would contradict
> > and _completely neutralize_ most of the crap from the GPLv3 that we
> > are talking about here ...
>
> And, per the same reasoning, some of the v2 provisions as well.

For a license to be legally enforceable it must be internally consistent. 
Without that internal consistency it becomes very easy to circumvent it. The 
GPLv2's definitions and defined scope - as per section 0 - define the limits 
of the license and are entirely consistent with the rest of it. What it 
*isn't* consistent with is the FSF's other "propaganda" and the wants of the 
FSF to make certain activities verboten in GPLv3.

> > dont you realize that declaring certain types of activities by hardware
> > makers as being "against freedom" is _exactly_ such an activity that the
> > GPLv2 did not attempt to control?
>
> No.  And some Linux hackers disagree with your assessment too.

And that is their right. However, it appears to a nearly unanimous consensus 
that it is the truth. It may not be liked by some people, but likes and 
dislikes don't matter.

> > censure, opression of free speech, out of control climate,
> > dictatorship, campaign financing laws, the WIN32 API and human
> > stupidity. By your argument we'd have to add prohibition against
> > those restrictions of freedom to the license too, right?
>
> -ENONSEQUITUR
>
> How do these stop a user's exercise of the four freedoms of a piece of
> software licensed under the GPL?
>
> > Your argument still leads to absurd results, even now that you've
> > modified it a few times already ...
>
> I hope you're not saying that my listening to you, recognizing
> mistakes in my arguments and fixing them up is a bad thing.
>
> But hey, at least I'm not modifying my arguments as much as you are!
> ;-)
>
> It's pretty easy to shoot a straw man and claim the original argument
> was broken.

Yep. I've done it to you on more than one occasion, Alexandre. The part that 
makes me laugh is that you still haven't realized it.

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Ingo Molnar]

* Alexandre Oliva <aoliva@redhat.com> wrote:

> On Jun 15, 2007, Ingo Molnar <mingo@elte.hu> wrote:
> 
> >>> it irreversibly cuts off certain people from being to distribute
> >>> GPLv3-ed software alongside with certain types of hardware that 
> >>> the FSF's president does not like.
> >>>
> >> That's not true.  They can just as well throw the key away and 
> >> refrain from modifying the installed software behind the users' 
> >> back.
> 
> > uhm, so you claim that my argument is false, and your proof for that 
> > is a "non-upgradeable Tivo"?? <sarcasm> That is a _great_ idea. Not 
> > being able to patch security holes. Not being able to fix bugs. Not 
> > being able to add new features. Makes complete sense.
> 
> Oh, so you think patching security holes, fixing bugs and adding new 
> features are good ideas?  What if you can't do it in your TiVo?

this has to be one of the most bizarre arguments i've read in this 
thread as of date. Are you seriously questioning the notion that it's a 
good and legitimate idea for a hardware vendor to make the system 
fixable, patchable and upgradable? Are you seriously suggesting that for 
a hardware vendor to be able to offer such a solution, if they are under 
the unescapable restriction of content providers that the system itself 
must be tamper-proof, it should not be able to use a GPL-ed kernel at 
all? Because that is what your arguments lead to, and that is what the 
GPLv3 implements. In case you didnt notice: RMS _does not want the Tivo 
to use a GPLv3 kernel_, simple as that, and the GPLv3 achieves that. He 
wants Tivo to either to go out of business or to go to WinCE or some 
other OS. Did you ever think about the meaning of the "anti" word in the 
"anti-Tivo" expression? Hint: it's not some friendly suggestion of 
cooperation and working together ;)

	Ingo

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 17, 2007, Ingo Molnar <mingo@elte.hu> wrote:

> * Alexandre Oliva <aoliva@redhat.com> wrote:

>> On Jun 15, 2007, Ingo Molnar <mingo@elte.hu> wrote:
>> 
>> >>> it irreversibly cuts off certain people from being to distribute
>> >>> GPLv3-ed software alongside with certain types of hardware that 
>> >>> the FSF's president does not like.
>> >>>
>> >> That's not true.  They can just as well throw the key away and 
>> >> refrain from modifying the installed software behind the users' 
>> >> back.
>> 
>> > uhm, so you claim that my argument is false, and your proof for that 
>> > is a "non-upgradeable Tivo"?? <sarcasm> That is a _great_ idea. Not 
>> > being able to patch security holes. Not being able to fix bugs. Not 
>> > being able to add new features. Makes complete sense.
>> 
>> Oh, so you think patching security holes, fixing bugs and adding new 
>> features are good ideas?  What if you can't do it in your TiVo?

> this has to be one of the most bizarre arguments i've read in this 
> thread as of date. Are you seriously questioning the notion that it's a 
> good and legitimate idea for a hardware vendor to make the system 
> fixable, patchable and upgradable?

No.  I'm questioning why the vendor could keep this privilege to
itself.

> Are you seriously suggesting that for a hardware vendor to be able
> to offer such a solution, if they are under the unescapable
> restriction of content providers that the system itself must be
> tamper-proof, it should not be able to use a GPL-ed kernel at all?

No, and I've already explained how I believe this can be accomplished
with the wording in the GPLv3dd4, although IANAL to tell whether
that's correct.

Just make the tivoization machinery require two keys: one that the
vendor keeps, one that the vendor gives to the user (maybe without
ever knowing it).  Neither one can install modifications alone, but
the user can approve modifications by the vendor, and the vendor can
approve modifications by the user.  This is still not ideal, but it at
least doesn't permit the vendor to remove features from under the
user.

> Because that is what your arguments lead to, and that is what the 
> GPLv3 implements.

You haven't really read that bit of dd3 or dd4, have you?

Or the various portions of this thread in which I showed your
assumptions are utterly broken?

> RMS _does not want the Tivo to use a GPLv3 kernel_,

I know you're not stupid, but I can't tell whether you're malicious or
just misinformed.

RMS does not want TiVo (or anyone else) to disrespect users' freedoms,
and installing technical measures to prevent users from adapting the
software to suit their needs and running their modifications is
disrespecting users freedoms.

That he is not opposed to the idea of TiVo using a GPLv3 kernel is
easy to see, if you take the time to read the draft instead of
spreading false assumptions about it:

  this requirement does not apply if neither you nor any third party
  retains the ability to install modified object code on the User
  Product

> Hint: it's not some friendly suggestion of cooperation and working
> together ;)

Hey, wouldn't this be just tit-for-tat? ;-)

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Jan Harkes]

On Sun, Jun 17, 2007 at 05:17:57AM -0300, Alexandre Oliva wrote:
> Just make the tivoization machinery require two keys: one that the
> vendor keeps, one that the vendor gives to the user (maybe without
> ever knowing it).  Neither one can install modifications alone, but
> the user can approve modifications by the vendor, and the vendor can
> approve modifications by the user.  This is still not ideal, but it at
> least doesn't permit the vendor to remove features from under the
> user.

So what features has Tivo removed (or threatened to remove) from the GPL
licensed parts? I think at some point they disabled the _undocumented_
skip feature in their own proprietary software, and ended up reenabling
it when a lot of customers complained. Even if those customers had the
ability to replace any GPL licensed parts, it would not have reenabled
the feature. And it was an undocumented easter egg type thing at that,
it isn't like they widely announced it in their advertising or as a
selling point.

So Google is using Linux right. What if they remove some feature? (let's
pick a randomg one, i.e. phone number lookup) Should I get a keycard for
their machine room to fix the problem, or maybe we should use some
secret sharing mechanism to prevent them from removing the feature.

> RMS does not want TiVo (or anyone else) to disrespect users' freedoms,
> and installing technical measures to prevent users from adapting the
> software to suit their needs and running their modifications is
> disrespecting users freedoms.

So if Tivo would allow you to boot your own kernel, but keeps the
harddisk encrypted if the booted kernel does not have the right
signature? In such a case you can run your own kernel and if you replace
the harddrive you can install all the applications you might want. You
cannot however use their software, any of the recorded content or obtain
any further guide data/service updates.

And how is that any different from taking an off-the-shelf PC and
booting your own kernel with Tivo's modifications? Or really different
from the current situation.

Tivo complied in as far as they made GPL licensed code available, you
can examine it, modify it, compile it. You just can't use it _in
combination with TiVo's own software and service_. I didn't think the
GPLv2 covered anything related to use and you have retained all the
freedoms the GPL promises.

> That he is not opposed to the idea of TiVo using a GPLv3 kernel is
> easy to see, if you take the time to read the draft instead of
> spreading false assumptions about it:
> 
>   this requirement does not apply if neither you nor any third party
>   retains the ability to install modified object code on the User
>   Product

So they keep the system locked down, but include perl/python/emacs and
distribute updates in the form of scripts/source code which are either
interpreted or compiled to a ramfs filesystem at boot. Time to add
another exception?

Jan

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 17, 2007, Jan Harkes <jaharkes@cs.cmu.edu> wrote:

> On Sun, Jun 17, 2007 at 05:17:57AM -0300, Alexandre Oliva wrote:
>> Just make the tivoization machinery require two keys: one that the
>> vendor keeps, one that the vendor gives to the user (maybe without
>> ever knowing it).  Neither one can install modifications alone, but
>> the user can approve modifications by the vendor, and the vendor can
>> approve modifications by the user.  This is still not ideal, but it at
>> least doesn't permit the vendor to remove features from under the
>> user.

> So what features has Tivo removed (or threatened to remove) from the GPL
> licensed parts?

Why does it matter?  The point is that a tivoizer *could* do that.

> So Google is using Linux right. What if they remove some feature?

Are you claiming Google is tivoizing something in their internal
infrastructure?  They're not distributing or conveying that software,
so, nothing wrong with that.

Or are you talking about their search appliance, which I know nearly
nothing about?

> So if Tivo would allow you to boot your own kernel, but keeps the
> harddisk encrypted if the booted kernel does not have the right
> signature?

  The information must suffice to ensure that the continued
  functioning of the modified object code is in no case prevented or
  interfered with solely because modification has been made.

> And how is that any different from taking an off-the-shelf PC and
> booting your own kernel with Tivo's modifications?

TiVo did not sell me that off-the-shelf PC with the Free Software in
it.  It (hypothetically) sold me a computer with technical measures
meant to restrict my ability to adapt the software it shipped to my
own needs and to run it for any purpose, while it can still do that.
That's a difference.

>>   this requirement does not apply if neither you nor any third
>>   party retains the ability to install modified object code on the
>>   User Product

> So they keep the system locked down, but include perl/python/emacs and
> distribute updates in the form of scripts/source code which are either
> interpreted or compiled to a ramfs filesystem at boot. Time to add
> another exception?

Intent behind this?: weasel out of the obligations of the license.
Anyone, probably even a US court, might very well see it that way.

They retain the ability to modify the software, so they ought to pass
it on to the user.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Scott Preece]

On 6/15/07, Alexandre Oliva <aoliva@redhat.com> wrote:
> On Jun 15, 2007, Ingo Molnar <mingo@elte.hu> wrote:
>
> > it irreversibly cuts off certain people from being to distribute
> > GPLv3-ed software alongside with certain types of hardware that the
> > FSF's president does not like.
>
> That's not true.  They can just as well throw the key away and refrain
> from modifying the installed software behind the users' back.
---

This characterization misses something important.  For many product
devices, like cell phones, the modification is never "behind the
user's back", but is done because the user has requested it (to fix a
problem or add a new feature). If you go to ROM-based software, the
user loses, because problems can't be fixed. For certain kinds of
problems the user might be able to get a replacement device, but
potentially involving losing any data stored on the device.

The FSF's approval of this distinction (ROM versus replaceable) places
the FSF's particular principles over users interests, for no
particular reason - if the manufacturer believes that it cannot
legally allow software modification, all the restriction does is force
them either to make the software unmodifiable (which advances freedom
not at all) or to use software under a different license (which
advances freedom not at all). The result? The user STILL has no
freedom to modify the software and the community around the software
is diminished.

To go back to the "behind your back" claim, the only cases I know
where the software is replaced behind the users' back are cases were
the updates are done by a service (usually not operated by the device
manufacturer) that the user has voluntarily requested (like TiVo
program guides or cable system subscriptions), which is generally a
cases outside the scope of the license in any case.

scott

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 15, 2007, "Scott Preece" <sepreece@gmail.com> wrote:

> On 6/15/07, Alexandre Oliva <aoliva@redhat.com> wrote:
>> On Jun 15, 2007, Ingo Molnar <mingo@elte.hu> wrote:
>> 
>> > it irreversibly cuts off certain people from being to distribute
>> > GPLv3-ed software alongside with certain types of hardware that the
>> > FSF's president does not like.
>> 
>> That's not true.  They can just as well throw the key away and refrain
>> from modifying the installed software behind the users' back.

> This characterization misses something important.  For many product
> devices, like cell phones, the modification is never "behind the
> user's back"

Okay, take out the "behind the users' back", it makes no difference.
That was just to highlight the frequent evil intentions behind keeping
the keys.

I wonder if giving half the key to the user and keeping the other half
would be enough to satisfy the GPLv3 language while still enabling the
vendor and user to update the software together.

> The FSF's approval of this distinction (ROM versus replaceable) places
> the FSF's particular principles over users interests, for no
> particular reason

Over *users* interest?  How so?

> if the manufacturer believes that it cannot legally allow software
> modification, all the restriction does is force them either to make
> the software unmodifiable (which advances freedom not at all) or to
> use software under a different license (which advances freedom not
> at all).

Right.


But if the manufacturer believes that it can legally allow it, and
wants to be able to install, software modifications, then it must
decide between giving that up and letting the user do it as well.  And
this is where the users interests may prevail.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Scott Preece]

On 6/15/07, Alexandre Oliva <aoliva@redhat.com> wrote:
> On Jun 15, 2007, "Scott Preece" <sepreece@gmail.com> wrote:
>
> >> That's not true.  They can just as well throw the key away and refrain
> >> from modifying the installed software behind the users' back.
>
> > This characterization misses something important.  For many product
> > devices, like cell phones, the modification is never "behind the
> > user's back"
>
> Okay, take out the "behind the users' back", it makes no difference.
> That was just to highlight the frequent evil intentions behind keeping
> the keys.
---

Yes, but in highlighting the possibility of evil intentions you
distort the fact that usually there are no such evil intentions...

---
>
> I wonder if giving half the key to the user and keeping the other half
> would be enough to satisfy the GPLv3 language while still enabling the
> vendor and user to update the software together.
---

I think that's a possibility. I don't see how it's functionally
different from the usual case where the manufacturer can't modify the
device without the user's consent simply because the user has physical
access to the device and the manufacturer doesn't.

---
>
> > The FSF's approval of this distinction (ROM versus replaceable) places
> > the FSF's particular principles over users interests, for no
> > particular reason
>
> Over *users* interest?  How so?
---

Users benefit from the ability to get software updates, from the
manufacturer, to resolve problems, fix security vulnerabilities, and
provide updated functionality.

---
>
> > if the manufacturer believes that it cannot legally allow software
> > modification, all the restriction does is force them either to make
> > the software unmodifiable (which advances freedom not at all) or to
> > use software under a different license (which advances freedom not
> > at all).
>
> Right.
>
>
> But if the manufacturer believes that it can legally allow it, and
> wants to be able to install, software modifications, then it must
> decide between giving that up and letting the user do it as well.  And
> this is where the users interests may prevail.
---

You're harping on the "cannot legally", which is fine but irrelevant.
Whether it's a legal requirement or a business decision, the result is
the same - neither forcing the manufacturer to make the device
non-updatable nor forcing the manufacturer to use different software
benefits anyone. I don't know of interesting cases where the
manufacturer makes the device non-modifiable out of sheer
bloody-mindedness.

I don't believe that the existence of this clause will lead to more
manufacturers making their devices modifiable - there are too many
other options if they think that non-modifiability is important to
them.

[Note that I *do* think it's perfectly appropriate that authors who
feel that they don't want their work used in such devices should be
able to license them in line with that belief. I just don't think it
has any practical value aside from making them feel better.]

scott

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 15, 2007, "Scott Preece" <sepreece@gmail.com> wrote:

> On 6/15/07, Alexandre Oliva <aoliva@redhat.com> wrote:
>> > The FSF's approval of this distinction (ROM versus replaceable) places
>> > the FSF's particular principles over users interests, for no
>> > particular reason

>> Over *users* interest?  How so?

> Users benefit from the ability to get software updates, from the
> manufacturer, to resolve problems, fix security vulnerabilities, and
> provide updated functionality.

Which they could have the option to do themselves if the manufacturer
didn't prohibit them from doing so.

>> > if the manufacturer believes that it cannot legally allow software
>> > modification, all the restriction does is force them either to make
>> > the software unmodifiable (which advances freedom not at all) or to
>> > use software under a different license (which advances freedom not
>> > at all).

>> Right.

>> But if the manufacturer believes that it can legally allow it, and
>> wants to be able to install, software modifications, then it must
>> decide between giving that up and letting the user do it as well.  And
>> this is where the users interests may prevail.

> Whether it's a legal requirement or a business decision, the result is
> the same - neither forcing the manufacturer to make the device
> non-updatable nor forcing the manufacturer to use different software
> benefits anyone.

I agree.  But that's an incomplete picture.

It's the other part of the picture, that you left out twice, that is
the case that is good for the users *and* for the community.

> I don't believe that the existence of this clause will lead to more
> manufacturers making their devices modifiable - there are too many
> other options if they think that non-modifiability is important to
> them.

> [Note that I *do* think it's perfectly appropriate that authors who
> feel that they don't want their work used in such devices should be
> able to license them in line with that belief. I just don't think it
> has any practical value aside from making them feel better.]

They can do that with GPLv3.  And those who don't want to stop this
can then add a special permission.  And then everybody wins.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Scott Preece]

On 6/15/07, Alexandre Oliva <aoliva@redhat.com> wrote:
> On Jun 15, 2007, "Scott Preece" <sepreece@gmail.com> wrote:
>

> > Whether it's a legal requirement or a business decision, the result is
> > the same - neither forcing the manufacturer to make the device
> > non-updatable nor forcing the manufacturer to use different software
> > benefits anyone.
>
> I agree.  But that's an incomplete picture.
>
> It's the other part of the picture, that you left out twice, that is
> the case that is good for the users *and* for the community.
---

I don't think I "left it out". The point is that if the manufacturer
is unwilling to give the right to modify, no change in the language is
going to cause the user to have that right.

scott

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 16, 2007, "Scott Preece" <sepreece@gmail.com> wrote:

> On 6/15/07, Alexandre Oliva <aoliva@redhat.com> wrote:
>> On Jun 15, 2007, "Scott Preece" <sepreece@gmail.com> wrote:
>> 

>> > Whether it's a legal requirement or a business decision, the result is
>> > the same - neither forcing the manufacturer to make the device
>> > non-updatable nor forcing the manufacturer to use different software
>> > benefits anyone.
>> 
>> I agree.  But that's an incomplete picture.
>> 
>> It's the other part of the picture, that you left out twice, that is
>> the case that is good for the users *and* for the community.
> ---

> I don't think I "left it out". The point is that if the manufacturer
> is unwilling to give the right to modify, no change in the language is
> going to cause the user to have that right.

If the alternatives are worse for the manufacturer than letting the
user have it, then it will have the intended effects.  In the other
cases, it won't make much of a difference for anyone else.

The question is: how does tivoization help the community (under the
tit-for-tat reasoning)?  Does it help more than anti-tivoization?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Tim Post]

On Fri, 2007-06-15 at 19:52 -0500, Scott Preece wrote:

> 
> Yes, but in highlighting the possibility of evil intentions you
> distort the fact that usually there are no such evil intentions...
> 

I don't think you can use "usually" and "fact" together like that. Why
is it so bad to account for them since they (do) surface and (could)
increase significantly in frequency? 

For me, the (could) is enough to act upon, regardless of the current
likely hood of it happening. Things change frequently. 

This, unfortunately comes pre-distorted depending on what you believe.
All of us are right but we still don't agree. Quite a fluke. 

That's the problem.

Best,
--Tim

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Scott Preece]

On 6/15/07, Tim Post <tim.post@netkinetics.net> wrote:
> On Fri, 2007-06-15 at 19:52 -0500, Scott Preece wrote:
>
> >
> > Yes, but in highlighting the possibility of evil intentions you
> > distort the fact that usually there are no such evil intentions...
> >
>
> I don't think you can use "usually" and "fact" together like that. Why
> is it so bad to account for them since they (do) surface and (could)
> increase significantly in frequency?
---

I agree that it is possible to have different definitions of "evil"
and "ethical"...

scott

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Ingo Molnar]

* Alexandre Oliva <aoliva@redhat.com> wrote:

> > if the manufacturer believes that it cannot legally allow software 
> > modification, all the restriction does is force them either to make 
> > the software unmodifiable (which advances freedom not at all) or to 
> > use software under a different license (which advances freedom not 
> > at all).
> 
> Right.
> 
> But if the manufacturer believes that it can legally allow it, and 
> wants to be able to install, software modifications, then it must 
> decide between giving that up and letting the user do it as well.  And 
> this is where the users interests may prevail.

with the little tiny problem that this is not what the GPLv3 actually 
implements. Our point from the very beginning: the GPLv3 "outlaws" 
certain hardware restrictions _even if they are fully legitimate_. Yes, 
of course, it also outlaws 'bad' uses of DRM. The GPLv3 tries to carve 
out some known 'good' uses of DRM (to stop the GPLv3 from being 
_totally_ unusable in vast areas of the marketplace), but that limited 
opt-in approach can in no way be the right solution (think about it as a 
whitelist - wouldnt you want to be able to _add_ to that whitelist?? The 
GPLv3 hardcodes it.)

In other words: i dont want the police to start shooting innocent people 
in the streets, in their pursuit of criminals. Yes, this means criminals 
have an easier job getting away - but _that_ is the price of freedom!

and all these problems of the GPLv3 DRM language derives from the same 
root issue: RMS is trying to make a manual call about what _technology 
use_ is 'good' and what is 'evil'. For some of these calls we might even 
agree. But most fundamentally, a license should _never_ get into the 
business of trying to 'judge' what _use_ is 'good' and what is 'evil'. 
As you can see it on this list alone, some people see Tivo's intentions 
as legitimate, they did the DRM to stay in business but still be able to 
use free software, employ free software developers, show Mythbox how to 
do this stuff on Linux, etc. But the GPLv3 completely destroys Tivo's 
ability to use Linux, were Linux to be under the GPLv3. And by doing 
that, those contested provisions of the GPLv3 itself become a tool 
against "freedom".

You tried to find a workaround for that, by suggesting the 'dont do 
security fixes then', 'use a split key', 'use a rent model' solutions, 
but dont you realize that by suggesting those you are explicitly against 
the intent of RMS, who wants to _stop_ Tivo from being able to do DRM? 
Dont you think it speaks volumes of the GPLv3's quality that you have to 
go out and search for a _workaround_, for a _back door in the license_, 
you have to _go against the intent of RMS_, to be able to implement 
something simple as upgradability? And that even after several attempts 
you have yet to find a solution that actually enables the Tivo? (and 
that is not an accident: RMS _does not want_ the Tivo to use a GPLv3'd 
kernel, and the GPLv3 is structured that way.)

	Ingo

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 17, 2007, Ingo Molnar <mingo@elte.hu> wrote:

> * Alexandre Oliva <aoliva@redhat.com> wrote:

>> > if the manufacturer believes that it cannot legally allow software 
>> > modification, all the restriction does is force them either to make 
>> > the software unmodifiable (which advances freedom not at all) or to 
>> > use software under a different license (which advances freedom not 
>> > at all).
>> 
>> Right.
>> 
>> But if the manufacturer believes that it can legally allow it, and 
>> wants to be able to install, software modifications, then it must 
>> decide between giving that up and letting the user do it as well.  And 
>> this is where the users interests may prevail.

> with the little tiny problem that this is not what the GPLv3 actually 
> implements.

Except that it does.  Go read it, then come back and admit you were
mistaken and spreading lies about it.

The vendor must decide between respecting the freedom of the user, or
stopping itself from modifying the software too.

> But most fundamentally, a license should _never_ get into the 
> business of trying to 'judge' what _use_ is 'good' and what is 'evil'. 

What it does is to seek to carry out its mission (*) of defending
users' freedoms.  Obstacles that are placed to impede the user from
enjoying the freedoms are supposed to not be permitted by the GPL.

(*) it seems that understanding "spirit of a license" is very
difficult for you; does the term "mission" help you understand what
the GPL means when it says "similar in spirit"?

> But the GPLv3 completely destroys Tivo's ability to use Linux, were
> Linux to be under the GPLv3.

Sorry, wrong.  Barring nonsense.

> You tried to find a workaround for that, by suggesting the 'dont do 
> security fixes then', 'use a split key', 'use a rent model' solutions, 

FTR, rent model wasn't me, and it doesn't escape the GPLv3dd4
obligations IIUC, IANAL.

> but dont you realize that by suggesting those you are explicitly against 
> the intent of RMS, who wants to _stop_ Tivo from being able to do DRM? 

You misunderstand not only the spirit of the license, but his
intentions.  Oh, wait!  They're the same, that's why.

Respect and defend users' freedoms.

Repeat after me until it sinks.

I know you're not stupid.  Why do you pretend to be?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bernd Paysan]

[-- Attachment #1: Type: text/plain, Size: 2474 bytes --]

On Thursday 14 June 2007 12:38, Ingo Molnar wrote:
> that's fine, but the fundamental question is: where is the moral
> boundary of the power that the copyright license gives? The FSF seems to
> believe "nowhere, anything that copyright law allows us to achieve our
> goals is a fair game" - and the GPLv3 shows that belief. I dont
> subscribe to that view. I think the proper limit is the boundary where
> the limit of the software is - because that's the only sane and globally
> workable way to stop the power-hungry. I.e. the information we produce
> is covered by the rules of the GPL. It might be used in ways
> inconvenient to us, it might be put on hardware we dont like (be that a
> Tivo, a landmine or an abortion instrument) but that does not change the
> fundamental fact: it's outside the _moral scope_ of our power.

Where is the boundary between hard- and software? I'm employed as hardware 
designer, and for this purpose, I write programs in a hardware description 
language, which can be converted into hardware through a synthesis 
software. I write firmware, which is assembled into binary and gets placed 
on on-chip memory (ROM or NVM). I've even studied computer science, and 
electric engineering was just a side-course. I know how transistors work, 
and how gates are implemented in terms of transistors, but in essense, it's 
not that relevant unless you want to do analog circuits. Usually, during 
the development phase, we put the Verilog into an FPGA, where the 
configuration file still is obviously "software" in any sense it can be. 
I've even released descriptions of some parts of the work I do under GPL 
for people to put it into their own FPGAs.

There is no boundary between hard- and software in the sense of that 
hardware is something fundamentally different. Hardware is just another way 
to implement programs, and it uses other languages (but SystemC even looks 
quite close to C). If there is a boundary, it's way below the distinction 
between a Tivo and a PC, because these two basically consist of a 
processor, some RAM, some flash, a harddisk, and a video driver.

What's true: We don't have the moral power to define *where* the software 
goes, but we have the moral power to define *how* users can change the 
software when they own the hardware (the physical representation).

-- 
Bernd Paysan
"If you want it done right, you have to do it yourself"
http://www.jwdt.com/~paysan/

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Ingo Molnar]

* Bernd Paysan <bernd.paysan@gmx.de> wrote:

> On Thursday 14 June 2007 12:38, Ingo Molnar wrote:
> > that's fine, but the fundamental question is: where is the moral
> > boundary of the power that the copyright license gives? The FSF seems to
> > believe "nowhere, anything that copyright law allows us to achieve our
> > goals is a fair game" - and the GPLv3 shows that belief. I dont
> > subscribe to that view. I think the proper limit is the boundary where
> > the limit of the software is - because that's the only sane and globally
> > workable way to stop the power-hungry. I.e. the information we produce
> > is covered by the rules of the GPL. It might be used in ways
> > inconvenient to us, it might be put on hardware we dont like (be that a
> > Tivo, a landmine or an abortion instrument) but that does not change the
> > fundamental fact: it's outside the _moral scope_ of our power.
> 
> Where is the boundary between hard- and software? [...]

this is largely irrelevant to my argument: the FSF is clearly trying to 
extend the scope of the GPL to restrict the distribution of certain 
hardware+software combinations. The FSF is not really arguing that the 
boundary between software and hardware is diffuse. (which btw. it 
clearly is) The FSF simply wants to be able to say via the GPLv3: "to be 
able to distribute GPL-ed software, the hardware is required to do this 
and this".

please note an important thing here: "required to do this and this" is 
the _precise antithesis_ of "freedom". The only significant restriction 
on freedom the GPLv2 allows is that the covered work (the software) is 
not to be restricted. And that is a fair deal. Even if any additional 
restrictions would otherwise be for the "common good" and would further 
"freedom" in creative ways.

If the FSF's argument and approach was correct then it would be fine to 
add these restrictions to GPLv4:

 - do not distribute non-GPL-compatible software with GPL software on 
   the same hardware.

 - send at least 10 free samples of the hardware to the FSF 
   headquarters. (after all true freedom is only achieved if developers 
   are not only allowed to modify the hardware, but are allowed to test 
   it as well, for the freedom of the community.)

 - donate $10 to the FSF.

 - spray "Linus sucks because he stole RMS's GNU thunder in the 90s and
   never gave it back!" graffiti on 3 separate walls in your
   neighborhood.

Each of these items is an additional restriction on either the 
hardware+software combination that is being distributed or on the person 
who does the distribution, and each of these items - some abstractly, 
some more directly - advance a notion of the "four GNU freedoms" in some 
way. And each of these items has a basis in copyright law and might be 
legally put into a license and might be enforceable. (ok, probably not 
the last item ;)

think about it, the list of things that one can do via license to 
"achieve more freedom" just doesnt stop! My point is: it has to stop at 
the only boundary that makes sense, and which boundary is clearly 
spelled out in the spirit and in the letter of the GPLv2: "our work is 
our work, your work is your work".

Any additional restriction to "help achieve more freedom" just puts us 
into divisive political and moralistic games that will only fracture us, 
that will eventually erode the value of our software and hence makes any 
'power' we have over that software meaningless in the real world. In the 
end no-one but the Microsofts of this world will win.

	Ingo

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alan Milnes]

On 14/06/07, Ingo Molnar <mingo@elte.hu> wrote:

>    My point is: it has to stop at the only boundary that makes sense, and which boundary
> is clearly spelled out in the spirit and in the letter of the GPLv2: "our work is our work, your
> work is your work".

Agreed - if you want to take my work you are welcome as long as you
contribute back your changes.  That's the deal that GPL2 enforces and
why it has been so successful. GPL3 is a very different beast with a
much wider agenda, which makes it far more difficult to achieve
consensus on what it should contain.

Personally I would have liked to seen a GPL2.x which fixes some of the
issues but stays true to the more limited objective.

Alan

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bernd Paysan]

[-- Attachment #1: Type: text/plain, Size: 1549 bytes --]

On Thursday 14 June 2007 16:08, Alan Milnes wrote:
> Agreed - if you want to take my work you are welcome as long as you
> contribute back your changes.  That's the deal that GPL2 enforces and
> why it has been so successful.

That may be a side effect of the GPL, but it's actually not how the GPLv2 
works (nor is it the intention). "Contribute back" means upstream. There's 
no such provision in the GPLv2, you contribute only downstream. And there 
are cases where you don't need to contribute at all.

E.g. the kernel hacking I'm doing at the moment: I have bought a uClinux 
blackfin board, for testing my digital audio amplifier. For that, I took 
one of the blackfin alsa audio drivers, and changed it so that it could 
talk to my digital audio amplifier. I'm not distributing this software, 
it's a complete in-house project, so I'm not obliged to contribute back. At 
the moment, I'm the only person in the world who has both access to the 
digital audio amplifier and the blackfin board, so releasing this driver in 
that early stage is a rather pointless excercise.

I think this above explains fairly well the "misunderstandings" that are 
appearing here. The GPL is not reflective (tit-for-tat), it's transient. If 
there's a loop in the transient propagation, it becomes reflective through 
the loop, but not by itself. This was the case in GPLv1, is the case in 
GPLv2, and will be the case in GPLv3.

-- 
Bernd Paysan
"If you want it done right, you have to do it yourself"
http://www.jwdt.com/~paysan/

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Rob Landley]

On Thursday 14 June 2007 11:44:07 Bernd Paysan wrote:
> On Thursday 14 June 2007 16:08, Alan Milnes wrote:
> > Agreed - if you want to take my work you are welcome as long as you
> > contribute back your changes.  That's the deal that GPL2 enforces and
> > why it has been so successful.
>
> That may be a side effect of the GPL, but it's actually not how the GPLv2
> works (nor is it the intention). "Contribute back" means upstream. There's
> no such provision in the GPLv2, you contribute only downstream. And there
> are cases where you don't need to contribute at all.

And the Linux kernel community has been familiar with this situation all 
along.  It's the bargain the kernel developers struck with each other a 
decade and a half ago.

Now the FSF is coming along and being Darth Vader: "I am altering the bargain.  
Pray I don't alter it any further."

> I think this above explains fairly well the "misunderstandings" that are
> appearing here. The GPL is not reflective (tit-for-tat), it's transient. If
> there's a loop in the transient propagation, it becomes reflective through
> the loop, but not by itself. This was the case in GPLv1, is the case in
> GPLv2, and will be the case in GPLv3.

That's not specifically a limitation of the GPL, that's a limitation of 
copyright law which forms the basis of the GPL.  It covers distribution, not 
usage.

GPLv2 eliminates the case where I have a modified binary I contributed to, but 
can't see the source code of those modifications.  This has the pragmatic 
effect of greatly reducing forking in a project, such as the Emacs/Lucid 
Emacs fork that inspired the "Emacs license" that became GPLv1.

Rob

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Rob Landley <rob@landley.net> wrote:

> Now the FSF is coming along and being Darth Vader: "I am altering
> the bargain.  Pray I don't alter it any further."

1) it can't possibly do that.  the Linux license is something that
only the Linux developers can decide.

2) I don't know how the FSF is approaching the Linux developers, but
what I've been personally trying to do in this infinite thread was
mainly to set the record straight that v3 did not change the spirit of
the license, like some have claimed.

3) Another thing I've tried to do was to try to figure out why Linux
developers seem to consider v2 better than v3 for their own goals.  I
must admit I failed.  The presented reasons don't seem to distinguish
v2 from v3 to me, or rather make v3 sound better.

It's disappointing that I took so much of everyone's time without
achieving any of my goals.  I hope it was at least useful or
enlightening to some.

I'll now try to step out of the discussion, but I guess I'm just as
addicted to flames.  I don't see that it's getting anywhere, and I
don't particularly enjoy the name calling.  And then, I was politely
invited to go away...

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Thu, 14 Jun 2007, Alexandre Oliva wrote:
> 
> It's disappointing that I took so much of everyone's time without
> achieving any of my goals.

What do you expect, when you tried to entertain a legal picture of the 
GPLv2 that even the FSF counsel doesn't believe in?

I will state one more time: I think that what Tivo did was and is:

 (a) perfectly legal wrt the GPLv2 (and I have shown multiple times why 
     your arguments don't hold logical water - if you actually followed 
     them yourself, you wouldn't be using a redhat.com email address!)

 (b) not just legally right, but perfectly morally right too (it wasn't 
     some underhanded "trick" thing - it was following the spirit _and_ 
     the letter of the law)

 (c) the only reasonable thing many companies *can* do in the face of laws 
     and regulations and entities like the RIAA/MPAA.

and you should admit that the fact that the FSF counsel says that it 
couldn't sue Tivo in the US, it means that while my standpoint may not be 
the _only_ possible one, I'm certainly not "confused" about (a) above.

The (b) and (c) points are not "legal" points, they are about the fact 
that quite often, morality and practicality are independent of legality, 
and you should never see law as being the *only* thing that matters. So 
the reason I bring them up is that it wasn't just "legally ok", they also 
had good *reasons* for doing it, and there was no hanky-panky about it!

In fact, I consider Tivo one of the good guys, because they were one of 
the few people that had things like the GPLv2 actually printed out and 
clearly stated IN THE MAIN PAPER MANUAL. In the very first version of 
their box. Without anybody twisting any arms at all. 

IOW, Tivo really did everything right. I personally think that they were 
even classy about it.

And that's my opinion. THINK about that for a moment. THINK about the fact 
that I am the original copyright holder in the main software project they 
used, and that I state that as neither having ever gotten paid _or_ owning 
any stock what-so-ever in Tivo. 

Dammit, if I cannot say that I think what they did was fine, who can?

So pause there for a moment, and really *think* about the above. Stop 
seeing Tivo as "the devil". 

[ Wait a few seconds here, thinking! ]

Now, we both agree that GPLv3 would change the situation wrt Tivo, don't 
we?

[ Wait a few more seconds here, thinking about what that means, taking the 
  above into consideration ]

..so given that I think that what Tivo did is *fine*, the GPLv3 "solution" 
is not a solution at all, is it?

Quite the reverse. It's a unnecessary restriction that doesn't actually 
solve anything at all, it just adds problems of its own.

And yet you claim that you cannot understand why I (and others) would 
consider the GPLv3 to be a "worse" license. It is *obviously* worse to 
anybody who thought that "Tivoization" wasn't a problem in the first 
place!

.. but I guess you'll ignore that argument, the way you ignored all the 
other ones too, and continue to blame your lack of understanding on me 
being "confused" about the issue.

			Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:

> On Thu, 14 Jun 2007, Alexandre Oliva wrote:
>> 
>> It's disappointing that I took so much of everyone's time without
>> achieving any of my goals.

> What do you expect, when you tried to entertain a legal picture of the 
> GPLv2 that even the FSF counsel doesn't believe in?

I don't think I made significant legal arguments.  My points were
about the spirit of the license.  That's not legal at all.  That's
moral and ethical background.

Your aggressive response directed at the FSF came as quite a surprise
to me, given the way I got into this conversation.

> I will state one more time: I think that what Tivo did was and is:

>  (a) perfectly legal wrt the GPLv2

I respectfully disagree, and I know I'm not alone in this assessment.
I know other kernel developers agree with it.  And they're as entitled
to claim failure to comply with the license as you are.

>      (and I have shown multiple times why 
>      your arguments don't hold logical water

You countered one of the various arguments I have, and you failed at
that.  It was another, quite different argument, that got me to
realize it didn't work legally.  But this says nothing about
compliance with the spirit of the license.

>  (b) not just legally right, but perfectly morally right too

I guess we'll have to agree to disagree on this one.

>  (c) the only reasonable thing many companies *can* do in the face
>  of laws and regulations and entities like the RIAA/MPAA.

No, TiVO could just throw the key away.  Why doesn't it?

> and you should admit that the fact that the FSF counsel says that it 
> couldn't sue Tivo in the US, it means that while my standpoint may not be 
> the _only_ possible one, I'm certainly not "confused" about (a) above.

I don't think I've ever claimed you were confused about (a).  I said
we disagreed.  That's quite different.

You were confused between the legalese and the spirit.

> The (b) and (c) points are not "legal" points, they are about the fact 
> that quite often, morality and practicality are independent of legality, 
> and you should never see law as being the *only* thing that matters.

Aah, now we're getting somewhere!

> Dammit, if I cannot say that I think what they did was fine, who can?

If you and all other Linux copyright holders agreed about it, sure.

Just like you could all grant it an additional permission, just so
that they feel safe about it.

> Now, we both agree that GPLv3 would change the situation wrt Tivo, don't 
> we?

Yes.  They'd have to give up the ability to update the software, or
pass it on to the user.  If they can't do the latter, they could still
do the former.  How bad would this be for them, do you know?

> And yet you claim that you cannot understand why I (and others) would 
> consider the GPLv3 to be a "worse" license.

That's because when you talk about why GPLv2 is better, you always
talked about virtues that are just as present in v3, and that AFAICT
would in fact by increased by v3.

> .. but I guess you'll ignore that argument, the way you ignored all the 
> other ones too, and continue to blame your lack of understanding on me 
> being "confused" about the issue.

I'm sorry if I come off that way.  But we really look at this issue
from very different perspectives, and it's difficult for me to try to
see it your way.  Please cut me some slack here.  I'm trying hard, but
there is so much noise and so many hard feelings that seeing what the
real issues are is not that easy.

Thanks for your understanding,

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Fri, 15 Jun 2007, Alexandre Oliva wrote:
> 
> Yes.  They'd have to give up the ability to update the software, or
> pass it on to the user.  If they can't do the latter, they could still
> do the former.  How bad would this be for them, do you know?

In other words, you advocate license for technical programs that causes 
people to make bad technical choices?

Yeah, that's real smart. That's a sign of true intellect, isn't it?

Wrong.

		Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 15, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:

> On Fri, 15 Jun 2007, Alexandre Oliva wrote:
>> 
>> Yes.  They'd have to give up the ability to update the software, or
>> pass it on to the user.  If they can't do the latter, they could still
>> do the former.  How bad would this be for them, do you know?

> In other words, you advocate license for technical programs that causes 
> people to make bad technical choices?

I do place ethical issues over technical ones, if that's what you're
asking.


And then, why should the vendor have any say on the software that runs
on the hardware I purchased from them, after the purchase?

Heck, I'd feel *safer* if I knew they couldn't modify the code in my
box without my permission.  Do you support WGA or the Sony Rootkit?
How is this any different?

But if they want to keep the ability to change the software in my box,
I want that for myself as well.  If for no other reason, in case they
mess things up.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bernd Paysan]

[-- Attachment #1: Type: text/plain, Size: 2114 bytes --]

On Friday 15 June 2007 04:19, Linus Torvalds wrote:
> I will state one more time: I think that what Tivo did was and is:
>
>  (a) perfectly legal wrt the GPLv2 (and I have shown multiple times why
>      your arguments don't hold logical water - if you actually followed
>      them yourself, you wouldn't be using a redhat.com email address!)

Linus, Harald Welte managed to get the keys from Siemens, which sold 
a "tivoized" Linux router. As typical for German courts, he got the key in 
a settlement, so there's no citable court verdict (totally, he AFAIK got 
two court verdicts up to now). But the way settlements work in Germany 
suggests that it's likely that the court would have decided that way*.

So there is no verdict on this question, but a strong hint that you are 
wrong, at least under German law (which is not the best money can buy ;-). 
And as Tivo doesn't sell their crap into Germany, we can't test it with 
them.

It's you who think the GPLv2 is tit-for-tat, not the FSF, and it's not in 
the text of the GPLv2; this is your private misinterpretation. The GPLv2 
is "transitive", i.e. everybody receives the same rights, nobody is 
entitled to take rights from others (the only person to choose is the 
author), and this certainly includes technical means (even if it does not 
explicitely say so), and the GPLv3 is just the same (it's just more 
readable). All arguments from your side put up are straw men like "hardware 
vs. software" and such.

BTW: Hardware as it is done today (chips, printed circuit boards, etc.) is 
copyrighted as well, as it is much cheaper to copy hardware than to develop 
it.

*) German judges want to resolve civil cases by settlements. They let the 
parties pass a few arguments, and indicate which way the judge possibly 
would decide to help them settle fast. Settlement is way cheaper than a 
court verdict, so most sane people choose to settle (unfortunately it's 
often the insane people who go to court ;-).

-- 
Bernd Paysan
"If you want it done right, you have to do it yourself"
http://www.jwdt.com/~paysan/

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alan Cox]

> 2) I don't know how the FSF is approaching the Linux developers, but
> what I've been personally trying to do in this infinite thread was
> mainly to set the record straight that v3 did not change the spirit of
> the license, like some have claimed.

The FSF have certainly tried to talk to me a bit about it - mostly about
some product called GNU/Linux which I had to tell them I'd never heard of
and wasn't involved in ;)

> 
> 3) Another thing I've tried to do was to try to figure out why Linux
> developers seem to consider v2 better than v3 for their own goals.  I
> must admit I failed.  The presented reasons don't seem to distinguish
> v2 from v3 to me, or rather make v3 sound better.

At least one important one I think is this:

A large number of people contributed to the GPLv2 kernel. They did so on
the basis there was an agreement about how the result could and would be
used. The GPLv3 changes that agreement, whether for good or bad depends
on who you are and what you do.

What right does Linus or anyone else have to change the rules and
unavoidably harm some of the people who contributed on the basis of the
previous licence. Any community project is built around a set of
expectations and beliefs encoded in culture, licences, documents and so
on.

The kernel community was built around GPLv2. A large number of the people
involved did so for pragmatic not FSF reasons and are not part of FSF
culture. The fact that community isn't interested in GPL3 should not be a
suprise, nor should it be seen as it seems you see it to be a failure of
the GPL3. 

GPLv2 is how we've done it, it has been for fourteen years and numerous
people have contributed on that basis. Should we kick some of them out of
that community because a third party says "new license good". What
matters more to the project itself - respect for those who work on it and
their beliefs or an FSF attempt to strengthen free software protection ?
Thats an "ends and means" type question but I think it explains the
fundamental question very well.

Alan

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 15, 2007, Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:

>> 3) Another thing I've tried to do was to try to figure out why Linux
>> developers seem to consider v2 better than v3 for their own goals.  I
>> must admit I failed.  The presented reasons don't seem to distinguish
>> v2 from v3 to me, or rather make v3 sound better.

> What right does Linus or anyone else have to change the rules

None.  It would have to be an agreement between all parties involved.
A difficult one, everyone knows.

But see, this is a distraction.  It doesn't even begin to address the
relevant (to me) question: why Linux developers seem to consider v2
better than v3 for their own goals.

I can appreciate the difficulties that there would be for switching
from v2 to v3.  This in itself might be a reason to not even try to
switch to another license, no matter how much better it could possibly
be.

But it doesn't give any hint whatsoever as to why v2 is better than
v3.  In fact, it simply avoids addressing that point.


Now, of course, each individual contributor may have different reasons
to be part of the Linux community, and each individual contributor may
have chosen v2 or v2+ or any other v2-compatible set of licensing
terms for different reasons.

I'd very much like to hear (err read), from those who think v2 serves
their reasons to contribute to Linux better than v3, why that is so.

Thanks,

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, "Alan Milnes" <alan@itfoundation.org.uk> wrote:

> Agreed - if you want to take my work you are welcome as long as you
> contribute back your changes.  That's the deal that GPL2 enforces and
> why it has been so successful.

Where did you get this impression that GPLv2 enforces this deal?

It doesn't, and this is *exactly* why I dispute the claim that GPLv2
is tit-for-tat.

> GPL3 is a very different beast with a much wider agenda,

The agenda is *precisely* the same: ensure that all users are free to
modify and share the licensed software.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Michael Poole]

Ingo Molnar writes:

> this is largely irrelevant to my argument: the FSF is clearly trying to 
> extend the scope of the GPL to restrict the distribution of certain 
> hardware+software combinations. The FSF is not really arguing that the 
> boundary between software and hardware is diffuse. (which btw. it 
> clearly is) The FSF simply wants to be able to say via the GPLv3: "to be 
> able to distribute GPL-ed software, the hardware is required to do this 
> and this".

Most people arguing for the expansive interpretation do not really
care what hardware is combined with what software.  They care about
the ability for the user (in the GPLv2's terms, someone who receives
GPL'ed software) to have comparable ability to modify and
(re-)distribute the software as the software distributor does.  The
issue of GPLed software on DRMed hardware applies equally to digital
video recorders, where the hardware and software distributor are
usually the same, and video game consoles, where they are not.

There is no good reason to treat a "GPL-incompatible" hardware
platform (for example, incompatible due to restrictions on the keys to
generate digital signatures) differently than a "GPL-incompatible"
patent area.  If a software distributor cannot simultaneously comply
with the GPL and his other obligations, he should either not
distribute the software or be prepared to face the liability from
breaching his obligations.

Michael Poole

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Rob Landley]

On Thursday 14 June 2007 07:27:59 Bernd Paysan wrote:
> Where is the boundary between hard- and software?

Software's the bit that's infinitely replicable at zero cost.  Hardware tends 
not to be.

Rob

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bernd Paysan]

[-- Attachment #1: Type: text/plain, Size: 1433 bytes --]

On Friday 15 June 2007 01:08, Rob Landley wrote:
> On Thursday 14 June 2007 07:27:59 Bernd Paysan wrote:
> > Where is the boundary between hard- and software?
>
> Software's the bit that's infinitely replicable at zero cost.  Hardware
> tends not to be.

There's no "zero cost" for software replication, either. You have to pay for 
your media, even if today's price for a GB harddisk space is just 20 
Euro-cents. You have to pay for your bandwidth (and even if it's a 
flat-rate, the maximum amount of data you can get through is 
bandwidth*(seconds per month) for one month fee). Hardware is replicated as 
well as software, the cost for hardware replication is higher than for 
software replication, because more things are to do. The basical principle 
of producing a CD-ROM and a chip is exactly the same: lithography. 
You "print" it. A decade ago, ES2 had made chips by direct e-beam 
lithography, so the offset of the mask costs were eliminated.

With an ES2-like process, you could have your "free software CPU", where you 
design modifications yourself, send the file to the fab, and get your 
customized chip back for essentially the same price as a non-customized 
version (supposed all the tool-chain would be free software, and not 
horrible expensive Cadence/Synopsys/Mentor software).

-- 
Bernd Paysan
"If you want it done right, you have to do it yourself"
http://www.jwdt.com/~paysan/

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Ingo Molnar <mingo@elte.hu> wrote:

> I think the proper limit is the boundary where the limit of the
> software is - because that's the only sane and globally workable way
> to stop the power-hungry.

But see, I'm not talking about getting permission to hack the
hardware.  I'm only talking about getting permission to hack the Free
Software in it.

It's your position that mingles the issues and permits people to use
the hardware to deprive users of freedom over the software that
they're entitled to have.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Thu, 14 Jun 2007, Alexandre Oliva wrote:
> 
> But see, I'm not talking about getting permission to hack the
> hardware.  I'm only talking about getting permission to hack the Free
> Software in it.

You have that permission. You can hack on it all you want.

Oh, but you want to hack the hardware to accept it? That's a totally 
different issue. If so, buy a Neuros OSD box.

Really. Go to google, and type in "Neuros OSD".

Do it *now*, and then stop wasting our time. You *can* do what you want to 
do.

		Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Greg KH]

On Thu, Jun 14, 2007 at 10:46:55AM -0700, Linus Torvalds wrote:
> 
> Oh, but you want to hack the hardware to accept it? That's a totally 
> different issue. If so, buy a Neuros OSD box.
> 
> Really. Go to google, and type in "Neuros OSD".
> 
> Do it *now*, and then stop wasting our time. You *can* do what you want to 
> do.

Unfortunatly that device has some closed source kernel drivers, so you
don't really have "full" control over the system :(

thanks,

greg k-h

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Neshama Parhoti]

Forgive me for a little off-topic question but I have a difficulty
to understand a technical issue about this all.

The Linux Kernel cannot easily switch licenses because of the
large amount of people involved in it (i.e. contributed code on which
they have copyright).

But many of FSF's GNU projects are similar - for example GCC has contributors
from many many companies and individuals, from which I presume there
are who might object to GPLv3.

So how come they can so easily move to GPLv3 ?
Don't they have to have permission from all of those contributors (many
of which are Linux companies and distributors who might prefer staying
at GPLv2) ?

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Greg KH]

On Thu, Jun 14, 2007 at 09:42:16PM +0300, Neshama Parhoti wrote:
>  Forgive me for a little off-topic question but I have a difficulty
>  to understand a technical issue about this all.
> 
>  The Linux Kernel cannot easily switch licenses because of the
>  large amount of people involved in it (i.e. contributed code on which
>  they have copyright).
> 
>  But many of FSF's GNU projects are similar - for example GCC has 
>  contributors
>  from many many companies and individuals, from which I presume there
>  are who might object to GPLv3.
> 
>  So how come they can so easily move to GPLv3 ?

The FSF required copyright assignment to themselves in order to accept
the changes from the developers.  So the FSF owns the whole copyright
and can change things whenever they want, to whatever license they want.

This is the exact opposite of the kernel in which all of the original
contributors own the copyright.

Hope this helps,

greg k-h

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Greg KH <greg@kroah.com> wrote:

> The FSF required copyright assignment to themselves in order to accept
> the changes from the developers.

For many strategic projects, but not all of them.

> So the FSF owns the whole copyright and can change things whenever
> they want, to whatever license they want.

This is not true.  Have you ever read the copyright assignment
contract?  It very clearly constrains the ways the FSF can release the
code.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bill Nottingham]

Neshama Parhoti (pneshama@gmail.com) said: 
> But many of FSF's GNU projects are similar - for example GCC has 
> contributors
> from many many companies and individuals, from which I presume there
> are who might object to GPLv3.

FSF requires copyright assignment to the FSF on things like the compiler.

Bill

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alan Cox]

> So how come they can so easily move to GPLv3 ?
> Don't they have to have permission from all of those contributors (many
> of which are Linux companies and distributors who might prefer staying
> at GPLv2) ?

The FSF uses copyright assignments to ensure the entire project is under
FSF control. Linux does not - there are benefits to both approaches.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Chris Friesen]

Alexandre Oliva wrote:

> But see, I'm not talking about getting permission to hack the
> hardware.  I'm only talking about getting permission to hack the Free
> Software in it.

No you're not...you're talking about being able to hack the software 
*and load it back onto the original hardware*.

> It's your position that mingles the issues and permits people to use
> the hardware to deprive users of freedom over the software that
> they're entitled to have.

The software license controls the software.  If the hardware has 
restrictions on it that limit what software it will run, then that is 
unrelated to the software license.

There is nothing stopping you from taking the code for the tivo, 
modifying it, distributing it, or even running it on other hardware.

Suppose I had some machine that will only run microsoft-signed binaries. 
  Would it be at all related to any software license that this machine 
won't let me run linux?

Chris

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, "Chris Friesen" <cfriesen@nortel.com> wrote:

> Alexandre Oliva wrote:
>> But see, I'm not talking about getting permission to hack the
>> hardware.  I'm only talking about getting permission to hack the Free
>> Software in it.

> No you're not...you're talking about being able to hack the software
> *and load it back onto the original hardware*.

Yes.  You wouldn't impose restrictions on modifying the software like
that, now would you?  Even though the GPL says you can't impose
further restrictions on modification and distribution.

>> It's your position that mingles the issues and permits people to use
>> the hardware to deprive users of freedom over the software that
>> they're entitled to have.

> The software license controls the software.  If the hardware has
> restrictions on it that limit what software it will run, then that is
> unrelated to the software license.

As in, the license controls the software.  If a patent creates
restrictions that limit what you can do with the software, then that
is unrelated to the software license.

As in, the license controls the software.  If a discriminatory
contract limits what you can do with the software, then that is
unrelated to the software license.

As in, the license controls the software.  If I send you the source
code, but it happens to be protected by a key that only the hardware
can decode, and it won't decode for you, then that is unrelated to the
software license.

Is that so, really?

> There is nothing stopping you from taking the code for the tivo,
> modifying it, distributing it, or even running it on other hardware.

True.  But TiVO is still imposing further restrictions on how I can
modify the software stored in their device, while reserving that
ability to itself.  This is wrong.  This is not "in kind".  This is
not "tit-for-tat".  Tit-for-tat is: if they can, then I can too, and
if I can't, then they can't either.

> Suppose I had some machine that will only run microsoft-signed
> binaries. Would it be at all related to any software license that this
> machine won't let me run linux?

That would be an unfortunate machine to have, but if Linux or some
other GPLed software was not shipped in it, then I don't see how this
is relevant to this discussion.  It's not about the hardware, it's
about the software in it, and about passing on the freedoms related
with it.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Thursday 14 June 2007 15:13:31 Alexandre Oliva wrote:
> On Jun 14, 2007, "Chris Friesen" <cfriesen@nortel.com> wrote:
> > Alexandre Oliva wrote:
> >> But see, I'm not talking about getting permission to hack the
> >> hardware.  I'm only talking about getting permission to hack the Free
> >> Software in it.
> >
> > No you're not...you're talking about being able to hack the software
> > *and load it back onto the original hardware*.
>
> Yes.  You wouldn't impose restrictions on modifying the software like
> that, now would you?  Even though the GPL says you can't impose
> further restrictions on modification and distribution.

replace != modify

>
> >> It's your position that mingles the issues and permits people to use
> >> the hardware to deprive users of freedom over the software that
> >> they're entitled to have.
> >
> > The software license controls the software.  If the hardware has
> > restrictions on it that limit what software it will run, then that is
> > unrelated to the software license.
>
> As in, the license controls the software.  If a patent creates
> restrictions that limit what you can do with the software, then that
> is unrelated to the software license.

No - because this case is covered in GPLv2. Lose the straw-men.

> As in, the license controls the software.  If a discriminatory
> contract limits what you can do with the software, then that is
> unrelated to the software license.

Incorrect. This is, again, covered by the GPLv2. Straw-man argument.

> As in, the license controls the software.  If I send you the source
> code, but it happens to be protected by a key that only the hardware
> can decode, and it won't decode for you, then that is unrelated to the
> software license.

Straw-man. Situation covered by the GPLv2.

> Is that so, really?
>
> > There is nothing stopping you from taking the code for the tivo,
> > modifying it, distributing it, or even running it on other hardware.
>
> True.  But TiVO is still imposing further restrictions on how I can
> modify the software stored in their device, while reserving that
> ability to itself.  This is wrong.  This is not "in kind".  This is
> not "tit-for-tat".  Tit-for-tat is: if they can, then I can too, and
> if I can't, then they can't either.

But that right has never been guaranteed by the GPL. It might have been the 
*intent* of RMS when he wrote GPLv1 and the *intent* of the FSF when they 
wrote GPLv2, but intent is worth exactly *NOTHING* in the law *UNLESS* that 
intent is spelled out.

Anyway, as I've pointed out before: replace != modify

You can *replace* parts of a program and it will be a modification, you can 
*replace* components of a piece of Hardware and it will be a modification but 
replacing one software component of a device with another is *NOT* a 
modification. Why? Because the hardware hasn't changed at all - the hardware 
is merely there so the software can perform its job. And since you are 
*replacing* the *ENTIRE* piece of software, it isn't a modification of the 
software.

> > Suppose I had some machine that will only run microsoft-signed
> > binaries. Would it be at all related to any software license that this
> > machine won't let me run linux?
>
> That would be an unfortunate machine to have, but if Linux or some
> other GPLed software was not shipped in it, then I don't see how this
> is relevant to this discussion.  It's not about the hardware, it's
> about the software in it, and about passing on the freedoms related
> with it.

Exactly. However, you are making it about the hardware by making the claim 
that "replacing a program, in its entirety, with another is a modification". 
It isn't. A modification is when you replace or change a *portion* of a 
program. By your logic I could write an operating system that is 100% binary 
compatible with Linux and I'd be *required* to release it under the GPL, 
because, even though it *replaces* Linux, it's still a "modification".

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> On Thursday 14 June 2007 15:13:31 Alexandre Oliva wrote:
>> On Jun 14, 2007, "Chris Friesen" <cfriesen@nortel.com> wrote:
>> > Alexandre Oliva wrote:
>> >> It's your position that mingles the issues and permits people to use
>> >> the hardware to deprive users of freedom over the software that
>> >> they're entitled to have.

>> > The software license controls the software.  If the hardware has
>> > restrictions on it that limit what software it will run, then that is
>> > unrelated to the software license.

>> As in, the license controls the software.  If a patent creates
>> restrictions that limit what you can do with the software, then that
>> is unrelated to the software license.

> No - because this case is covered in GPLv2. Lose the straw-men.

It's not a straw man.  See, I was just showing that there's precedent
to ensuring that other tricks can't be used to deny users the freedoms
that the GPL is meant to defend.

By pointing out this is in the GPLv2, you acknowledge the point I
wanted to make.


So what is it that makes hardware so different that it can be used as
a trick to deny users freedoms, if other tricks can't?

> But that right has never been guaranteed by the GPL. It might have been the 
> *intent* of RMS when he wrote GPLv1 and the *intent* of the FSF when they 
> wrote GPLv2, but intent is worth exactly *NOTHING* in the law *UNLESS* that 
> intent is spelled out.

That's the different between legal terms and the spirit.  And the
promise of the GPL is to retain the spirit, to defend the freedoms.

>> That would be an unfortunate machine to have, but if Linux or some
>> other GPLed software was not shipped in it, then I don't see how this
>> is relevant to this discussion.  It's not about the hardware, it's
>> about the software in it, and about passing on the freedoms related
>> with it.

> Exactly. However, you are making it about the hardware by making the claim 
> that "replacing a program, in its entirety, with another is a modification". 
> It isn't. A modification is when you replace or change a *portion* of a 
> program. By your logic I could write an operating system that is 100% binary 
> compatible with Linux and I'd be *required* to release it under the GPL, 
> because, even though it *replaces* Linux, it's still a "modification".

I'm not sure I agree with the reasoning here, but I'm already
convinced that the argument about modification by replacement won't
fly.

But then again I ask you: why do you think TiVO is making these
hardware locks?  What do they want to cause or stop?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Jesper Juhl]

On 15/06/07, Alexandre Oliva <aoliva@redhat.com> wrote:
[snip]
>
> So what is it that makes hardware so different that it can be used as
> a trick to deny users freedoms, if other tricks can't?
>
[snip]

Why can't you understand that the GPL v2 is a *software* license, it
doesn't cover hardware at all.

If I take some GPLv2 software, modify it and then distribute it on a
CD-ROM and provide the source code as well, then I have complied with
the terms of the license.
If I take the same software, make the same modifications and
distribute the software in a ROM inside some piece of hardware, but
still allow people access to a copy of the source code used to build
whatever I put inside that ROM, then I've also complied with the
license.
In neither case can you modify the copy on the hardware (be it ROM
chip or CD-ROM), but that's not required by the license. As long as
you have access to the source code it's OK. The license says nothing
about you having to be able to update it on the hardware. The license
only says you need access to the source code.

No one is taking away your freedom to change the source or
redistribute it or whatever. The only thing locked hardware prevents
you from doing is installing modified software on that specific piece
of hardware, but that is completely outside the scope of the
*software* license.

[snip]
>
> But then again I ask you: why do you think TiVO is making these
> hardware locks?  What do they want to cause or stop?
>
I can't know for a fact what TiVO wants, but I can guess.

1) Maybe they want to prevent you installing modified software on
their hardware, then contacting them when you break it, costing them
money in customer support etc.

2) Perhaps they don't want to risk being liable if you modify the
software on their box in a way that allows you to use it as a means to
break the law.

3) Maybe they don't want you to modify the software running on their
hardware in such a way as to use the software to obtain intimate
details about their hardware that could be used by a competitor to
create a product superiour to theirs.

All quite valid reasons in my opinion.


-- 
Jesper Juhl <jesper.juhl@gmail.com>
Don't top-post  http://www.catb.org/~esr/jargon/html/T/top-post.html
Plain text mails only, please      http://www.expita.com/nomime.html

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alan Cox]

> Why can't you understand that the GPL v2 is a *software* license, it
> doesn't cover hardware at all.

The GPLv2 is a copyright license not a software licence, indeed there is
no such thing as a 'software licence'. It deals with the circumstances
and manner in which you are permitted (by the author) to make copies of
their work, to modify their work and in some cases to perform their work
(plus other sundry rights). Copyright law doesn't care whether the object
in question is as abstract as computer source code (providing it has been
'fixated' in some form) or a two hundred foot high art installation - or
a combination of the two.

So irrespective of the whole pointless debate going on you are trying to
draw lines that don't exist in the first place.

> I can't know for a fact what TiVO wants, but I can guess.

You could also do your research.

> All quite valid reasons in my opinion.

and all wrong.

Look up the owning and controlling interests in Tivo and you'll find the
correct reason - stopping you doing evil things like keeping movies
you've recorded or uploading them to the internet [which ironically of
course is the entire effect of the whole 'convergence' thing]

Alan

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Jesper Juhl]

On 15/06/07, Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:
> > Why can't you understand that the GPL v2 is a *software* license, it
> > doesn't cover hardware at all.
>
> The GPLv2 is a copyright license not a software licence, indeed there is
> no such thing as a 'software licence'. It deals with the circumstances
> and manner in which you are permitted (by the author) to make copies of
> their work, to modify their work and in some cases to perform their work
> (plus other sundry rights). Copyright law doesn't care whether the object
> in question is as abstract as computer source code (providing it has been
> 'fixated' in some form) or a two hundred foot high art installation - or
> a combination of the two.
>

Right. My bad.

> So irrespective of the whole pointless debate going on you are trying to
> draw lines that don't exist in the first place.
>
> > I can't know for a fact what TiVO wants, but I can guess.
>
> You could also do your research.
>
I have absolutely no idea where to go look something like that up :-(

> > All quite valid reasons in my opinion.
>
> and all wrong.
>
> Look up the owning and controlling interests in Tivo and you'll find the
> correct reason - stopping you doing evil things like keeping movies
> you've recorded or uploading them to the internet [which ironically of
> course is the entire effect of the whole 'convergence' thing]
>
Hmm, wouldn't that be my guess nr. 2?  A way to use the hardware to
break the law...


Anway, the whole point of my post was mainly to /try/ and say that the
GPL gives you a right to obtain source code for modifications, but it
doesn't say anything about being able to run a compiled version of
that source on any specific hardware.


-- 
Jesper Juhl <jesper.juhl@gmail.com>
Don't top-post  http://www.catb.org/~esr/jargon/html/T/top-post.html
Plain text mails only, please      http://www.expita.com/nomime.html

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Friday 15 June 2007 02:59:31 Jesper Juhl wrote:
<snip>
> > > All quite valid reasons in my opinion.
> >
> > and all wrong.
> >
> > Look up the owning and controlling interests in Tivo and you'll find the
> > correct reason - stopping you doing evil things like keeping movies
> > you've recorded or uploading them to the internet [which ironically of
> > course is the entire effect of the whole 'convergence' thing]
>
> Hmm, wouldn't that be my guess nr. 2?  A way to use the hardware to
> break the law...
>
>
> Anway, the whole point of my post was mainly to /try/ and say that the
> GPL gives you a right to obtain source code for modifications, but it
> doesn't say anything about being able to run a compiled version of
> that source on any specific hardware.

And you are correct. It is also clear, thanks to language directly in the 
GPLv2 itself, that there is no "intent" of the license to cover that 
situation.

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 15, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> On Friday 15 June 2007 02:59:31 Jesper Juhl wrote:

>> it doesn't say anything about being able to run a compiled version
>> of that source on any specific hardware.

> And you are correct. It is also clear, thanks to language directly
> in the GPLv2 itself, that there is no "intent" of the license to
> cover that situation.

You're again confusing legal terms with the intent.  The legal terms
provide an indication of the intent, but the preamble, along with the
free software definition it alludes to, do an even better job at that.

That said, the letter of the GPL explicitly says that the act of
running the program is not restricted, and that this is outside the
scope of the license.  It's a copyright license, and per US law,
running software is not regulated by copyright; this is not so
elsewhere, and local interpretation indicates that the intent of the
license is indeed to grant unlimited permission to run the program and
modified versions thereof, based on the free software definition.


But then, when someone says "I won't let you run modified versions of
this software on this hardware I'm selling you", is this not a further
restriction on the exercise of the rights granted in the license?

And, per the spirit, if the manufacturer can still install and run
modified versions of the software on that hardware, is it not failing
to comply with the spirit of passing on all the rights that you have?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Friday 15 June 2007 15:37:04 Alexandre Oliva wrote:
> On Jun 15, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> > On Friday 15 June 2007 02:59:31 Jesper Juhl wrote:
> >> it doesn't say anything about being able to run a compiled version
> >> of that source on any specific hardware.
> >
> > And you are correct. It is also clear, thanks to language directly
> > in the GPLv2 itself, that there is no "intent" of the license to
> > cover that situation.
>
> You're again confusing legal terms with the intent.  The legal terms
> provide an indication of the intent, but the preamble, along with the
> free software definition it alludes to, do an even better job at that.

And the preamble, not being part of the active portion of the license, has 
absolutely *ZERO* bearing. Just as it is not the *intent* of RMS, the FSF or 
*ANY* person (or legal entity) that had a hand in crafting the GPLv2 or GPLv3 
which is looked at when determining the "intent" of the license. It is the 
intent of the person and/or "legal entity" that has placed their work under 
said license.

What is so hard to understand about that ?

> That said, the letter of the GPL explicitly says that the act of
> running the program is not restricted, and that this is outside the
> scope of the license.  It's a copyright license, and per US law,
> running software is not regulated by copyright; this is not so
> elsewhere, and local interpretation indicates that the intent of the
> license is indeed to grant unlimited permission to run the program and
> modified versions thereof, based on the free software definition.

And that may be what courts in Brazil believe, independent of the fact that 
the license itself *intentionally* limits itself to "copying, distribution 
and modification". That legal decision, in fact, may not have been motivated 
by the actual belief that that was the intent of the license - it could (and, 
from looking at the situation and available facts, might actually have been) 
motivated by political reasons.

>
> But then, when someone says "I won't let you run modified versions of
> this software on this hardware I'm selling you", is this not a further
> restriction on the exercise of the rights granted in the license?

In Brazil, because the courts there have rendered a judgment that the license 
requires the unlimited running of the covered work. But that is *unfairly* 
applying a license on a piece of software to the hardware on which it runs. 
Based on your logic the hardware manufacturer would have to enable people to 
run code compiled for an entirely different processor. Not that it matters in 
the least.

> And, per the spirit, if the manufacturer can still install and run
> modified versions of the software on that hardware, is it not failing
> to comply with the spirit of passing on all the rights that you have?

Not in the least. They have the rights to "copy, modify and distribute" 
the "source code for a work". That is *EXACTLY* the set of rights they have 
to the code, and it is *EXACTLY* the set of rights they pass on. The GPL does 
not apply to any *BINARY* form of the work, except for the fact that you are 
required to provide the source code that was used to generate the binary.

The GPL *clearly* defines "source code" as:
"the preferred form of the work for making modifications to it"
It goes on to state:
"For an executable work, complete source code means all the source code for 
all modules it contains, plus any associated interface definition files, plus 
the scripts used to control compilation and installation of the executable."

(Note that, since the "signing" of the TiVO kernel is part of the installation 
they *should* be including the script that does the signing. However, since 
the SHA1 key that is part of the signing process is *not* a "script" (even in 
the loosest possible definition of the term) they do not have to provide it.)

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 15, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> On Friday 15 June 2007 15:37:04 Alexandre Oliva wrote:
>> On Jun 15, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
>> > On Friday 15 June 2007 02:59:31 Jesper Juhl wrote:
>> >> it doesn't say anything about being able to run a compiled version
>> >> of that source on any specific hardware.
>> >
>> > And you are correct. It is also clear, thanks to language directly
>> > in the GPLv2 itself, that there is no "intent" of the license to
>> > cover that situation.
>> 
>> You're again confusing legal terms with the intent.  The legal terms
>> provide an indication of the intent, but the preamble, along with the
>> free software definition it alludes to, do an even better job at that.

> And the preamble, not being part of the active portion of the license, has 
> absolutely *ZERO* bearing. Just as it is not the *intent* of RMS, the FSF or 
> *ANY* person (or legal entity) that had a hand in crafting the GPLv2 or GPLv3 
> which is looked at when determining the "intent" of the license. It is the 
> intent of the person and/or "legal entity" that has placed their work under 
> said license.

No disagreement.  You keep forgetting that I'm not here to say what
Linux licensing means or doesn't mean.

I'm here to point out that GPLv3 does not break the spirit of the GPL.

> What is so hard to understand about that ?

/echo

> (Note that, since the "signing" of the TiVO kernel is part of the
> installation they *should* be including the script that does the
> signing. However, since the SHA1 key that is part of the signing
> process is *not* a "script" (even in the loosest possible definition
> of the term) they do not have to provide it.)

It's not build script, it's just regular source code, indeed.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Ingo Molnar]

* Alexandre Oliva <aoliva@redhat.com> wrote:

> >> You're again confusing legal terms with the intent.  The legal 
> >> terms provide an indication of the intent, but the preamble, along 
> >> with the free software definition it alludes to, do an even better 
> >> job at that.
> 
> > And the preamble, not being part of the active portion of the 
> > license, has absolutely *ZERO* bearing. Just as it is not the 
> > *intent* of RMS, the FSF or *ANY* person (or legal entity) that had 
> > a hand in crafting the GPLv2 or GPLv3 which is looked at when 
> > determining the "intent" of the license. It is the intent of the 
> > person and/or "legal entity" that has placed their work under said 
> > license.
> 
> No disagreement.  You keep forgetting that I'm not here to say what 
> Linux licensing means or doesn't mean.

it is _you_ forgetting to read what you wrote just 1 mail ago above. 
_Read_ it:

  "The legal terms provide an indication of the intent, but the 
   preamble, along with the free software definition it alludes to, do 
   an even better job at that.".

Your point was totally bogus, and you have been pointed out that your 
point was bogus. And your answer - instead of admitting that you were 
wrong once again (i'm not even asking you to apologize for wasting our 
time) - to pretend that there is "no disagreement" and to patronize your 
discussion partner with a "you keep forgetting ..." phrase and a 
non-sequitor statement? How low can this discussion get? I'm truly 
amazed ...

	Ingo

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 15, 2007, Ingo Molnar <mingo@elte.hu> wrote:

> * Alexandre Oliva <aoliva@redhat.com> wrote:

>> >> You're again confusing legal terms with the intent.  The legal 
>> >> terms provide an indication of the intent, but the preamble, along 
>> >> with the free software definition it alludes to, do an even better 
>> >> job at that.

>> > And the preamble, not being part of the active portion of the 
>> > license, has absolutely *ZERO* bearing. Just as it is not the 
>> > *intent* of RMS, the FSF or *ANY* person (or legal entity) that had 
>> > a hand in crafting the GPLv2 or GPLv3 which is looked at when 
>> > determining the "intent" of the license. It is the intent of the 
>> > person and/or "legal entity" that has placed their work under said 
>> > license.

>> No disagreement.  You keep forgetting that I'm not here to say what 
>> Linux licensing means or doesn't mean.

> it is _you_ forgetting to read what you wrote just 1 mail ago above. 
> _Read_ it:

>   "The legal terms provide an indication of the intent, but the 
>    preamble, along with the free software definition it alludes to, do 
>    an even better job at that.".

You still don't seem to get the difference between spirit and letter,
and the difference between author of the license and licensor of the
software (I guess I wasn't clear about this).

When I talk about the spirit of the GPL, I'm talking about how its
authors designed it, how they described their motivations in the
preamble, in the Free Software Definition, in speeches explaining it,
etc.

When I talked about the meaning of the Linux licensing above, I'm
talking about the intent of the Linux authors when choosing the GPLv2.

They aren't necessarily the same.

Some Linux authors may have read the preamble and understood something
else.  Some may have simply skipped it, and focused in the legal
terms.  Some even understood the legal terms in different senses.  But
they have all agreed to license their code under GPLv2, whatever the
motivations each one of them had, and none of them has to match the
spirit of the GPL.


Now, the spirit of the GPL, the intent behind its design, is something
that may be entirely different.  And when I say that GPLv3 didn't
change the spirit of the GPL, I'm saying that from the perspective of
someone who understands very deeply the philosphy and motivations
behind it.


Please understand that these are two separate issues.  That the spirit
doesn't change, and that you don't share that spirit, or that you
didn't think that was the spirit, doesn't justify a claim that the
revision is changing the spirit.  It's not.  The spirit is, and has
always been, to defend users' freedoms (the 4 freedoms of the Free
Software definition), such that Free Software remains Free.


The fact that they can be different means that GPLv3 may not match the
goals you had when you chose GPLv2 for your contributions.  That's
unfortunate.  But this doesn't mean that GPLv3 is changing the spirit
of the GPL.  It's merely exposing that your goals don't match the
spirit of the GPL.  There's nothing wrong about this, GPLv2 is not a
bad license, just like the LGPLv2 is not a bad license.  They just
don't defend users' freedoms as well as GPLv3 will do.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Ingo Molnar]

* Alexandre Oliva <aoliva@redhat.com> wrote:

> On Jun 15, 2007, Ingo Molnar <mingo@elte.hu> wrote:
> 
> > * Alexandre Oliva <aoliva@redhat.com> wrote:
> >
> > > On Jun 15, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> >> >> >> It is also clear, thanks to language directly in the GPLv2 
> >> >> >> itself, that there is no "intent" of the license to cover 
> >> >> >> that situation.
> 
> >> >> You're again confusing legal terms with the intent.  The legal 
> >> >> terms provide an indication of the intent, but the preamble, along 
> >> >> with the free software definition it alludes to, do an even better 
> >> >> job at that.
> 
> >> > And the preamble, not being part of the active portion of the 
> >> > license, has absolutely *ZERO* bearing. Just as it is not the 
> >> > *intent* of RMS, the FSF or *ANY* person (or legal entity) that had 
> >> > a hand in crafting the GPLv2 or GPLv3 which is looked at when 
> >> > determining the "intent" of the license. It is the intent of the 
> >> > person and/or "legal entity" that has placed their work under said 
> >> > license.
> 
> >> No disagreement.  You keep forgetting that I'm not here to say what 
> >> Linux licensing means or doesn't mean.
> 
> > it is _you_ forgetting to read what you wrote just 1 mail ago above. 
> > _Read_ it:
> 
> >   "The legal terms provide an indication of the intent, but the 
> >    preamble, along with the free software definition it alludes to, do 
> >    an even better job at that.".

[...]

> Now, the spirit of the GPL, the intent behind its design, is something 
> that may be entirely different.  And when I say that GPLv3 didn't 
> change the spirit of the GPL, I'm saying that from the perspective of 
> someone who understands very deeply the philosphy and motivations 
> behind it.

why do you start babbling about the GPLv3 when the false statement you 
made was about the GPLv2 - as visible _very_ clearly in the first quote 
above??  Again, as a reminder, this point was presented to you (see the 
quotes above), in the discussion about whether the Tivo is fine by the 
GPLv2 or not:

' It is also clear, thanks to language directly in the GPLv2
  itself, that there is no "intent" of the license to cover that
  situation. '

to which you replied with this falsehood:

' You're again confusing legal terms with the intent.  The legal
  terms provide an indication of the intent, but the preamble, 
  along with the free software definition it alludes to, do an even 
  better job at that. '

and you have simply been pointed out that what you say is trivially 
false - the simple legal fact is that the GPLv2 does not "embedd" more 
than what is its letter and, to the lesser extent that letter may be 
ambigious, what Linus' (and other copyright holders') intent is and was. 
Not RMS's external intentions or the "free software definition" you 
mention.

how about just simply admitting that you were wrong about this, instead 
of putting up (and beating down) yet another non-sequitor strawman 
argument? Is that really that hard and embarrasing to do? I've yet to 
see a _single_ instance of you admitting in this thread that "oops, it 
seems i was really wrong about this point. Sorry.". Or are you one of 
those perfect humans who are never wrong? ;-)

	Ingo

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 17, 2007, Ingo Molnar <mingo@elte.hu> wrote:

> Again, as a reminder, this point was presented to you (see the
> quotes above), in the discussion about whether the Tivo is fine by
> the GPLv2 or not:

That's false.  I've explicitly avoided discussions on whether the
legal terms of GPLv2 permit tivoization.

What I've done was to discuss whether the tivoization was in line with
mission (since you seem to have a problem grasping the notion of
spirit of a license) of the license, because some people (yourself
included) claimed GPLv3 changed the spirit.

That others tried to steer the discussion away from this apparently
incomprehensible concept of spirit of the license, to the point of
their and your getting utterly confused and making nonsensical claims
about inconsistencies in my reasoning, moving it into the legal terms
that I was not willing to discuss, is not my fault.  My participation
here was about intent, about spirit, about mission of the GPL.

As I've already made it clear, this doesn't necessarily match intent
of people who chose GPLv2 as the license, and that's fine.  Copyright
holders know the spirit as they understood it, and that's how they
meant to license their work.

But the spirit (mission) of the GPL is the one the FSF wrote about in
the preamble.  No amount of "but the legal terms say such and such" or
"but I'm the licensor" distractions can change that.

> ' You're again confusing legal terms with the intent.  The legal
>   terms provide an indication of the intent, but the preamble, 
>   along with the free software definition it alludes to, do an even 
>   better job at that. '

> and you have simply been pointed out that what you say is trivially 
> false - the simple legal fact is that the GPLv2 does not "embedd" more 
> than what is its letter and, to the lesser extent that letter may be 
> ambigious, what Linus' (and other copyright holders') intent is and was. 
> Not RMS's external intentions or the "free software definition" you 
> mention.

Which shows you don't understand the notion of "spirit of license" (as
opposed to intent of licensing, which I AFAIK invented today to try to
dispell this confusion), and that the fact that the letter of the
license doesn't have bearing about the intent of the author of the
license, which is what the spirit of the license is about.

> how about just simply admitting that you were wrong about this,

I'm not.  You are.  Really.  Until you understand the difference
between "letter of license", "intent of licensor" and "spirit of
license", you won't be able to understand this.

> I've yet to see a _single_ instance of you admitting in this thread
> that "oops, it seems i was really wrong about this point. Sorry.".

I've done that, IIRC more than once.  Sorry to disappoint you.

> Or are you one of those perfect humans who are never wrong? ;-)

:-)  I wish :-)

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Tim Post]

On Sun, 2007-06-17 at 05:42 -0300, Alexandre Oliva wrote:

> Which shows you don't understand the notion of "spirit of license" (as
> opposed to intent of licensing, which I AFAIK invented today to try to
> dispell this confusion), and that the fact that the letter of the
> license doesn't have bearing about the intent of the author of the
> license, which is what the spirit of the license is about.
> 
> > how about just simply admitting that you were wrong about this,
> 
> I'm not.  You are.  Really.  Until you understand the difference
> between "letter of license", "intent of licensor" and "spirit of
> license", you won't be able to understand this.

I just can't stand this anymore.

Might I submit that nobody can be wrong in a hypothetical? The chances
of the kernel changing to GPL v3 are the same as fort knox being ripped
off. IM GLAD! This means I can continue to merge.

Folks I will summarize myself one last time and stop disrupting your
list. I (personally) think enough time has been taken from productivity
and I'm becoming awfully concerned about the integrity of Alexandre's
ass. He's not here to irritate you, he's here to do what he thinks is
right. I admire him for that considering the adversity.

(its not a mantra, you can keep reading, I promise!)

If I found advantage in a bug In Linux, and you fixed it, I have a
choice to not apply the patch so that the bug remains useful to me.

The FSF found needs to produce another license to serve all of the
people who have interest in the FSF. This does not mean you need to
*apply* it. Its there if you want it, but its up to you.

Just as I told Linus, This is a kernel, not *&$(*# sed, and its a
decision I'm happy was not taken lightly.

NOBODY should be influencing you with what you do with your
contributions!! Make up your own minds in your own ways or have fun in
the parrot cage where you sit in your own shit until someone else cleans
it up. Who likes that?

I am asking as strongly as one can ask, *please* put this to bed.
Continuing at this point can serve no useful purpose and everyone
involved shares a common intent to be useful! Can we agree on that
spirit?

Opinions have been expressed over, and over and over again. Everyone
knows what everyone thinks. 

We're all doing more harm than good. Who hates waste? I do.

Best,
--Tim

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Friday 15 June 2007 17:45:16 Alexandre Oliva wrote:
> On Jun 15, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> > On Friday 15 June 2007 15:37:04 Alexandre Oliva wrote:
> >> On Jun 15, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> >> > On Friday 15 June 2007 02:59:31 Jesper Juhl wrote:
> >> >> it doesn't say anything about being able to run a compiled version
> >> >> of that source on any specific hardware.
> >> >
> >> > And you are correct. It is also clear, thanks to language directly
> >> > in the GPLv2 itself, that there is no "intent" of the license to
> >> > cover that situation.
> >>
> >> You're again confusing legal terms with the intent.  The legal terms
> >> provide an indication of the intent, but the preamble, along with the
> >> free software definition it alludes to, do an even better job at that.
> >
> > And the preamble, not being part of the active portion of the license,
> > has absolutely *ZERO* bearing. Just as it is not the *intent* of RMS, the
> > FSF or *ANY* person (or legal entity) that had a hand in crafting the
> > GPLv2 or GPLv3 which is looked at when determining the "intent" of the
> > license. It is the intent of the person and/or "legal entity" that has
> > placed their work under said license.
>
> No disagreement.  You keep forgetting that I'm not here to say what
> Linux licensing means or doesn't mean.
>
> I'm here to point out that GPLv3 does not break the spirit of the GPL.

And you have been given evidence that it does. It may not break the spirit as 
you, or anyone else that believes that the FSF is the singular fountain of 
truth, see it. But it does break the spirit as a large fraction of the 
populace see it. Why do they like GPL3? Because they don't like what TiVO 
did - they see it and say "but that's not fair", so when they see that GPL3 
makes it a license violation they accept GPL3 without understanding the 
bigger picture.

To *anyone* who is considered an "Adult" in *ANY* nation and still wants to 
scream "its not fair" when something doesn't go their way or somebody does 
something they don't like I have this to say: Grow up! Life *ISN'T* fair - at 
all.

> > What is so hard to understand about that ?
>
> /echo
>
> > (Note that, since the "signing" of the TiVO kernel is part of the
> > installation they *should* be including the script that does the
> > signing. However, since the SHA1 key that is part of the signing
> > process is *not* a "script" (even in the loosest possible definition
> > of the term) they do not have to provide it.)
>
> It's not build script, it's just regular source code, indeed.

Not even. If it's anything it's "input to a program". QED: It isn't covered by 
the license. Hell, give me a week and access to a signed TiVO kernel and I'll 
produce a signed kernel that is functionally equivalent - it won't run on a 
TiVO, but I'll have replicated the "signing script" and process that TiVO 
uses. It isn't until you start extending the definition of source code in 
strange ways that the key becomes "source code". Hell, for all *anyone* knows 
(that isn't employed by TiVO) they could enter the key *manually* - ie: 
interactive input.

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Fri, 15 Jun 2007, Daniel Hazelton wrote:
> 
> And the preamble, not being part of the active portion of the license, has 
> absolutely *ZERO* bearing.

That's not true. Again, ianal, etc etc, but:

"Intent" *does* matter, and if you wrote down the intent at the time you 
entered some legal agreement, that actually also has non-zero bearing (as 
it can be used to _show_ intent more clearly than claiming fifteen years 
later "but, your honour, I _intended_ to do something else").

BUT!

Intent only matters if the contract reading itself isn't otherwise clear!

In other words, intent can never *override* the contract. But if there is 
some actual legally unclear area, the intent can be used by a 
judge/jury/arbitrator to _guide_ them to the decision. And the better 
documented that intent is, the better it is in that respect. Intent 
_after_ the fact is not very meaningful, is it?

So the preamble *does* matter legally, it just matters a lot less than the 
actual legal wording itself.

Put another way: the "intent" of the parties can be, and is, used to read 
the contract, but it's actually a mutual thing for a contract (since you 
have two or more parties in question - the intent of _one_ party doesn't 
really matter, if he cannot show that the other party agreed with that 
intent! And to get back to rms' personal history: _his_ intent, and _his_ 
personal history, and _his_ other documents don't matter for an 
agreement that doesn't include him).

So as such, a preamble can be used to show the intent of the parties, and 
thus to guide any reading.

Of course, so can "real life". If a certain reading of a contract doesn't 
make sense (and that has nothing to do with "intent"), then that obviously 
cannot be what either party really agreed to.

Oh, and documented intent can be used for laches, ie it can be a _defense_ 
against an accusation. If somebody has publicly stated some intent, he 
can't really complain later when somebody else tried to fulfill his 
intent, can he?

			Linus

PS. "Intent" can obviously have a _totally_ different legal meaning too. 
The difference between "murder" and "manslaughter" is about pre-meditation 
- ie "intent". But that's not an issue in licenses/contracts, except for a 
very specific _kind_ of "contract" - the kind the Mob puts out on people 
it wants to get rid of ;)

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Ingo Molnar]

* Linus Torvalds <torvalds@linux-foundation.org> wrote:

> > And the preamble, not being part of the active portion of the 
> > license, has absolutely *ZERO* bearing.
> 
> That's not true. Again, ianal, etc etc, but:
> 
> "Intent" *does* matter, and if you wrote down the intent at the time 
> you entered some legal agreement, that actually also has non-zero 
> bearing (as it can be used to _show_ intent more clearly than claiming 
> fifteen years later "but, your honour, I _intended_ to do something 
> else").

yeah. What comes up periodically in GPLv3 discussions as 'proof' of what 
the GPL means are totally detached statements of the FSF and of RMS, 
often written a decade _after_ the GPL has been chosen for a license of 
the Linux kernel. (the whole anti-Tivo line was invented well after the 
fact.) And those statements have little bearing on the interpretation of 
the license of GPL-ed works. (unless, of course, the author of a GPL-ed 
work agrees with those statements and intends them to be his 
interpretation of the license.)

	Ingo

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alan Cox]

> And the preamble, not being part of the active portion of the license, has 
> absolutely *ZERO* bearing. Just as it is not the *intent* of RMS, the FSF or 

Wrong (again)

The pre-amble is incredibly important as is the intent of the license
creator and even more so of the author.

When trying to solve a dispute the process starts with the legal
equivalent of banging the two parties head together in the hope they see
sense. If that fails then the legal wording is considered in detail.
Where it is ambiguous the surrounding context is considered in order to
understand the probable intent of the case.

Finally the stated intent of the author is considered in defence (The
doctrine of estoppel), and at least in UK law whether their intent was
honest (The doctrine of clean hands)

Legal disputes almost always end up about the things that are not clear
(if they were clear one side would shut up and put up) so the preambles
and statements are terribly important when this occurs, along with the
context and history.

Thus for example the fact Linus has said he believes what Tivo does is ok
means he can no longer sue Tivo for doing it. They are relying on his
promise.

Alan

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

[ Damn. I moved you to my flamers list, and then I started reading it. I'm 
  addicted to flaming. Sue me. I really do enjoy it too much. If I didn't 
  do software development, my full-time job would probably be to troll 
  various internet sites and try to set up flame wars. I'm bad, I know. 

  It's an addiction. I'm not proud. ]

On Thu, 14 Jun 2007, Alexandre Oliva wrote:
> 
> But then again I ask you: why do you think TiVO is making these
> hardware locks?  What do they want to cause or stop?

Actually, they didn't want to lock down the hardware at all. The first 
versions of the Tivo was really quite hackable - and people started 
hacking them.

They were basically forced to add lockdown by the content vendors. You can 
call them evil for "caving in", but hey, it was their whole market. They 
really had no choice. Being a company actually limits you in some ways..

If you don't want to cave in to content providers, use a regular PC and 
soemthing like MythTV. You will probably also have to use the analogue 
hole, and will have a really hard time unscrambling digital cable TV 
signals of your own, but hey, you can see it as a challenge. At least in 
places where it's not illegal.

And yes, there are bad laws in the US. But blaming Tivo for them is 
ludicrous. And the *laws* won't get fixed by software licensing either, 
quite the reverse. The GPLv3 will just make free software that uses it 
*less* relevant in that space, rather than more.

For example, I'd rather have some GPLv2'd DVD player software that does 
*not* come with a de-css key (I can get that key myself quite easily), and 
that thus gets distributed in a "useless" form, than have a GPLv3'd DVD 
player that cannot be distributed at all, because it needs the magic 
unlocking key, and distributing the css key is illegal in some countries.

Or if I was an mplayer developer (which I'm not - so I have absolutely 
*zero* say in the mplayer license - please don't take this as anythign 
like that), I'd prefer for mplayer to be GPLv2, simply because that way I 
could see my software in some high-end (legal) DVD players that actually 
complied with the insane laws that exist. Sure, to comply with the laws 
and not get sued, they might have to limit the hardware, but hey, in other 
saner places of the world (like Finland), you can use the GPLv2'd software 
legally *without* those concerns.

See? The more permissive license actually allows more people to get 
involved. And the only thing that really *matters* (the source code) can 
be distributed and improved on by all these different people, even if some 
of them may have their hands bound by legal issues.

Btw, the same is true of things like FCC rules in the US. All that is evil 
does not come from the RIAA and MPAA. It's entirely possible that a 
cellphone manufacturer would have to lock down the control logic that sets 
the power levels - and that is something that is against the license of 
the GPLv3.

So the GPLv3 actually _hinders_ people who might otherwise help the 
community from helping, by making the license so strict that those people 
(who are nice people, but have their options limited by stupid laws and 
regulations) cannot use the GPLv3.

			Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alan Cox]

> If you don't want to cave in to content providers, use a regular PC and 
> soemthing like MythTV. You will probably also have to use the analogue 
> hole

Only in the USA. In most of the world its considered quite normal that
you can plug a USB disk into your PVR, save stuff to it and then plug it
into your PC.

Alan

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:

> It's an addiction. I'm not proud.

I guess this makes it two of us :-(

> They were basically forced to add lockdown by the content vendors.

They can do that.  They will still be able to do that with v3.

All they have to do is to throw away the keys that enable themselves
to modify the code further.

> For example, I'd rather have some GPLv2'd DVD player software that does 
> *not* come with a de-css key

libdvdread and libdvdcss are separate packages.

> So the GPLv3 actually _hinders_ people who might otherwise help the 
> community from helping, by making the license so strict that those people 
> (who are nice people, but have their options limited by stupid laws and 
> regulations) cannot use the GPLv3.

Just like v2 hinders their many customers.

Are you so sure v2 is better in this regard?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Ingo Molnar]

* Alexandre Oliva <aoliva@redhat.com> wrote:

> On Jun 14, 2007, Ingo Molnar <mingo@elte.hu> wrote:
> 
> > I think the proper limit is the boundary where the limit of the 
> > software is - because that's the only sane and globally workable way 
> > to stop the power-hungry.
> 
> But see, I'm not talking about getting permission to hack the 
> hardware.  I'm only talking about getting permission to hack the Free 
> Software in it.
> 
> It's your position that mingles the issues and permits people to use 
> the hardware to deprive users of freedom over the software that 
> they're entitled to have.

where does this false sense of entitlement come from? The hardware maker
ows you nothing but what is written into the GPLv2. Not more, not less.

(In fact, most hardware makers that utilize free software today give
back _substantially more_ to the community than the license requires!
For example they are currently the largest employers of free software
developers - although nothing in the license forces them to do so. Why?
Because the economic rules that the GPLv2 creates are healthy.)

you are not "entitled" to dictate the hardware's design (or any other 
copyrighted work's design), even if the license gives you the power to 
do so. By your argument we'd have to put the following items into the 
license too:

 - free on-site training for free software developers about the 
   hardware's inner workings. (It is justified to teach free software
   the same know-how as in-house engineers of the hardware maker. 
   Without this, users are hindered in their freedom to use and 
   effectively modify (fix) the software.)

 - free access to all the hardware diagnostics tools that the hardware 
   maker has. (Without that it might be impossible to modify the 
   software as efficiently as the hardware maker's own engineers can do 
   it.)

 - free samples of the hardware to be sent to free software developers,
   upon request. (The hardware maker's own engineers have free access to 
   samples. Otherwise free software users might not get the same level 
   of driver support as the hardware maker can achieve.)

 - free access to the hardware manufacturing equipment. (If i wish to 
   modify the free software in a way that requires more RAM than the 
   hardware has, i need access to the manufacturing equipment to produce 
   a new version of the hardware that can run that free software. The 
   hardware maker has this right and flexibility to modify the software, 
   so i should have that same right too.)

see how quickly your argument becomes totally ludicrous, if brought to 
its logical conclusion?

This "right to modify" and "have the same rights as the hardware maker" 
arguments are _totally_ bogus, they were made up after the fact, just 
because quite apparently RMS had a fit over Tivo and started this verbal 
(and legal) vendetta. The FSF is now attempting to rewrite history and 
pretends that this "always was in the GPLv2" and applies this newly 
thought up concept to the GPLv3 in a way that substantially departs from 
the spirit of the GPLv2. Which spirit the GPLv2 explicitly promised to 
uphold in Section 9. Which could make any contrary section of the GPLv3 
unenforceable, when applied to "GPLv2 or later" licensed software.

	Ingo

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Al Viro]

On Thu, Jun 14, 2007 at 09:55:17PM +0200, Ingo Molnar wrote:
> This "right to modify" and "have the same rights as the hardware maker" 
> arguments are _totally_ bogus, they were made up after the fact, just 
> because quite apparently RMS had a fit over Tivo and started this verbal 
> (and legal) vendetta. The FSF is now attempting to rewrite history and 
> pretends that this "always was in the GPLv2" and applies this newly 
> thought up concept to the GPLv3 in a way that substantially departs from 
> the spirit of the GPLv2. Which spirit the GPLv2 explicitly promised to 
> uphold in Section 9. Which could make any contrary section of the GPLv3 
> unenforceable, when applied to "GPLv2 or later" licensed software.

That, BTW, is perhaps the worst problem with v2 (inherited by v3).
WTF _is_ "the spirit of the license" and who gets to decide if two
licenses are in the same spirit?  As soon as we get to "well, original
authors of the license are the final authority on that", we are
in the "I've always said ..." country.

Look, humans _suck_ at revision control, especially that of our
intentions and opinions.  It doesn't even require malice, all
ancedotes about spouses/mothers-in-law/etc. nonwithstanding.

We all easily fall into belief that we had always meant what we mean
now; that even if we said something different, it was just a poor
wording; that if we had known what we know now, we would certainly
had come to the same conclusions we have come to now.

"In the same spirit" is just about the weakest requirement in that
area.  I.e. the most prone to drift, especially when one is an ideologist
and thus has severely decayed integrity to start with.  Call it a
professional disease of crystal ball users - or a prerequisite for
playing a visionary, if you will ;-/

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Ingo Molnar]

* Al Viro <viro@ftp.linux.org.uk> wrote:

> On Thu, Jun 14, 2007 at 09:55:17PM +0200, Ingo Molnar wrote:
> > This "right to modify" and "have the same rights as the hardware maker" 
> > arguments are _totally_ bogus, they were made up after the fact, just 
> > because quite apparently RMS had a fit over Tivo and started this verbal 
> > (and legal) vendetta. The FSF is now attempting to rewrite history and 
> > pretends that this "always was in the GPLv2" and applies this newly 
> > thought up concept to the GPLv3 in a way that substantially departs from 
> > the spirit of the GPLv2. Which spirit the GPLv2 explicitly promised to 
> > uphold in Section 9. Which could make any contrary section of the GPLv3 
> > unenforceable, when applied to "GPLv2 or later" licensed software.
> 
> That, BTW, is perhaps the worst problem with v2 (inherited by v3). WTF 
> _is_ "the spirit of the license" and who gets to decide if two 
> licenses are in the same spirit? [...]

yeah. I see this as: "RMS does not want to let go of the community". 
This clause amounts to "power to relicense" _vast_ amounts of free 
software and this is by far the worst problem with the "GPLv3 process".

The GPLv3 process was pretended to be "open", but regardless of what the 
"GPL comittees" said, in the end it was one person: the president of the 
FSF (Richard Stallman) who singlehandedly decided what went into the 
GPLv3 draft and what not. For example he singlehandedly has ignored all 
the criticism that the the "Tivo" section has received.

And note how hypocritic RMS's position is here. Where is that freedom 
when it comes to the licensing process? Why does RMS have more rights 
over modifications to the license than all the other free software 
developers have? Should not he give that freedom to others too? Shouldnt 
there be a fair and just election, a vote? You know, that democracy 
thing.

And with his current attitude he affects somewhere around of 1 billion 
lines of free software. Via a license that is just a few hunded lines 
long.

I believe RMS should accept the fact that most of that code was written 
without people having bought into his ideology, and he should accept 
_responsibility_ for the power he has acquired by genius or by accident 
(your choice) and he should try to _understand_ how those people tick - 
instead of trying to further his own personal agenda.

He shouldnt say what amounts to "oh, my original intent was this and 
that, if you didnt understand it and still wrote code and used the 
default 'or later' license, it's your damn fault".

He should accept that what happened happened, after he wrote 100,000 
lines of original GNU code another ten thousand people wrote about a 
_ten thousand times more_ code. He should also accept that the "open 
source" community is about many other things, and it is alot more varied 
than his thinking is. He does not have to _like_ Tivo, but he should try 
to _understand_ them, and he should be compassionate about other 
people's right to have their own opinion and their own approaches to 
freedom. It is very clear that he has not even attempted to do that so 
far.

And the best way to start would be to significantly limit the 'same 
spirit' clause by putting in something like this:

 'in the event of a section of this license being rules unenforceable by 
  a court of law the FSF has the option to modify the license in the 
  most minimal fashion to make that section enforceable again'.

but this would mean RMS would have to give up power irreversibly. Will 
that ever happen?

	Ingo

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Ingo Molnar <mingo@elte.hu> wrote:

> * Alexandre Oliva <aoliva@redhat.com> wrote:

> you are not "entitled" to dictate the hardware's design (or any other 
> copyrighted work's design),

Agreed.

> By your argument we'd have to put the following items into the
> license too:

No, you're confusing two very different situations.

In the case of TiVO, it's getting out of its way to make sure users
can't enjoy one of the freedoms that the license says it ought to pass
on.

In the cases you mentioned, the company would have to get out of its
way to put the other parties on equal grounds.

The former is bad, it's against the spirit of the license, it's a
further restriction.

The latter would be nice to have, but it would be wrong to demand it.

You're picturing the difference between blocking the way such that you
can't get there, and actually taking you there.  What the GPL seeks is
just that you don't get in the way.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Ingo Molnar]

* Alexandre Oliva <aoliva@redhat.com> wrote:

> On Jun 14, 2007, Ingo Molnar <mingo@elte.hu> wrote:
> 
> > * Alexandre Oliva <aoliva@redhat.com> wrote:
> 
> > you are not "entitled" to dictate the hardware's design (or any 
> > other copyrighted work's design),
> 
> Agreed.

hey, that's progress. If you concede this single point then your 
arguments about the Tivo situation all fall like domino stones. Just 
watch it happen please:

> > By your argument we'd have to put the following items into the 
> > license too:
> 
> No, you're confusing two very different situations.
>
> In the case of TiVO, it's getting out of its way to make sure users 
> can't enjoy one of the freedoms that the license says it ought to pass 
> on.

the GPLv2 license says no such thing, and you seem to be mighty confused 
about how software licenses work.

the GPL applies to software. It is a software license.

the Tivo box is a piece of hardware.

a disk is put into it with software copied to it already: a bootloader, 
a Linux kernel plus a handful of applications. The free software bits 
are available for download.

the Tivo box is another (copyrighted) work, a piece of hardware.

so how can, in your opinion, the hardware that Tivo produces, "take 
away" some right that the user has to the GPL-ed software? Because they 
distribute the software and the hardware in the same package, and 
because the hardware (as _ALL_ hardware on this planet) has certain 
limitations? It was _your_ choice to buy that particular 
hardware+software combination, with whatever limitations the hardware 
has. One such limitation of the hardware might be that its color is 
butt-ugly pink. Another limitation might be that the buttons on it are 
too small for elderly people to press. A third limitation might be that 
it's not a general purpose computer and that it's not freely 
programmable by the end user. Bugger, what did you expect? Why didnt you 
buy a green PVR? Why didnt you buy a PVR with larger buttons? Why didnt 
you buy a general purpose computer? Did perhaps the Tivo look like a 
nice general purpose PC to you when you bought it?

> In the cases you mentioned, the company would have to get out of its 
> way to put the other parties on equal grounds.

how about quoting what i wrote and rebutting it specifically if you 
disagree with it, instead of writing a non-sequitor generality? You are 
involved in compiler development, so you should have the mental ability 
to follow logical arguments and you should be able to conduct a 
meaningful and objective discussion. Lets look at one of the examples i 
gave you:

> > - free access to all the hardware diagnostics tools that the 
> > hardware maker has. (Without that it might be impossible to modify 
> > the software as efficiently as the hardware maker's own engineers 
> > can do it.)

by your argument, the user has some "right to modify the software", on 
that piece of hardware it bought which had free software on it, correct? 
By your argument, the "right to modify the software" becomes meaningless 
if you cannot soft-upgrade your Tivo, if you have solder off the ROM to 
install your own ROM with a bootloader that does not do the SHA1 check, 
correct?

But by that _very same argument_, you are hindered _much more_ by not 
having proper hardware diagnostics tools and no access to hardware 
specifications. If you dont know how the hardware works, you cannot fix 
bugs in the software. So by your argument, the user has an inherent 
right to get on equal footing with the hardware manufacturer to modify 
the software on that specific hardware? There's no ifs and when. "having 
to solder off the ROM" is a "restriction on modifiability" just as much 
as "having less information about the hardware's inner workings". In 
fact, ask any kernel developer, "having to solder off the ROM" is a lot 
_smaller_ restriction than "having no information about the hardware's 
inner workings".

	Ingo

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alan Cox]

> the GPLv2 license says no such thing, and you seem to be mighty confused 
> about how software licenses work.

There is no such thing as a software licence. It is a copyright licence.

> the GPL applies to software. It is a software license.

You can GPL a new graphical logo you painted on your toilet seat, you can
GPL hardware designs. It might not be a good licence for either but it is
a valid licence.

> the Tivo box is a piece of hardware.

A Tivo box is a collection of literary works protected by copyright,
designs protected by design patents and copyright, names and logos
protected by trademarks, functionalities protected by patents and many
more things. These are the things that restrict what I may do with it and
how I may treat it. The collection of bits of metal and sand aren't
really of relevance in terms of licencing.

If it was a generic housebrick with none of these things attached then
within the law I can do what I like with it including copying it. A book
is a copyright work but the copyright is about the literary work and the
fact it is on paper is largely irrelevant. What determines your usage
rights for those pieces of paper are the literary work it carries not the
pieces of paper (unless made of a new patented paper material or similar)

> a disk is put into it with software copied to it already: a bootloader, 
> a Linux kernel plus a handful of applications. The free software bits 
> are available for download.

Except the keys - which may nor may not be required depending upon how a
court (not a mailing list) interprets the phrases

"The source code for a work means the preferred form of the work for
making modifications to it"

and

"For an executable work, complete source code means all the source code
for all modules it contains, plus any associated interface definition
files, plus the scripts used to control compilation and installation of
the executable."

If you ask the legal profession about this seriously the answer you get
is bluntly "There is no caselaw I am aware of", which means that nobody
knows. Obviously Tivo and their legal counsel have formed an opinion and
have based their actions upon that opinion.

> the Tivo box is another (copyrighted) work, a piece of hardware.

You can't copyright hardware. Sorry but if you are going to try and have
a detailed logical argument you need to start from a rigorous base point.

Alan

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Ingo Molnar]

* Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:

> > the GPLv2 license says no such thing, and you seem to be mighty 
> > confused about how software licenses work.
> 
> There is no such thing as a software licence. It is a copyright 
> licence.

a "software license" is a common shortcut for "copyright license for 
copies of software works". It's a commonly used phrase. In fact it is 
used by the FSF itself too:

   http://www.fsf.org/licensing/essays/free-sw.html

   "To decide whether a specific software license qualifies as a free 
                                 ^^^^^^^^^^^^^^^^
    software license, we judge it based on these criteria to determine 
    whether it fits their spirit as well as the precise words."

> > the GPL applies to software. It is a software license.
> 
> You can GPL a new graphical logo you painted on your toilet seat, you 
> can GPL hardware designs. It might not be a good licence for either 
> but it is a valid licence.

yeah - the GPL can be applied to most types of works recognized by 
copyright law.

> > the Tivo box is a piece of hardware.
> 
> A Tivo box is a collection of literary works protected by copyright, 
> designs protected by design patents and copyright, names and logos 
> protected by trademarks, functionalities protected by patents and many 
> more things. These are the things that restrict what I may do with it 
> and how I may treat it. The collection of bits of metal and sand 
> aren't really of relevance in terms of licencing.

If you are into technicalities then you fail to achieve that "rigorous 
base" by a wide margin. The Tivo box is not "a collection of literary 
works", it is a piece of matter, that also happens to contain fixated 
copies of literary (and other) works. The Tivo box is just one copy of 
those works - it is not "a collection of literary works". (Only if there 
was just a single Tivo box on the planet then could that box itself be 
meaningfully called a collection of works - a single and unique "master 
copy" of a work can be called the work itself.)

and that distinction, although fine, is very important. Look at GPLv2 
section 0:

" 0. This License applies to any program or other work which contains a 
  notice placed by the copyright holder saying it may be distributed 
  under the terms of this General Public License. "

the work is not the copy! The work is a more 'abstract' entity. The word 
"copyright" comes straight from that: the right to create specific 
copies of the work. And that's another reason why it's nonsensical to 
suggest that somehow the GPLv2 gives us the right to influence the 
hardware environment that the copy of the kernel got fixated into. We 
dont. ( unless that hardware environment too is a copy of a GPL-ed work 
or it is a copy of a work that is a modification of or derives from a 
GPL-ed work - but in the Tivo case it isnt. It's a collection of copies 
of works and derivation does not "jump" from the harddisk to the 
hardware. )

More down the technicalities road: the Tivo box also contains many items 
that are not copies of works protected by copyright: common types of 
screws that are not original forms of expression that are creative 
enough enough to gain copyright protection. Or numbers painted on 
various places. Or computer-originated random output. Copies of works 
that have entered the public domain and thus are not under the scope of 
copyright protection.

Neither is the Tivo box "collection of functionalities protected by 
patents", if then it is an embodiment of a method and apparatus, which 
invention is under patent protection (there are other types of patents 
as well), or which invention might not be under patent protection but 
have a patent application pending. (which might or might not issue at 
the end of the patent application process.)

> > a disk is put into it with software copied to it already: a bootloader, 
> > a Linux kernel plus a handful of applications. The free software bits 
> > are available for download.
> 
> Except the keys - which may nor may not be required depending upon how a
> court (not a mailing list) interprets the phrases
> 
> "The source code for a work means the preferred form of the work for
> making modifications to it"

i think it is clear what is intended with this section: that for example 
using automatic tools to strip out comments and obfuscating the source 
code does not fly, because what matters is the _form of the software_ 
the developer usually makes his modifications under. So this in essence 
defines the scope of the actual source code that must be made available 
so that it works on a general purpose computer, not the specific 
hardware environment under which the developer operates.

so i believe it is a ... fairly creative bending of the wording of this 
section to attempt to extend it to the hardware environment. You dont 
get my ssh keys either [*] that i use on my test-boxes, and those test 
boxes are very much part of the preferred way for me to produce kernel 
patches. But you get my kernel patches for sure! [ that is, if they dont 
crash the testboxes :-) ] Am i violating the GPLv2?

> and
> 
> "For an executable work, complete source code means all the source 
> code for all modules it contains, plus any associated interface 
> definition files, plus the scripts used to control compilation and 
> installation of the executable."

i think it is pretty clear what this section intends: not allowing 
people to become cute by stripping out of makefiles from the source 
tarball. If someone else tries to run that software on a general purpose 
computer (which has enough physical resources for that purpose) it 
should be possible.

But to read this to require a toaster that a piece of free software came 
installed on to be modifiable by the licensee who choses to excercise 
his rights under the GPL, in the same way as the original developer was 
able to modify that toaster is ... quite creative too i think, and leads 
to many absurd results.

> If you ask the legal profession about this seriously the answer you 
> get is bluntly "There is no caselaw I am aware of", which means that 
> nobody knows. Obviously Tivo and their legal counsel have formed an 
> opinion and have based their actions upon that opinion.

i guess i'll take Linus' word that the FSF's own lawyers agreed that the 
distribution of the Tivo box does not break the GPLv2. (although the 
cynic in me might say that this could be a self-serving position on 
their behalf done for tactical reasons, to increase the perceived 
'justification' for the GPLv3.)

	Ingo

[*] actually, you can get them if you want to, because i very much trust 
    you :-)

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alan Cox]

> > A Tivo box is a collection of literary works protected by copyright, 
> > designs protected by design patents and copyright, names and logos 
> > protected by trademarks, functionalities protected by patents and many 
> > more things. These are the things that restrict what I may do with it 
> > and how I may treat it. The collection of bits of metal and sand 
> > aren't really of relevance in terms of licencing.
> 
> If you are into technicalities then you fail to achieve that "rigorous 
> base" by a wide margin. The Tivo box is not "a collection of literary 
> works", it is a piece of matter, that also happens to contain fixated 

The physical matter is irrelevant. I am perfectly entitled to own, shape
and fiddle with sand and bits of metal. If I wish to remove the software
from the tivo, melt it down and cast the result into the shape of an
obscene gesture and wear it at the tivo shareholder meeting so be it. At
that point it would be my work made from melting the tivo that was the
protected work - same matter.

> copies of literary (and other) works. The Tivo box is just one copy of 
> those works - it is not "a collection of literary works". (Only if there 
> was just a single Tivo box on the planet then could that box itself be 
> meaningfully called a collection of works - a single and unique "master 
> copy" of a work can be called the work itself.)

Each copy is an instance of the work. My copy does not change its status,
nor its legal situation if someone rounds up every other tivo and melts
them down. I guess if you want to be pedantic the Tivo contains "an
instance of the work"

> the work is not the copy! The work is a more 'abstract' entity. The word 
> "copyright" comes straight from that: the right to create specific 
> copies of the work. And that's another reason why it's nonsensical to 
> suggest that somehow the GPLv2 gives us the right to influence the 
> hardware environment that the copy of the kernel got fixated into. We 

We have every legal right to do so. I am perfectly permitted to try
to grant you the right to reproduce my work only if you pay me $25 and the
reproductions are provided in a silver box with flashy blue lights. I am
perfectly permitted as author of a work to tell you "no". You as box
maker are perfectly at liberty to tell me where to go stick my offer and
just not use my work.

I can influence your hardware all I like. What I cannot do is influence
you in any way if you decide not to take any action involving my
copyright. Nor can I through copyright require certain kinds of condition
(eg control other works on the same media) as that requires contract law
and a proper contract, nor certain things that are deemed to be unlawful
by the state (The GPL gives me the right to modify the code to break into
the DoD, steal all their secrets and mail them to the Iraqi government,
the law of the USA not unsuprisingly takes that right away).

> More down the technicalities road: the Tivo box also contains many items 
> that are not copies of works protected by copyright: common types of 
> screws that are not original forms of expression that are creative 
> enough enough to gain copyright protection. Or numbers painted on 
> various places. Or computer-originated random output. Copies of works 
> that have entered the public domain and thus are not under the scope of 
> copyright protection.

And this matters because ?

> Neither is the Tivo box "collection of functionalities protected by 
> patents", if then it is an embodiment of a method and apparatus, which 
> invention is under patent protection (there are other types of patents 
> as well), or which invention might not be under patent protection but 
> have a patent application pending. (which might or might not issue at 
> the end of the patent application process.)

Ok I guess thats a question of level of abstraction, like being "an
instance"

> > "The source code for a work means the preferred form of the work for
> > making modifications to it"
> 
> i think it is clear what is intended with this section: that for example 

The Lawyers don't. As experts in their field I generally trust their view
on this. Also remember that lawyers assess legality not morality so there
are other questions to ask than "will I get sued".

> But to read this to require a toaster that a piece of free software came 
> installed on to be modifiable by the licensee who choses to excercise 
> his rights under the GPL, in the same way as the original developer was 
> able to modify that toaster is ... quite creative too i think, and leads 
> to many absurd results.

Agreed. But GPLv2 has many absurdities such as the way it handles
copyight notices. It wasn't designed when GUI apps were the norm, it
predates web hosted services and the GPL mobile phone was, I suspect, not
on the drafters radar let alone in their pocket.

If my toaster is ROM based then it is difficult to argue that the
preferred form for modification is anything but the code and usual build
files. If the system is writable then it is possible if not reasonable to
argue that the preferred form includes the information needed to load
that modified image, or should do so.

What this means for the FSF goals if Tivo get up one morning and switch
their system firmware to ROM however is interesting 8)

Alan

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 15, 2007, Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:

> What this means for the FSF goals if Tivo get up one morning and switch
> their system firmware to ROM however is interesting 8)

I'm not the FSF, and I don't speak for it, but it seems to me that
this would be "mission accomplished".

The goal AFAIK is not to force people to enable others to hack the
hardware or software to their liking.  The goal is respect for the
freedoms, it's not making it more difficult for others to do what you
can and want to do.  I guess it also goes under the name "Golden
Rule".  Others might phrase it as tit-for-tat, or quid pro quo.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bernd Schmidt]

Alexandre Oliva wrote:
> On Jun 15, 2007, Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:
> 
>> What this means for the FSF goals if Tivo get up one morning and switch
>> their system firmware to ROM however is interesting 8)
> 
> I'm not the FSF, and I don't speak for it, but it seems to me that
> this would be "mission accomplished".

This is insane.  You start with a lofty ideal involving "freedom", and
when you end up with a meaningless technicality (and in technical terms
a change for the worse) you consider it a victory?

Yes, I know that this is what happens in politics (look here, our laws
had an effect!), but I have more respect for you than to think you fall
for these kinds of games.  I do not wish to revise my opinion.


Bernd

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 16, 2007, Bernd Schmidt <bernds_cb1@t-online.de> wrote:

> Alexandre Oliva wrote:
>> On Jun 15, 2007, Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:
>> 
>>> What this means for the FSF goals if Tivo get up one morning and switch
>>> their system firmware to ROM however is interesting 8)
>> 
>> I'm not the FSF, and I don't speak for it, but it seems to me that
>> this would be "mission accomplished".

> This is insane.  You start with a lofty ideal involving "freedom", and
> when you end up with a meaningless technicality (and in technical terms
> a change for the worse) you consider it a victory?

It accomplishes the mission in that everyone is on the same grounds.
Same freedom for everyone.  If the vendor tries to keep a privilege
over the software to itself, denying it to its customers, it's failing
to comply with the spirit of the license.  It's really this simple.
Is this so hard to understand?

The goal is not to push vendors away from GPLed software.  If they
can't permit modification of the software, that's fine, they can still
accomplish this.

What they can't do is deny it to customers while they retain it to
themselves.  This is unfair, this is wrong, and this disrespects
users' freedoms.  Therefore, the GPL should not permit it.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Al Viro]

On Sat, Jun 16, 2007 at 01:57:59PM -0300, Alexandre Oliva wrote:
> On Jun 16, 2007, Bernd Schmidt <bernds_cb1@t-online.de> wrote:
> 
> > Alexandre Oliva wrote:
> >> On Jun 15, 2007, Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:
> >> 
> >>> What this means for the FSF goals if Tivo get up one morning and switch
> >>> their system firmware to ROM however is interesting 8)
> >> 
> >> I'm not the FSF, and I don't speak for it, but it seems to me that
> >> this would be "mission accomplished".
> 
> > This is insane.  You start with a lofty ideal involving "freedom", and
> > when you end up with a meaningless technicality (and in technical terms
> > a change for the worse) you consider it a victory?
> 
> It accomplishes the mission in that everyone is on the same grounds.
> Same freedom for everyone.  If the vendor tries to keep a privilege
> over the software to itself, denying it to its customers, it's failing
> to comply with the spirit of the license.  It's really this simple.
> Is this so hard to understand?
> 
> The goal is not to push vendors away from GPLed software.  If they
> can't permit modification of the software, that's fine, they can still
> accomplish this.
> 
> What they can't do is deny it to customers while they retain it to
> themselves.  This is unfair, this is wrong, and this disrespects
> users' freedoms.  Therefore, the GPL should not permit it.

How the hell does that improve the situation for users?  Alexandre,
please realize that you are preaching to non-believers.  I realize
that you have accepted the FSF credo, but if you want that conversation
to go anywhere you have to separate the things you believe in from
the things you can rationally explain.  Apologetics of your variety is 
not going to cut it.  _Can_ you separate the things relying on your
beliefs from the things that can stand on their own?  If you can't
do that, please stop wasting everyone's time and bandwidth.  It's
a secular maillist; what any of us might happen to believe in is personal
and frankly, none of your damn business.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 16, 2007, Al Viro <viro@ftp.linux.org.uk> wrote:

> How the hell does that improve the situation for users?

Maybe it doesn't.  How does it make it worse?

Maybe just providing an incentive for the vendor to respect users'
freedoms will do the trick, and *some* vendors will do, while those
who can't will keep the status quo.

And then we're likely to be better off.

> I realize that you have accepted the FSF credo, but if you want that
> conversation to go anywhere you have to separate the things you
> believe in from the things you can rationally explain.

I've already explained what the spirit of the GPL is.

I've already explained that the anti-Tivoization provision is in line
with it.


I've already asked in what sense Tivoization makes for a better
tit-for-tat, and got no reply whatsoever, rational or otherwise.

I have already hinted at why it makes things worse.


You don't have to believe what I believe to analyze the arguments
rationally, just like I don't have to believe what you believe to
analyze your arguments rationally.

We may still get to different conclusions as to what is better, if we
have different values guiding us.

But whatever conclusion you arrive at won't change the plain fact that
Tivoization is against the spirit of the GPL, because it is a means to
restrict users' freedoms that the GPL is designed to defend.


It's really this simple.  I'm not trying to convince you of anything
other than that the spirit of the GPL is not being changed at all.
You don't have to agree with that spirit in order to accept this
simple fact.  And while people keep on spreading this lie, I'll be
inclined to point out that it's false.


See, this is not about promoting GPLv3, or "pushing it down your
throats", as some have claimed.  This feeling is just a symptom of the
high rejection for the FSF ideology, that appears to blind so many
smart people from rational reasoning on matters that touch the FSF
ideology.

This is not even about showing that the letter of GPLv2 prohibited
Tivoization.  My arguments concerning Tivoization were all about the
spirit of the license, and unfortunately so many people seem unable to
tell the spirit from the letter that they keep on moving the
discussion to legal technicalities, and then they shoot straw men and
feel happy that they shot an argument.  But the argument stands
untouched, and the straw man was already dead before.  Wasted time.


As far as I'm concerned, Linux is released under GPLv2, and that's a
good thing.  It's unlikely to change.  I wish it changed for better,
but that's just me, and my contributions to Linux in term of code are
really minimal.  I have no say on that.


But as someone involved in the GPLv3 development, it saddens me when
people lie about it.  I feel it's my moral obligation to set the
record straight.  And that's what I've been trying to do.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Jesper Juhl]

On 16/06/07, Alexandre Oliva <aoliva@redhat.com> wrote:
> On Jun 16, 2007, Al Viro <viro@ftp.linux.org.uk> wrote:
>
> > How the hell does that improve the situation for users?
>
> Maybe it doesn't.  How does it make it worse?
>
Now not even the vendor can upgrade the software in the hardware and
fix problems for the user. The user loses.

-- 
Jesper Juhl <jesper.juhl@gmail.com>
Don't top-post  http://www.catb.org/~esr/jargon/html/T/top-post.html
Plain text mails only, please      http://www.expita.com/nomime.html

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Sat, 16 Jun 2007, Jesper Juhl wrote:
>
> Now not even the vendor can upgrade the software in the hardware and
> fix problems for the user. The user loses.

You're seeing that the wrong way.

The correct response is (and I quote from the manual, pick one talking 
point at random):

   "This is a great step for freedom, as the users now have exactly the
    same rights as the vendors."

   "When we talk about free software, we don't talk about 'free as in 
    beer', we talk about 'free as in buggy and unfixable'"

   "You're now at least no less free than anybody else!"

   "Oh, except for the fact that those other people still design the 
    hardware you are using, and the programs you watch. But we have a 
    plan for that too! We will make the GPLv4 outlaw Disney and Britney 
    Spears!"

   "In order to protect your freedoms, we sometimes have to take some 
    freedoms away. In particular, the freedom of critical thinking got 
    revoked last year, because people were just too 'confused'"

   "There are no American Infidels in Baghdad. Never!"

There's a long list of those things, but sadly I didn't have time to copy 
them all when I sneaked into the FSF main office in my ninja suit under 
the cover of darkness.

		Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Al Viro]

On Sat, Jun 16, 2007 at 03:12:57PM -0700, Linus Torvalds wrote:
>    "In order to protect your freedoms, we sometimes have to take some 
>     freedoms away. In particular, the freedom of critical thinking got 
>     revoked last year, because people were just too 'confused'"

ITYM "upgraded to Loyalty To The Cause, which is the best antidote against
doubt and confusion plaguing the so-called rationalists".

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 16, 2007, "Jesper Juhl" <jesper.juhl@gmail.com> wrote:

> On 16/06/07, Alexandre Oliva <aoliva@redhat.com> wrote:
>> On Jun 16, 2007, Al Viro <viro@ftp.linux.org.uk> wrote:
>> 
>> > How the hell does that improve the situation for users?
>> 
>> Maybe it doesn't.  How does it make it worse?
>> 
> Now not even the vendor can upgrade the software in the hardware and
> fix problems for the user. The user loses.

Assuming the vendor's intent as for patching the software is to help
the user.  If the vendor doesn't want to let the user do that
independently, why should this assumption hold?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Sat, 16 Jun 2007, Alexandre Oliva wrote:
> 
> I've already explained what the spirit of the GPL is.

No. You've explained one thing only: that you cannot see that people don't 
*agree* on the "spirit".

You think that there is only one "spirit", and that you own the code-book, 
and that your spirit is thus the only right one.

This is where we started. The same way you seem to think that "freedom" 
has only the meaning *you* and the FSF give it, and that somehow the 
spirit of the GPL includes the "four freedoms" that aren't even 
_mentioned_ in it.

THAT IS NOT TRUE.

But equally importantly, it's not even *relevant*. Nobody is suing the FSF 
for contract violation for changing the spirit. Yes, people have brought 
out the argument that the GPLv3 actually even changes the spirit, and you 
don't seem to realize that people can have different opinions. You just 
repeat YOUR OWN OPINION about the spirit over and over again.

But even if the spirit changes, so what? The GPL doesn't actually say 
"same in spirit". It says "similar in spirit", implying that the spirit is 
"similar". 

In other words, your arguments are wholly irrelevant.

> I've already explained that the anti-Tivoization provision is in line
> with it.

.. and we have already explained to you that it's irrelevant. 

So let's get back to the *real* issue:

 - The GPLv2 was ok with Tivo.

   I really tried to explain to you *why* that was, but by now, I can't be 
   bothered any more. Even if you cannot understand it, just accept it. 

   And if you have a hard time accepting it, just accept the fact that the 
   FSF thinks Tivo cannot be sued, which is just another way of saying 
   "they didn't actually break the license".

 - *I* think Tivo is fine. Other people think Tivo is fine. Other people 
   have told you they think what Tivo did is fine. Some people have even 
   said that they don't like Tivo, but that they don't think the license 
   should stop Tivo.

 - The GPLv3 tries to stop Tivo.

Instead of mumbling about your spirit and feelings (I need to be a whole 
lot more drunk before I start caring), how about you look at those three 
statements, and then admit that you see why the people in bullet#2 think 
that

	GPLv2 is a better license than GPLv3

I don't *care* how you mangle the "spirit of the GPLv2", because that was 
never the issue.

What I care about is that the GPLv3 is a _worse_license_ than GPLv2, and 
that I'd be stupid to select the worse of two licenses, wouldn't I?

So just stop *whining* about this.

The GPLv3 is the worse license, for anybody who thinks that what Tivo did 
should not be against the license. It really is that simple.

And again: you don't even have to *like* Tivo to realize that the license 
itself shouldn't try to spell out some anti-Tivo measures. As I've _also_ 
tried to explain, the anti-Tivo measures are actually more than just "anti 
Tivo". They are also "anti-anything-else-that-might-want-to-lock-down-a-
specific-version-for-security-or-regulatory-reasons".

But in the end, it really hinges on a very simple concept:

 - Not everybody thinks like you or agrees with you.

 - In particular, the original copyright author in the kernel does *not* 
   think like you, and *realized* that he doesn't really like the FSF 
   religious agenda years and years ago, and made sure that the FSF cannot 
   control the licensing of the Linux kernel.

If you don't accept those two simple *facts*, I don't know what's wrong 
with you.

			Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 17, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:

> On Sat, 16 Jun 2007, Alexandre Oliva wrote:
>> 
>> I've already explained what the spirit of the GPL is.

> No. You've explained one thing only: that you cannot see that people don't 
> *agree* on the "spirit".

They don't have to.

Just like nobody but you can tell why you chose the GPLv2, nobody but
RMS can tell why he wrote the GPL.  And the intent behind writing the
GPL is what defines its spirit.

> Yes, people have brought out the argument that the GPLv3 actually
> even changes the spirit,

And that's the point that I'm fighting here.  It does not change the
spirit.  It's still ensuring that Free Software remains Free:
respecting and defending the four freedoms defined in the Free
Software definition.

>> I've already explained that the anti-Tivoization provision is in line
>> with it.

> .. and we have already explained to you that it's irrelevant. 

It is relevant.  It was the point that my participation was intended
to address.

I guess it is just too hard to accept that an FSFer could not be
trying to force GPLv3 down your throat or some other such nonsense.

>  - The GPLv2 was ok with Tivo.

There's disagreement about this, even among developers of the kernel
Linux, and you know it.

I know you're always right and I pretend to respect that ;-), but why
do you think your opinion should prevail over theirs?  Don't you
realize that they're as entitled as you are to enforce the license,
and in the way *they* (not you) perceive and meant to license their
code?

And then again, this is not something I'm overly concerned about.  I
probably don't have enough contributions to Linux for my take on it to
make any difference whatsoever.

This is not the real issue at all.  The real issue, that brought me
here and got you to name calling me and the FSFs, is that there were
false claims about the GPLv3 that I wanted to dispell, particularly
the point about its changing the spirit.  The anti-tivoization
provisions are in the spirit of the GPL, and so much so that a number
of people perceive them as already covered by GPLv2.

>  - The GPLv3 tries to stop Tivo.

A minor nit, but no, it doesn't.  It tries to stop the practice of
tivoization on programs licensed under the GPLv3.

TiVo has a number of choices, and so do other tivoizers, even if they
adopt software under the GPLv3.


> What I care about is that the GPLv3 is a _worse_license_ than GPLv2,

Even though anti-tivoization furthers the quid-pro-quo spirit that you
love about v2, and anti-tivoization is your only objection to v3?
That's what I don't understand.  This is so obviously contradictory to
me that it's almost funny, and you've so far dodged my questions about
this and refrainied from commenting on this contradiction so much that
it looks like it's a blind spot in your mind.

> I'd be stupid to select the worse of two licenses, wouldn't I?

Yes.  That's precisely why I don't understand your stance.  Because I
expect you to be intelligent, but starting from your stated motivation
for choosing GPLv2, and from the consequences of the anti-tivoization
provisions, you'd satisfy your motivations better with v3.

Tivoization reduces the motivation for customers of tivoized devices
to improve the software.  You end up with contributions from the
manufacturers alone, instead of from all the user community.

With explicit anti-tivoization provisions, you may very well lose
contributions from some tivoizers, but for those who change their
stance, you gain far more contributors.  You don't need a lot of
tivoizers to take the path of freedom for you to win big time in the
bottom line that you posed as the only relevant one.

You see why I don't understand your position?

> They are also "anti-anything-else-that-might-want-to-lock-down-a-
> specific-version-for-security-or-regulatory-reasons".

It's not, this is false.  "Lock down" is permitted.  It just won't
work if the business model depends on modifying stuff behind the
user's back.  But other cases of "lock down" are permitted:

  this requirement does not apply if neither you nor any third party
  retains the ability to install modified object code on the User
  Product (for example, the work has been installed in ROM).

>  - Not everybody thinks like you or agrees with you.

>  - In particular, the original copyright author in the kernel does *not* 
>    think like you, and *realized* that he doesn't really like the FSF 
>    religious agenda years and years ago, and made sure that the FSF cannot 
>    control the licensing of the Linux kernel.

I hereby acknowledge, one more time, that I accept these facts.

Since we're in such a good mood now, would you mind acknoledging some
other simple facts, such that we can end this discussion?

- the spirit of the GNU GPL, written by RMS in the FSF, is to keep
Free Software Free, respecting and defending the freedoms of users of
software licensed under the GPL

  It can serve other goals, and some people, yourself included, chose
  it for other reasons, but the intent, the spirit of the license is
  what its author intended it to be, just like the intent behind each
  contribution to Linux is whatever the author of the contribution
  meant it to be.

- GPLv3 does not change this spirit

  On the contrary, it advances this spirit.  Given that defending
  these freedoms is the mission of the FSF, it's no surprise that it
  does revise the GPL to do it.  It's not like it has a choice.


- Tivoization reduces the incentive for contributions

  Customers of tivoized devices can't enjoy or even test the benefits
  of their modifications to the software on the device where the
  modifications would be most useful for them.


- anti-tivoization provisions encourage tivoizers who can respect
  users' freedoms to do so

  If the choice is that or not being able to change the software for
  the user or adopting another platform, they may very well choose
  this option, and then you get not only more users and mind-share,
  but also far more contributors, and the community of developers that
  forms around the product benefits the former-tivoizer as well.


Are these so hard to accept?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Sunday 17 June 2007 01:09:01 Alexandre Oliva wrote:
> On Jun 17, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:
> > On Sat, 16 Jun 2007, Alexandre Oliva wrote:
> >> I've already explained what the spirit of the GPL is.
> >
> > No. You've explained one thing only: that you cannot see that people
> > don't *agree* on the "spirit".
>
> They don't have to.
>
> Just like nobody but you can tell why you chose the GPLv2, nobody but
> RMS can tell why he wrote the GPL.  And the intent behind writing the
> GPL is what defines its spirit.

"Charging for programs is an crime against humanity"

> > Yes, people have brought out the argument that the GPLv3 actually
> > even changes the spirit,
>
> And that's the point that I'm fighting here.  It does not change the
> spirit.  It's still ensuring that Free Software remains Free:
> respecting and defending the four freedoms defined in the Free
> Software definition.

See, you can't even keep the FSF's "Free Software Definition" and its 
inherent "religion" out of the discussion. Sure, the FSF can claim that the 
GPL is intended as a way to "defend" the "Four Freedoms" defined *BY* *THEM*, 
but unless alluded to in the license, the only bearing it can have, anywhere, 
is on the "intent" of the license, as seen by the FSF. And if the "ability to 
run a "covered work" on any piece of hardware" is "freedom 0" then binary 
distribution is in violation of the "spirit" - I can't run an x86 binary on a 
PPC. Isn't that a "designed in hardware restriction" that violates 
the "spirit" of the license ?
 
> >> I've already explained that the anti-Tivoization provision is in line
> >> with it.
> >
> > .. and we have already explained to you that it's irrelevant.
>
> It is relevant.  It was the point that my participation was intended
> to address.
>
> I guess it is just too hard to accept that an FSFer could not be
> trying to force GPLv3 down your throat or some other such nonsense.
>
> >  - The GPLv2 was ok with Tivo.
>
> There's disagreement about this, even among developers of the kernel
> Linux, and you know it.
>
> I know you're always right and I pretend to respect that ;-), but why
> do you think your opinion should prevail over theirs?  Don't you
> realize that they're as entitled as you are to enforce the license,
> and in the way *they* (not you) perceive and meant to license their
> code?
>
> And then again, this is not something I'm overly concerned about.  I
> probably don't have enough contributions to Linux for my take on it to
> make any difference whatsoever.
>
> This is not the real issue at all.  The real issue, that brought me
> here and got you to name calling me and the FSFs, is that there were
> false claims about the GPLv3 that I wanted to dispell, particularly
> the point about its changing the spirit.  The anti-tivoization
> provisions are in the spirit of the GPL, and so much so that a number
> of people perceive them as already covered by GPLv2.
>
> >  - The GPLv3 tries to stop Tivo.
>
> A minor nit, but no, it doesn't.  It tries to stop the practice of
> tivoization on programs licensed under the GPLv3.
>
> TiVo has a number of choices, and so do other tivoizers, even if they
> adopt software under the GPLv3.
>
> > What I care about is that the GPLv3 is a _worse_license_ than GPLv2,
>
> Even though anti-tivoization furthers the quid-pro-quo spirit that you
> love about v2, and anti-tivoization is your only objection to v3?
> That's what I don't understand.  This is so obviously contradictory to
> me that it's almost funny, and you've so far dodged my questions about
> this and refrainied from commenting on this contradiction so much that
> it looks like it's a blind spot in your mind.

Not in the least. They still have to release their changes if they want to 
sell their devices. Or are you so blinded by your belief that the FSF and RMS 
can't be wrong that you can't understand that?

> > I'd be stupid to select the worse of two licenses, wouldn't I?
>
> Yes.  That's precisely why I don't understand your stance.  Because I
> expect you to be intelligent, but starting from your stated motivation
> for choosing GPLv2, and from the consequences of the anti-tivoization
> provisions, you'd satisfy your motivations better with v3.

It is only *YOUR* opinion that the GPLv3 is the better license. As Linus has 
explained, he doesn't share your viewpoint.

> Tivoization reduces the motivation for customers of tivoized devices
> to improve the software.  You end up with contributions from the
> manufacturers alone, instead of from all the user community.

No, it reduces their motivation to improve the software on *those* devices. If 
they like the software enough to actually download the source, they probably 
also liked it enough to install it on their computer *AND* will modify it to 
make it work better on their computer.

> With explicit anti-tivoization provisions, you may very well lose
> contributions from some tivoizers, but for those who change their
> stance, you gain far more contributors.  You don't need a lot of
> tivoizers to take the path of freedom for you to win big time in the
> bottom line that you posed as the only relevant one.

With your argument about reduced motive shotten down this portion falls apart.

> You see why I don't understand your position?
>
> > They are also "anti-anything-else-that-might-want-to-lock-down-a-
> > specific-version-for-security-or-regulatory-reasons".
>
> It's not, this is false.  "Lock down" is permitted.  It just won't
> work if the business model depends on modifying stuff behind the
> user's back.  But other cases of "lock down" are permitted:
>
>   this requirement does not apply if neither you nor any third party
>   retains the ability to install modified object code on the User
>   Product (for example, the work has been installed in ROM).
>
> >  - Not everybody thinks like you or agrees with you.
> >
> >  - In particular, the original copyright author in the kernel does *not*
> >    think like you, and *realized* that he doesn't really like the FSF
> >    religious agenda years and years ago, and made sure that the FSF
> > cannot control the licensing of the Linux kernel.
>
> I hereby acknowledge, one more time, that I accept these facts.
>
> Since we're in such a good mood now, would you mind acknoledging some
> other simple facts, such that we can end this discussion?
>
> - the spirit of the GNU GPL, written by RMS in the FSF, is to keep
> Free Software Free, respecting and defending the freedoms of users of
> software licensed under the GPL

Agreed. The disagreement is about what that spirit is. I feel that its spirit 
is in the free and open exchange of ideas, as personified by the software 
people write. I *ALSO* feel that it's spirit lies in the phrase "do whatever 
you want with the software as long - but if you add your own ideas to it, 
give them back to the people like your inspiration was given to you."

You,  the FSF and, apparently, RMS, feel it is about the "Four Freedoms" as 
defined by RMS. I'm quite sure that my view is much more common among the 
people that *DON'T* think that the FSF and RMS are never wrong.

>   It can serve other goals, and some people, yourself included, chose
>   it for other reasons, but the intent, the spirit of the license is
>   what its author intended it to be, just like the intent behind each
>   contribution to Linux is whatever the author of the contribution
>   meant it to be.
>
> - GPLv3 does not change this spirit
>
>   On the contrary, it advances this spirit.  Given that defending
>   these freedoms is the mission of the FSF, it's no surprise that it
>   does revise the GPL to do it.  It's not like it has a choice.

Again, that is *your* version of the "spirit".

>
> - Tivoization reduces the incentive for contributions
>
>   Customers of tivoized devices can't enjoy or even test the benefits
>   of their modifications to the software on the device where the
>   modifications would be most useful for them.

I agree to the "can't enjoy or test" bits. But I don't believe that it reduces 
anything. Personally I feel that anything that exposes people to "Free 
Software" is a *BONUS*.

>
> - anti-tivoization provisions encourage tivoizers who can respect
>   users' freedoms to do so
>
>   If the choice is that or not being able to change the software for
>   the user or adopting another platform, they may very well choose
>   this option, and then you get not only more users and mind-share,
>   but also far more contributors, and the community of developers that
>   forms around the product benefits the former-tivoizer as well.
>
>
> Are these so hard to accept?

Yes. Because a number of your "facts" are massively flawed. Now, please, 
you've proven to me that you can't, in fact, do any *objective* thinking 
about this topic.

When you are ready to drop your pre-conceptions and think *objectively* about 
the topic, come back and talk. Until then, go away.

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 17, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> On Sunday 17 June 2007 01:09:01 Alexandre Oliva wrote:
>> On Jun 17, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:
>> > On Sat, 16 Jun 2007, Alexandre Oliva wrote:
>> >> I've already explained what the spirit of the GPL is.
>> >
>> > No. You've explained one thing only: that you cannot see that people
>> > don't *agree* on the "spirit".
>> 
>> They don't have to.
>> 
>> Just like nobody but you can tell why you chose the GPLv2, nobody but
>> RMS can tell why he wrote the GPL.  And the intent behind writing the
>> GPL is what defines its spirit.

> "Charging for programs is an crime against humanity"

Assuming he actually said that, I have no doubt that it would
pre-dates by far even the Free Software Definition, let alone the GPL.

> See, you can't even keep the FSF's "Free Software Definition" and its 
> inherent "religion" out of the discussion.

Of course not.  That's what the spirit of the GPL is all about.  And
the spirit of the GPL is what the discussion is all about for me.

> Sure, the FSF can claim that the GPL is intended as a way to
> "defend" the "Four Freedoms" defined *BY* *THEM*, but unless alluded
> to in the license, the only bearing it can have, anywhere, is on the
> "intent" of the license, as seen by the FSF.

Exactly!  And since the *Free* *Software* Foundation wrote the
license, and documented the goals in the preamble, referring to
keeping *free* *software* *free*, it is quite safe to say that this
*is* indeed the intent, the spirit of the GPL.

> And if the "ability to run a "covered work" on any piece of hardware
> is "freedom 0" then binary distribution is in violation of the
> "spirit" - I can't run an x86 binary on a PPC. Isn't that a
> "designed in hardware restriction" that violates the "spirit" of the
> license ?

It's not.  freedom and ability have two very different meanings.

Freedom to run the software for any purpose means that people won't
stop you from doing that.  It may take you some work, such as porting
the software, rebuilding it, etc.  But if, at the end of that effort,
you find that it will run on your development machine, but not in a
machine where the original software runs on, and that's because the
manufacturer imposed prohibitions on running unauthorized versions of
the software, then the manufacturer of the hardware is very clearly
disrespecting your freedom #0 WRT that software.

Demanding the ability to run the software for any purpose, without any
effort whatsoever, would indeed be nonsensical.

 
(BTW, covered work is a legal term, only present in the legal portion
of the license, which I'm actively avoiding, because I'm not a lawyer,
and my point is about the spirit.  but I'm sure I wrote that before
;-)

>> > I'd be stupid to select the worse of two licenses, wouldn't I?
>> 
>> Yes.  That's precisely why I don't understand your stance.  Because I
>> expect you to be intelligent, but starting from your stated motivation
>> for choosing GPLv2, and from the consequences of the anti-tivoization
>> provisions, you'd satisfy your motivations better with v3.

> It is only *YOUR* opinion that the GPLv3 is the better license.

Are you even reading what I write?

>> Tivoization reduces the motivation for customers of tivoized devices
>> to improve the software.  You end up with contributions from the
>> manufacturers alone, instead of from all the user community.

> No, it reduces their motivation to improve the software on *those*
> devices. If they like the software enough to actually download the
> source, they probably also liked it enough to install it on their
> computer *AND* will modify it to make it work better on their
> computer.

Sure, but that's a different point.  They could do that with or
without tivoization.

The point is that, if they have an issue with the program in the
device, and they'd like to improve it, but they find that they won't
be able to use their modification to get the device to do what they
want, they're less likely to make the change.

Now multiply this by all customers, and see how much you're losing by
permitting tivoization, assuming that at least some tivoizers would
change their minds towards respecting users' freedoms, if faced with
an anti-tivoization licensing provision.

> With your argument about reduced motive shotten down this portion falls apart.

A distraction doesn't shoot down an argument.

>> - the spirit of the GNU GPL, written by RMS in the FSF, is to keep
>> Free Software Free, respecting and defending the freedoms of users of
>> software licensed under the GPL

> Agreed. The disagreement is about what that spirit is.

What is the 'agreed' supposed to mean, then? ;-)

> I feel that its spirit is in the free and open exchange of ideas, as
> personified by the software people write. I *ALSO* feel that it's
> spirit lies in the phrase "do whatever you want with the software as
> long - but if you add your own ideas to it, give them back to the
> people like your inspiration was given to you."

You're entitled to have these motivations to release software under
GPLv2, or any other license that you believe furthers these goals.

But you have no say whatsoever on what intent RMS had when he wrote
the GPL, just like he has no say whatsoever on what intent you have
when you choose the GPL for your program.

> You, the FSF and, apparently, RMS, feel it is about the "Four
> Freedoms" as defined by RMS.

s/feel/know/

> I'm quite sure that my view is much more common among the people
> that *DON'T* think that the FSF and RMS are never wrong.

Others can choose the GPL for other reasons.  There's nothing wrong
with this.

What's wrong is to then complain that the GPL is changing the spirit,
just because the revised version allegedly no longer matches their
reasons.

>> - GPLv3 does not change this spirit

>> On the contrary, it advances this spirit.  Given that defending
>> these freedoms is the mission of the FSF, it's no surprise that it
>> does revise the GPL to do it.  It's not like it has a choice.

> Again, that is *your* version of the "spirit".

Again, this is the FSF version of the spirit, the only one that
matters as far as "the spirit of the GPL" is concerned.

>> - Tivoization reduces the incentive for contributions

>> Customers of tivoized devices can't enjoy or even test the benefits
>> of their modifications to the software on the device where the
>> modifications would be most useful for them.

> I agree to the "can't enjoy or test" bits.

See, no "argument shot down" ;-)

> But I don't believe that it reduces anything.

It reduces the incentive for these users to collaborate.

Or, if you want to put it in a positive tone, the ability to enjoy and
test their modifications would grow the number of contributors.  More
"giving back in kind".  More tit-for-tat.

> Personally I feel that anything that exposes people to "Free
> Software" is a *BONUS*.

I don't think tivoized software qualifies as Free Software any more.

Sure, they still get the sources and that's still Free Software, but
the tivoized binaries aren't.

> Yes. Because a number of your "facts" are massively flawed.

Because you say so?  Even though you agree with points after claiming
to have shot them down?  Now that's rational!

> Now, please, you've proven to me that you can't, in fact, do any
> *objective* thinking about this topic.

/me hands Daniel a mirror

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Sunday 17 June 2007 02:27:42 Alexandre Oliva wrote:
> On Jun 17, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> > On Sunday 17 June 2007 01:09:01 Alexandre Oliva wrote:
> >> On Jun 17, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:
> >> > On Sat, 16 Jun 2007, Alexandre Oliva wrote:
> >> >> I've already explained what the spirit of the GPL is.
> >> >
> >> > No. You've explained one thing only: that you cannot see that people
> >> > don't *agree* on the "spirit".
> >>
> >> They don't have to.
> >>
> >> Just like nobody but you can tell why you chose the GPLv2, nobody but
> >> RMS can tell why he wrote the GPL.  And the intent behind writing the
> >> GPL is what defines its spirit.
> >
> > "Charging for programs is an crime against humanity"
>
> Assuming he actually said that, I have no doubt that it would
> pre-dates by far even the Free Software Definition, let alone the GPL.

The quote actually pre-dates the FSF

> > See, you can't even keep the FSF's "Free Software Definition" and its
> > inherent "religion" out of the discussion.
>
> Of course not.  That's what the spirit of the GPL is all about.  And
> the spirit of the GPL is what the discussion is all about for me.

The "spirit" is no different than "intent". Different words that mean the 
exact same thing.

> > Sure, the FSF can claim that the GPL is intended as a way to
> > "defend" the "Four Freedoms" defined *BY* *THEM*, but unless alluded
> > to in the license, the only bearing it can have, anywhere, is on the
> > "intent" of the license, as seen by the FSF.
>
> Exactly!  And since the *Free* *Software* Foundation wrote the
> license, and documented the goals in the preamble, referring to
> keeping *free* *software* *free*, it is quite safe to say that this
> *is* indeed the intent, the spirit of the GPL.

The intent of the GPL, as seen by the FSF, *DOESN'T* *MATTER* *AT* *ALL* when 
the software isn't licensed by the FSF. Or did you forget that part of the 
discussion?

> > And if the "ability to run a "covered work" on any piece of hardware
> > is "freedom 0" then binary distribution is in violation of the
> > "spirit" - I can't run an x86 binary on a PPC. Isn't that a
> > "designed in hardware restriction" that violates the "spirit" of the
> > license ?
>
> It's not.  freedom and ability have two very different meanings.

This isn't what you've argued before. The hardware doesn't allow me to run the 
software, so its a designed-in limitation on the freedom of the end-user.

> Freedom to run the software for any purpose means that people won't
> stop you from doing that.  It may take you some work, such as porting
> the software, rebuilding it, etc.  But if, at the end of that effort,
> you find that it will run on your development machine, but not in a
> machine where the original software runs on, and that's because the
> manufacturer imposed prohibitions on running unauthorized versions of
> the software, then the manufacturer of the hardware is very clearly
> disrespecting your freedom #0 WRT that software.

But a PPC binary won't run on an x86 either :)

No, I'm in agreement with you here. But I'm smart enough to not buy something 
that does this to me.

> Demanding the ability to run the software for any purpose, without any
> effort whatsoever, would indeed be nonsensical.
>
>
> (BTW, covered work is a legal term, only present in the legal portion
> of the license, which I'm actively avoiding, because I'm not a lawyer,
> and my point is about the spirit.  but I'm sure I wrote that before
> ;-)

I use it because that is the term used in the GPLv2. And since the GPLv3 (dd4) 
is no longer specific to software "covered work" is the best choice.

> >> > I'd be stupid to select the worse of two licenses, wouldn't I?
> >>
> >> Yes.  That's precisely why I don't understand your stance.  Because I
> >> expect you to be intelligent, but starting from your stated motivation
> >> for choosing GPLv2, and from the consequences of the anti-tivoization
> >> provisions, you'd satisfy your motivations better with v3.
> >
> > It is only *YOUR* opinion that the GPLv3 is the better license.
>
> Are you even reading what I write?

Yes. But you are interpreting Linus' intentions using your own preconceived 
beliefs, rather than looking at the situation objectively. 

> >> Tivoization reduces the motivation for customers of tivoized devices
> >> to improve the software.  You end up with contributions from the
> >> manufacturers alone, instead of from all the user community.
> >
> > No, it reduces their motivation to improve the software on *those*
> > devices. If they like the software enough to actually download the
> > source, they probably also liked it enough to install it on their
> > computer *AND* will modify it to make it work better on their
> > computer.
>
> Sure, but that's a different point.  They could do that with or
> without tivoization.

Exactly. Tivoization doesn't make a damned bit of difference.

> The point is that, if they have an issue with the program in the
> device, and they'd like to improve it, but they find that they won't
> be able to use their modification to get the device to do what they
> want, they're less likely to make the change.

They complain to the manufacturer, file a report with a consumer watchdog 
agency and start advising people against buying the device.

> Now multiply this by all customers, and see how much you're losing by
> permitting tivoization, assuming that at least some tivoizers would
> change their minds towards respecting users' freedoms, if faced with
> an anti-tivoization licensing provision.

Apply the same logic to my above statement and tell me - how much money does 
the company lose ?

> > With your argument about reduced motive shotten down this portion falls
> > apart.
>
> A distraction doesn't shoot down an argument.

Quite thoroughly shot down.

> >> - the spirit of the GNU GPL, written by RMS in the FSF, is to keep
> >> Free Software Free, respecting and defending the freedoms of users of
> >> software licensed under the GPL
> >
> > Agreed. The disagreement is about what that spirit is.
>
> What is the 'agreed' supposed to mean, then? ;-)

It means that I agree that the GPL is about "respecting and defending 
freedoms"

> > I feel that its spirit is in the free and open exchange of ideas, as
> > personified by the software people write. I *ALSO* feel that it's
> > spirit lies in the phrase "do whatever you want with the software as
> > long - but if you add your own ideas to it, give them back to the
> > people like your inspiration was given to you."
>
> You're entitled to have these motivations to release software under
> GPLv2, or any other license that you believe furthers these goals.
>
> But you have no say whatsoever on what intent RMS had when he wrote
> the GPL, just like he has no say whatsoever on what intent you have
> when you choose the GPL for your program.

I never claimed I did. I was just pointing out that your belief that there is 
only one "spirit" of the license is complete and utter BS.

> > You, the FSF and, apparently, RMS, feel it is about the "Four
> > Freedoms" as defined by RMS.
>
> s/feel/know/

I said exactly what I meant.

> > I'm quite sure that my view is much more common among the people
> > that *DON'T* think that the FSF and RMS are never wrong.
>
> Others can choose the GPL for other reasons.  There's nothing wrong
> with this.
>
> What's wrong is to then complain that the GPL is changing the spirit,
> just because the revised version allegedly no longer matches their
> reasons.

It isn't, because from my own, and other peoples, viewpoint it *IS*.

> >> - GPLv3 does not change this spirit
> >>
> >> On the contrary, it advances this spirit.  Given that defending
> >> these freedoms is the mission of the FSF, it's no surprise that it
> >> does revise the GPL to do it.  It's not like it has a choice.
> >
> > Again, that is *your* version of the "spirit".
>
> Again, this is the FSF version of the spirit, the only one that
> matters as far as "the spirit of the GPL" is concerned.

So you admit you believe what the FSF says?

> >> - Tivoization reduces the incentive for contributions
> >>
> >> Customers of tivoized devices can't enjoy or even test the benefits
> >> of their modifications to the software on the device where the
> >> modifications would be most useful for them.
> >
> > I agree to the "can't enjoy or test" bits.
>
> See, no "argument shot down" ;-)
>
> > But I don't believe that it reduces anything.
>
> It reduces the incentive for these users to collaborate.
>
> Or, if you want to put it in a positive tone, the ability to enjoy and
> test their modifications would grow the number of contributors.  More
> "giving back in kind".  More tit-for-tat.

If the platform doesn't allow the running of modified binaries, why would the 
modifications matter? Sure, TiVO might like them - hell, they might even pay 
for them - but would anyone else?

So modifications for that "closed execution" platform might suffer, but that 
is the *ONLY* thing that will suffer.

> > Personally I feel that anything that exposes people to "Free
> > Software" is a *BONUS*.
>
> I don't think tivoized software qualifies as Free Software any more.
>
> Sure, they still get the sources and that's still Free Software, but
> the tivoized binaries aren't.

Yes, they are. But only because the law states that they are.

> > Yes. Because a number of your "facts" are massively flawed.
>
> Because you say so?  Even though you agree with points after claiming
> to have shot them down?  Now that's rational!

I agreed with *portions* of some of the statements. However, did I actually 
say which facts I found flawed?

> > Now, please, you've proven to me that you can't, in fact, do any
> > *objective* thinking about this topic.
>
> /me hands Daniel a mirror

I've been looking at this objectively the entire time. That you don't 
understand that is just more proof that you aren't.

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 17, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> The intent of the GPL, as seen by the FSF, *DOESN'T* *MATTER* *AT* *ALL* when 
> the software isn't licensed by the FSF. Or did you forget that part of the 
> discussion?

You're mixing up spirit of license with intent of licensing (or
something else, I don't remember exactly the term I invented to try to
make this distinction clear)

The former is the intent of the FSF for writing the GPL.

The latter is the motivations of the copyright holder to license his
work under that license.

The FSF has no say on the latter.  The copyright holders have no say
on the former.

When people claim the GPL changed its spirit, they're claiming the FSF
changed its intent.

It didn't.

Clear now?


>> > And if the "ability to run a "covered work" on any piece of hardware
>> > is "freedom 0" then binary distribution is in violation of the
>> > "spirit" - I can't run an x86 binary on a PPC. Isn't that a
>> > "designed in hardware restriction" that violates the "spirit" of the
>> > license ?
>> 
>> It's not.  freedom and ability have two very different meanings.

> This isn't what you've argued before. The hardware doesn't allow me
> to run the software, so its a designed-in limitation on the freedom
> of the end-user.

If it doesn't allow you to run the software *because* of a designed-in
limitation on the freedom of the end-user, then it's a disrespect for
the freedom.  I've never changed my position in this regard.  Maybe I
wasn't clear, or you misunderstood, or the network corrupted the bits
in transit ;-)

> No, I'm in agreement with you here. But I'm smart enough to not buy
> something that does this to me.

tit for tat or each for oneself?

>> (BTW, covered work is a legal term, only present in the legal portion
>> of the license, which I'm actively avoiding, because I'm not a lawyer,
>> and my point is about the spirit.  but I'm sure I wrote that before
>> ;-)

> I use it because that is the term used in the GPLv2.

Not in the preamble, which discusses the spirit I'm talking about.

> And since the GPLv3 (dd4) is no longer specific to software "covered
> work" is the best choice.

The preamble of GPLv3 is pretty much the same.

>> >> Tivoization reduces the motivation for customers of tivoized devices
>> >> to improve the software.  You end up with contributions from the
>> >> manufacturers alone, instead of from all the user community.
>> >
>> > No, it reduces their motivation to improve the software on *those*
>> > devices. If they like the software enough to actually download the
>> > source, they probably also liked it enough to install it on their
>> > computer *AND* will modify it to make it work better on their
>> > computer.
>> 
>> Sure, but that's a different point.  They could do that with or
>> without tivoization.

> Exactly. Tivoization doesn't make a damned bit of difference.

For this case.  But it's not the case where I claimed it made a
difference.

>> The point is that, if they have an issue with the program in the
>> device, and they'd like to improve it, but they find that they won't
>> be able to use their modification to get the device to do what they
>> want, they're less likely to make the change.

> They complain to the manufacturer, file a report with a consumer watchdog 
> agency and start advising people against buying the device.

And?  How is this going to achieve more contributions in kind?

>> Now multiply this by all customers, and see how much you're losing by
>> permitting tivoization, assuming that at least some tivoizers would
>> change their minds towards respecting users' freedoms, if faced with
>> an anti-tivoization licensing provision.

> Apply the same logic to my above statement and tell me - how much money does 
> the company lose ?

Dunno.  Not much?  I know I complain to hardware manufacturers that
ship broken BIOSes, to no avail.  It doesn't look like they care, or
that it makes a difference.

And then, I'm not talking about a case in which the thing is broken
(in which case the user might have a real case)

Think of improvements I'd like to make, that I probably won't do
because the hardware won't let me run it.  So you'll never see those
contributions I and all the other untivoized users could make, and you
won't ever know what you're missing.

>> > With your argument about reduced motive shotten down this portion falls
>> > apart.

>> A distraction doesn't shoot down an argument.

> Quite thoroughly shot down.

If you say so, it must be right, in spite of all objective evidence,
eh? :-)

>> >> - the spirit of the GNU GPL, written by RMS in the FSF, is to keep
>> >> Free Software Free, respecting and defending the freedoms of users of
>> >> software licensed under the GPL
>> >
>> > Agreed. The disagreement is about what that spirit is.
>> 
>> What is the 'agreed' supposed to mean, then? ;-)

> It means that I agree that the GPL is about "respecting and defending 
> freedoms"

But still, somehow, that's not its spirit, you say.  I don't get it.

>> What's wrong is to then complain that the GPL is changing the spirit,
>> just because the revised version allegedly no longer matches their
>> reasons.

> It isn't, because from my own, and other peoples, viewpoint it *IS*.

Which shows you don't know what the spirit really is.  It is, and it
has always been, what you agreed above that the GPL was about.

What others think the spirit is doesn't affect what the spirit is.  It
just says what others think the spirit is, and how off the mark they
are in their assessment of the spirit of the license (= intent behind
its creation)

>> >> - GPLv3 does not change this spirit

>> >> On the contrary, it advances this spirit.  Given that defending
>> >> these freedoms is the mission of the FSF, it's no surprise that it
>> >> does revise the GPL to do it.  It's not like it has a choice.

>> > Again, that is *your* version of the "spirit".

>> Again, this is the FSF version of the spirit, the only one that
>> matters as far as "the spirit of the GPL" is concerned.

> So you admit you believe what the FSF says?

I guess this should be pretty obvious that I believe this, yes.

I don't know that I can make a general assertion about my believing
what the FSF says, but I don't remember having had reasons to
disbelieve it.

>> Or, if you want to put it in a positive tone, the ability to enjoy and
>> test their modifications would grow the number of contributors.  More
>> "giving back in kind".  More tit-for-tat.

> If the platform doesn't allow the running of modified binaries, why would the 
> modifications matter? Sure, TiVO might like them - hell, they might even pay 
> for them - but would anyone else?

> So modifications for that "closed execution" platform might suffer, but that 
> is the *ONLY* thing that will suffer.

You're looking at the downside.  Look at the upside: all the
contributions you'll get from users of formerly-tivoized platforms.
Being able to put their efforts to work in their own self-interest, on
their own hardware, they are likely to give more back "in kind" than
the vendor ever could.

And if the vendor doesn't go that way, what you lose are the limited
contributions of that vendor.

Big potential win, small potential loss.  Sounds like a no-brainer to
me, really.

>> > Personally I feel that anything that exposes people to "Free
>> > Software" is a *BONUS*.

>> I don't think tivoized software qualifies as Free Software any more.

>> Sure, they still get the sources and that's still Free Software, but
>> the tivoized binaries aren't.

> Yes, they are. But only because the law states that they are.

I can't make sense of what you're saying.

Are you disputing that tivoized binaries are non-Free Software?

What law could possibly support this claim?

>> > Now, please, you've proven to me that you can't, in fact, do any
>> > *objective* thinking about this topic.
>> 
>> /me hands Daniel a mirror

> I've been looking at this objectively the entire time.

So you say.  I believe I have too.  That we disagree doesn't mean any
of us is not being objective.  It may mean we have different
backgrounds, we're talking past each other, we're not understanding
each other, and a number of other possibilities.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bernd Schmidt]

Alexandre Oliva wrote:
> On Jun 17, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:
>> No. You've explained one thing only: that you cannot see that people don't 
>> *agree* on the "spirit".
> 
> They don't have to.
> 
> Just like nobody but you can tell why you chose the GPLv2, nobody but
> RMS can tell why he wrote the GPL.  And the intent behind writing the
> GPL is what defines its spirit.

Given that a number of people who don't buy into FSF ideology (let's
call them "open source proponents" to contrast them with the "free
software people") have concluded that the GPLv2 achieves their personal
goals, and have chosen the GPLv2 as the license for their projects, I'd
argue that the spirit that is embodied in the GPLv2 is actually a larger
thing than what the FSF intended, and more inclusive.

When these same people now disagree with the GPLv3, it indicates that
something has been lost, and the spirit of the _license_ has changed.
The _intention_ behind writing the license may or may not have been the
same (who can tell, after 20-odd years?), but this is separate from the
spirit embodied in the license itself - the latter has, in my mind
anyway, clearly been changed.  You might prefer to say "clarified", but
it comes down to the same thing.

But personally, I find the discussion about whether the spirit changed
or not somewhat beside the point and not very interesting.  What's
really going to cause problems is the fact that the actual wording took
a turn for the worse.


Bernd

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 17, 2007, Bernd Schmidt <bernds_cb1@t-online.de> wrote:

> Alexandre Oliva wrote:
>> On Jun 17, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:
>>> No. You've explained one thing only: that you cannot see that people don't 
>>> *agree* on the "spirit".
>> 
>> They don't have to.
>> 
>> Just like nobody but you can tell why you chose the GPLv2, nobody but
>> RMS can tell why he wrote the GPL.  And the intent behind writing the
>> GPL is what defines its spirit.

> Given that a number of people who don't buy into FSF ideology (let's
> call them "open source proponents" to contrast them with the "free
> software people") have concluded that the GPLv2 achieves their personal
> goals, and have chosen the GPLv2 as the license for their projects, I'd
> argue that the spirit that is embodied in the GPLv2 is actually a larger
> thing than what the FSF intended, and more inclusive.

This sounds like a good argument, but it doesn't hold water.

Consider this: We manufacture bread toasters and sell them in the
market with great success.  They're big and bulky.  So the engineers
work on reducing its size, but in a way they can still fit perfectly a
slice of bread.  When we launch bread toaster, people complain that
this new product cannot toast bagels any more, that we've changed the
spirit of the bread toaster.


See?  Just because you could use it for other purposes doesn't make
the intent behind it any different.

> When these same people now disagree with the GPLv3, it indicates that
> something has been lost, and the spirit of the _license_ has changed.

It just shows that they've never agreed with the spirit of the license
in the first place.  They just saw it could do something else, and
used it for this reason.  There's nothing wrong about this.

What's wrong is to complain that those who introduced the license with
a specific and public intent, and that advancing that intent with a
new revision of the license, are changing the intent.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Sun, 17 Jun 2007, Alexandre Oliva wrote:
> 
> > What I care about is that the GPLv3 is a _worse_license_ than GPLv2,
> 
> Even though anti-tivoization furthers the quid-pro-quo spirit that you
> love about v2, and anti-tivoization is your only objection to v3?

You apparently do not understand "quid-pro-quo".

Another way of stating it might be "same for same".

A third way of stating it is "software for software". No, the romans never 
said that, but I just did, to make it just more obvious that the whole 
point is that you are expected to answer IN KIND!

I do *not* ask for hardware access.

I do *not* ask for money.

	And the reason I'm harping on "money" is that "money" is something 
	*different* from what I give out. I give out software. I don't 
	expect money in return.

	Money is *irrelevant*. It's allowed (and certainly much 
	appreciated), but it's not required.

See? Can you agree with that? Can you agree that that is actually part of 
what the whole "open source" spirit is all about (I'll avoid the word 
"free software", since you have defined it so rigorously personally that 
it makes no sense any more).

Now, replace "money" with "access to the hardware", and read the exact 
*same* sentences again:

	And the reason I'm harping on "access to hardware" is that "access 
	to hardware" is something *different* from what I give out. I give 
	out software. I don't expect access to hardware in return.

	Access to hardware is *irrelevant*. It's allowed (and certainly 
	much appreciated), but it's not required.

See? 

Exact same words. Exact same spirit. Just using "access to hardware" 
instead of "money".

You have been showing that you have a really hard time understanding that 
very *simple* argument.

> > I'd be stupid to select the worse of two licenses, wouldn't I?
> 
> Yes.  That's precisely why I don't understand your stance.

If you don't understand it after the above, I really can only say:

	"You are either terminally stupid, or you're not allowing yourself 
	 to see an obvious argument, because it destroys your world-view".

The latter is very possible. It's a very human thing. It's why apparently 
a lot of people in the US have a hard time believing in evolution. Are 
they terminally stupid? Yeah, that is quite possible. But it is also 
possible that they are of average intelligence, and they just cannot 
mentally _afford_ to follow the argument - it destroys the silyl stories 
they heard as children, and requires them to think too hard about the 
veracity of the source.

		Linus

PS. Since some people talked about the game theory aspects of 
"tit-for-tat", I'd like to point out that what is usually considered an 
even *better* strategy than "tit-for-tat" is actually "tit-for-tat with 
forgiveness".

In particular, "tit-for-tat with forgiveness" is considered better when 
there is ambiguity (like "communication difficulties" - does that sound 
familiar?) in the encouter. You allow some leeway, and don't always 
retaliate!

So the FSF is DOING THE WRONG THING! They are turning "tit-for-tat" not 
into "tit-for-tat with forgiveness", but into "tit-for-tat with preemptive 
strikes".

That is a *LOSING* strategy in game theory. So a game theorist could very 
well argue with good reason to believe he is right that the GPLv3 is 
actually a worse license even from a purely theoretical standpoint!

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 17, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:

> On Sun, 17 Jun 2007, Alexandre Oliva wrote:
>> 
>> > What I care about is that the GPLv3 is a _worse_license_ than GPLv2,
>> 
>> Even though anti-tivoization furthers the quid-pro-quo spirit that you
>> love about v2, and anti-tivoization is your only objection to v3?

> You apparently do not understand "quid-pro-quo".

> Another way of stating it might be "same for same".

> A third way of stating it is "software for software". No, the romans never 
> said that, but I just did, to make it just more obvious that the whole 
> point is that you are expected to answer IN KIND!

Yes.  And this was precisely what meant when I wrote "quid-pro-quo"
above.

> If you don't understand it after the above, I really can only say:

> 	"You are either terminally stupid, or you're not allowing yourself 
> 	 to see an obvious argument, because it destroys your world-view".

> The latter is very possible. It's a very human thing.

/me hands Linus a mirror


Serious, what's so hard to understand about:

  no tivoization => more users able to tinker their formerly-tivoized
  computers => more users make useful modifications => more
  contributions in kind

?

Sure, there's a downside too:

  no tivoization => fewer contributions from manufacturers that demand
  on tivoization


My perception is that the first easily dominates the second, and so
you are better off without tivoization.


  
> it is also possible that they are of average intelligence, and they
> just cannot mentally _afford_ to follow the argument - it destroys
> the silyl stories they heard as children, and requires them to think
> too hard about the veracity of the source.



> PS. Since some people talked about the game theory aspects of 
> "tit-for-tat", I'd like to point out that what is usually considered an 
> even *better* strategy than "tit-for-tat" is actually "tit-for-tat with 
> forgiveness".

> In particular, "tit-for-tat with forgiveness" is considered better when 
> there is ambiguity (like "communication difficulties" - does that sound 
> familiar?) in the encouter. You allow some leeway, and don't always 
> retaliate!

> So the FSF is DOING THE WRONG THING! They are turning "tit-for-tat" not 
> into "tit-for-tat with forgiveness", but into "tit-for-tat with preemptive 
> strikes".

Wrong.  It enables copyright holders to decide whether forgiveness is
appropriate, rather than forcing them to forgive.  Being forced to
forgive deception is not tit-for-tat, and it's a losing strategy.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Jesper Juhl]

On 17/06/07, Alexandre Oliva <aoliva@redhat.com> wrote:
[snip]
>
> Serious, what's so hard to understand about:
>
>   no tivoization => more users able to tinker their formerly-tivoized
>   computers => more users make useful modifications => more
>   contributions in kind
>
> ?
>
> Sure, there's a downside too:
>
>   no tivoization => fewer contributions from manufacturers that demand
>   on tivoization
>
>
> My perception is that the first easily dominates the second, and so
> you are better off without tivoization.
>

I have to disagree.

Let's say I'm the owner of a company selling some device that uses a
GPLv2 OS and some GPLv2 applications to do the job. Let's say that for
some reason I don't want the end users of my device to tinker with the
software inside my device.  Obviously I release the source for any
modifications I may have made, but I use the hardware to prevent users
from installing modified versions on the device (basically I TiVO'ize
the device).

Now I think you can agree to these things being positive:

 - My use of GPLv2 software in this device results in my employees
being exposed to open source software at work (who knows, some may
even start using such software at home as a result).  A good thing.

 - The source code with my own modifications that I distribute as
required by the GPLv2 can potentially be of use to other developers
working on other GPLv2 software and those other developers are free to
use those modifications.  Also a good thing.

 - When creating marketing material for my device I'll most likely
include information about the fact that I'm using WhatEverOS that is
GPL'ed as well as other Open Source components. This in turn results
in many people becoming aware that such software exist.  I have to say
that this is also a good thing.

 - When dealing with hardware companies supplying bits and pieces for
my device I'll probably push for components that already have open
source drivers, so my partners will find out there is value in having
open source drivers for their stuff and hopefully end up supporting
that.  Yet another good thing.

 - If I end up being happy with my choice of GPLv2 OS & GPLv2 apps
there's a, not insignificant, chance that I'll start helping out with
the development of those components or maybe sponsor other developers
with money to do so.  Again we have a positive benefit.

The only downside is that the end user purchasing the device can't
install modified versions of the software on it.

Now let's try it in a GPLv3 universe.  Since I can no longer create my
device without having to allow the end user to install modified
software on it I probably turn to some closed source OS like WinCE or
QNX (or maybe I use BSD, but now I can't be bothered to give my
modifications back any longer since their license says I don't have
to).  I'm still happy, I can still sell my device and make money just
like I used to, but:

 - My employees are no longer exposed to Open Source software at work.
 Bummer, no new users from there.

 - I no longer distribute the source for whatever modifications I make
in-house.  Damn, that's some nice software the community is missing
out on.

 - Now my adverticing material is highlighting the use of some
proprietary OS and apps.  What a drag, no more free adverticing for
open source software - actually, quite the opposite.

 - The next time I call my suppliers I just ask them to provide me
whatever closed source drivers they have for
BigCommercialClosedSourceOS and I'm happy since the drivers probably
work just fine.  Hmm, no more pressure on hardware companies to engage
in developing open drivers.

 - Now that I'm paying all this money in licenses for all this
proprietary software and not using any open source software at all,
there's zero to no chance I'll throw any money, developers or whatever
at any open source projects.  Dang, we just lost some corporate
funding.

So, as I see it, tivoization isn't all bad. In fact I think the
positives outweigh the negatives by quite a large margin.
Sure, with GPLv3 you may win the battle and force some manufacturers
away from your software if they can't/won't open their hardware up to
end user modifications. But you'll lose the war in that you'll be
killing the momentum that GPL'ed software has currently since you'll
be driving a lot of players away from it.

I'd rather have the few benefits we get from some company using GPL'ed
software in tivoized hardware than get nothing at all because the
GPLv3 drives that company into the arms of some proprietary vendor.


-- 
Jesper Juhl <jesper.juhl@gmail.com>
Don't top-post  http://www.catb.org/~esr/jargon/html/T/top-post.html
Plain text mails only, please      http://www.expita.com/nomime.html

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Chris Adams]

Once upon a time, Jesper Juhl <jesper.juhl@gmail.com> said:
>Let's say I'm the owner of a company selling some device that uses a
>GPLv2 OS and some GPLv2 applications to do the job. Let's say that for
>some reason I don't want the end users of my device to tinker with the
>software inside my device.  Obviously I release the source for any
>modifications I may have made, but I use the hardware to prevent users
>from installing modified versions on the device (basically I TiVO'ize
>the device).

BTW: Another reason a vendor might lock down the device is for security.
For example, Juniper routers (which now run a significant portion of the
"core" of the Internet) run FreeBSD on the routing engine.  They include
several GNU software utilities (for example gawk, diff, and gdb).
Starting with JUNOS 7.6 (IIRC), end-users can no longer build and run
their own binaries on the routing engine.  This means that the GPLv2
code cannot be modified in-place (similar to TiVo altough done using
different means).

The reason is that if there ever is a security hole in the routing
engine software (FreeBSD kernel, OpenSSH, etc.), it would be a really
bad thing if crackers could load arbitrary software (rootkits, spam
software, etc.) directly on Internet core routers.  If you think spam
zombies on cable modems or DSL are bad, imagine them on 100 megabit
links!

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bron Gondwana]

On Sun, Jun 17, 2007 at 04:58:40PM -0500, Chris Adams wrote:
> Once upon a time, Jesper Juhl <jesper.juhl@gmail.com> said:
> >Let's say I'm the owner of a company selling some device that uses a
> >GPLv2 OS and some GPLv2 applications to do the job. Let's say that for
> >some reason I don't want the end users of my device to tinker with the
> >software inside my device.  Obviously I release the source for any
> >modifications I may have made, but I use the hardware to prevent users
> >from installing modified versions on the device (basically I TiVO'ize
> >the device).
> 
> BTW: Another reason a vendor might lock down the device is for security.
> For example, Juniper routers (which now run a significant portion of the
> "core" of the Internet) run FreeBSD on the routing engine.  They include
> several GNU software utilities (for example gawk, diff, and gdb).
> Starting with JUNOS 7.6 (IIRC), end-users can no longer build and run
> their own binaries on the routing engine.  This means that the GPLv2
> code cannot be modified in-place (similar to TiVo altough done using
> different means).
> 
> The reason is that if there ever is a security hole in the routing
> engine software (FreeBSD kernel, OpenSSH, etc.), it would be a really
> bad thing if crackers could load arbitrary software (rootkits, spam
> software, etc.) directly on Internet core routers.  If you think spam
> zombies on cable modems or DSL are bad, imagine them on 100 megabit
> links!

To be fair here, this could also be accomplished by having to flip a
physical switch on the router, especially if you did something funky
like:

[---] push this button for a 5 minute access pass to upload new
      software through physical cable port 1.

More complex, but not unreasonable.

Bron.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Chris Adams]

Once upon a time, Bron Gondwana <brong@fastmail.fm> said:
> To be fair here, this could also be accomplished by having to flip a
> physical switch on the router, especially if you did something funky
> like:
> 
> [---] push this button for a 5 minute access pass to upload new
>       software through physical cable port 1.
> 
> More complex, but not unreasonable.

Well, there is no restriction on putting files on the routing engine's
storage devices (flash and hard drive); it is running OpenSSH, so
scp/sftp work fine, and you can drop to a shell easily.  The restriction
is that the kernel won't run unsigned binaries.

Also, flipping physical switches is pretty much an unreasonable
expectation for core router operation.  These are often in other
locations, sometimes other telcos' central offices (where you have to
pay to have "remote hands" do something and then hope they don't screw
it up).  You can easily go the entire life of a device where the primary
operators never physically see the device.

-- 
Chris Adams <cmadams@hiwaay.net>
Systems and Network Administrator - HiWAAY Internet Services
I don't speak for anybody but myself - that's enough trouble.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bron Gondwana]

On Sun, Jun 17, 2007 at 07:45:38PM -0500, Chris Adams wrote:
> Once upon a time, Bron Gondwana <brong@fastmail.fm> said:
> > To be fair here, this could also be accomplished by having to flip a
> > physical switch on the router, especially if you did something funky
> > like:
> > 
> > [---] push this button for a 5 minute access pass to upload new
> >       software through physical cable port 1.
> > 
> > More complex, but not unreasonable.
> 
> Well, there is no restriction on putting files on the routing engine's
> storage devices (flash and hard drive); it is running OpenSSH, so
> scp/sftp work fine, and you can drop to a shell easily.  The restriction
> is that the kernel won't run unsigned binaries.
> 
> Also, flipping physical switches is pretty much an unreasonable
> expectation for core router operation.  These are often in other
> locations, sometimes other telcos' central offices (where you have to
> pay to have "remote hands" do something and then hope they don't screw
> it up).  You can easily go the entire life of a device where the primary
> operators never physically see the device.

Every server I run is like that, but if something is important enough I
can remote control a robot over to push the button for me (actually, I
think they implement this under the hood by having a human read the
ticket I submit and go push the button for me manually, but that could
be my imagination.  So long as the button gets pushed the black box is
functioning)

Bron.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Andrea Arcangeli]

On Sun, Jun 17, 2007 at 04:58:40PM -0500, Chris Adams wrote:
> The reason is that if there ever is a security hole in the routing
> engine software (FreeBSD kernel, OpenSSH, etc.), it would be a really
> bad thing if crackers could load arbitrary software (rootkits, spam
> software, etc.) directly on Internet core routers.  If you think spam
> zombies on cable modems or DSL are bad, imagine them on 100 megabit
> links!

Not sure if it's a good example, keep in mind that at the first
exploitable software bug any hardware DRM breaks apart.

But since you made a BSD-embedded example, this shows how the only
really important thing is that by using linux instead of BSD, they
can't make huge improvements or important security bugfixes to the
routing engine, without us being able to incorporate them in our "home
firewalls", that's the whole difference with BSD and it explains the
spirit of the gpl pretty well and in the end why linux by definition
can receive more contributions and in turn be technically
superior. Whatever the vendor does with the gpl code is generally up
to him, and if it uses the closed approach it'll allow somebody else
to sell a "open" router (potentially at an higher price). Economy
101. The worry that nobody will step in and sell an "open" equivalent
is a red herring. Infact I wouldn't be so certain that openmoko would
exist if the current linux cellphones would be already totally open!
Now I know this all probably sounds boring talk, but I think it's much
closer to reality than the prospect of a trusted computing and/or DRM
apocalypse.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 17, 2007, "Jesper Juhl" <jesper.juhl@gmail.com> wrote:

> On 17/06/07, Alexandre Oliva <aoliva@redhat.com> wrote:
> [snip]

>> Serious, what's so hard to understand about:

>> no tivoization => more users able to tinker their formerly-tivoized
>> computers => more users make useful modifications => more
>> contributions in kind

> I have to disagree.

Your analysis stopped at the downside of prohibiting tivoization.  You
didn't analyze the potential upsides, so you may indeed come to
different conclusions, and they may very well be wrong.

It's very human to look only at the potential downside of an action
and conclude it's a bad action.

But it's more rational to look at the potential upside as well,
evaluate the likelihood of each in the grand scheme of things, and
then decide whether the potential upside will make up for the
potential downside.


> Let's say that for some reason I don't want the end users of my
> device to tinker with the software inside my device.

Ok, keep the *want* in mind.  This is very important.


> Now I think you can agree to these things being positive:

Yes, even if I'd phrase them slightly differently.

> The only downside is that the end user purchasing the device can't
> install modified versions of the software on it.

And therefore you severely limit the number of end users who might
turn into contributors because of self interest in hacking the device
to suit their needs.


> Now let's try it in a GPLv3 universe.  Since I can no longer create my
> device without having to allow the end user to install modified
> software on it

False assumption.  You can create the device using GPLv3 software in
it.  So your acccounting of necessary downsides is only one of the
possibilities.  The other possibility would be to have the program in
ROM, of course, which would come with a completely different set of
downsides, but that would retain all of the "these things being
positive" you mentioned above.


And, remember, since you merely don't *want* the end user of the
device to tinker with the software, you have the option to do let them
do that.

And, if you do, they may find in themselves reasons and incentives to
change the software in the device, and the improvements are likely to
get back to the community and thus back to you.  Everybody wins.

This is the upside that you left out from your analysis, and from
every other analysis that set out to "prove" that anti-tivoization is
bad that I've seen so far.

It appears that people are so concerned about whatever little they
might lose from requiring respect for users' freedoms that they don't
even consider what they might win, and that they *would* win if at
least some of the vendors were to make an choice more favorable to
their users and the community.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Jesper Juhl]

On 18/06/07, Alexandre Oliva <aoliva@redhat.com> wrote:
> On Jun 17, 2007, "Jesper Juhl" <jesper.juhl@gmail.com> wrote:
>
> > On 17/06/07, Alexandre Oliva <aoliva@redhat.com> wrote:
> > [snip]
>
> >> Serious, what's so hard to understand about:
>
> >> no tivoization => more users able to tinker their formerly-tivoized
> >> computers => more users make useful modifications => more
> >> contributions in kind
>
> > I have to disagree.
>
> Your analysis stopped at the downside of prohibiting tivoization.  You
> didn't analyze the potential upsides,

Maybe that's because I don't really see any up sides.

As I see it, if we prevent tivoization, then the most likely outcome
will be that a very few number of vendors will switch to ROM based
solutions or similar (everyone lose, both vendor and user), a few
vendors that currently tivoize hardware may open up their hardware but
I doubt that will be very many, and the vast majority of vendors will
move to *BSD or proprietary software since they simply can't or won't
open up their hardware.

So no, I don't think there are any upsides. We'll lose a huge number
of developers, testers and users inside the business comunity and
we'll lose a lot of exposure (like "hey, did you know TiVO actually
runs Linux inside? Isn't that cool?)...  Gaining a few hobyists at the
expense of driving a huge number of businesses away from GPL'ed
software does not look like an upside to me.


>so you may indeed come to
> different conclusions, and they may very well be wrong.
>
Just because I come to a different conclusion than you doesn't
nessesarily make it wrong.

> It's very human to look only at the potential downside of an action
> and conclude it's a bad action.
>
And you believe yourself to be immune to that - right?


> > Let's say that for some reason I don't want the end users of my
> > device to tinker with the software inside my device.
>
> Ok, keep the *want* in mind.  This is very important.
>
No, it is not. When I wrote that I meant "don't want" as in "really
don't want to since it'll destroy our business" or "really really
don't want to since we would be breaking the law" etc.

> > Now I think you can agree to these things being positive:
>
> Yes, even if I'd phrase them slightly differently.
>
> > The only downside is that the end user purchasing the device can't
> > install modified versions of the software on it.
>
> And therefore you severely limit the number of end users who might
> turn into contributors because of self interest in hacking the device
> to suit their needs.
>
Most people don't care about hacking their devices, and of the few who
do only a subset have the skill and only a subset of those will
actually contribute anything back. This is a *small* set of people and
gaining that small set at the expense of losing the large number of
contributers from various companies doesn't make sense to me.


> > Now let's try it in a GPLv3 universe.  Since I can no longer create my
> > device without having to allow the end user to install modified
> > software on it
>
> False assumption.  You can create the device using GPLv3 software in
> it.

Not as long as I want to prevent the user from tampering with it, no.

>So your acccounting of necessary downsides is only one of the
> possibilities.  The other possibility would be to have the program in
> ROM, of course, which would come with a completely different set of
> downsides, but that would retain all of the "these things being
> positive" you mentioned above.
>
But do you really expect a vendor to put a device on the market where
they also lock themselves out of upgrading it and releasing new
software for it?  That's just rediculous.


> And, remember, since you merely don't *want* the end user of the
> device to tinker with the software, you have the option to do let them
> do that.
>
See above.

> And, if you do, they may find in themselves reasons and incentives to
> change the software in the device, and the improvements are likely to
> get back to the community and thus back to you.  Everybody wins.
>
For a few select individuals that may be true. But for the majority of
the population it won't mean a thing.

> This is the upside that you left out from your analysis, and from
> every other analysis that set out to "prove" that anti-tivoization is
> bad that I've seen so far.
>
I'm sorry, but I don't think it holds water.

> It appears that people are so concerned about whatever little they
> might lose from requiring respect for users' freedoms that they don't
> even consider what they might win, and that they *would* win if at
> least some of the vendors were to make an choice more favorable to
> their users and the community.

Contrary to you, I don't believe any significant number of companies
will do that. It's simply better for business to just use other
software in that case.

-- 
Jesper Juhl <jesper.juhl@gmail.com>
Don't top-post  http://www.catb.org/~esr/jargon/html/T/top-post.html
Plain text mails only, please      http://www.expita.com/nomime.html

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 20, 2007, "Jesper Juhl" <jesper.juhl@gmail.com> wrote:

> On 18/06/07, Alexandre Oliva <aoliva@redhat.com> wrote:
>> Your analysis stopped at the downside of prohibiting tivoization.  You
>> didn't analyze the potential upsides,

> Maybe that's because I don't really see any up sides.

You do:

> a few vendors that currently tivoize hardware may open up their
> hardware but I doubt that will be very many

You just don't think they'd prevail over the downsides.  *This* is an
opinion I can respect, even if it's as much of a guestimate as mine.
I'm sure both are highly influenced by personal opinions, wishful
thinking and fears.  This is all very human.

>> so you may indeed come to different conclusions, and they may very
>> well be wrong.

> Just because I come to a different conclusion than you doesn't
> nessesarily make it wrong.

Agreed.  I didn't say they were.  I said they could be.  Can you prove
they're right?  Do you even have any supporting evidence to back your
guestimates?  Heck, you may even have more than I do.

I openly admit mine is mostly theoretical.  I extrapolate the initial
success of GNU+Linux on the PC environment, due in a large part to the
ability for users to tinker with their computers, and expect it not to
be so significantly different for other kinds of computers.

For sure you'll get a far lower *percentage* of hackers in consumer
devices than on PCs, whose users used to be far more
technically-inclined and thus more propense to become hackers when
GNU+Linux started than these days.

But then I think of all of these computer users who helped make
GNU+Linux what it is today, and other hackers that hadn't discovered
this inclination before because they haven't had access to hackable
computers.  They could be tinkering with their DVRs, cell phones,
wireless routers et al, and bringing the same kind of exciting
community development to these kinds of computers.

I'm saddened that the major Linux developers are willing to trade all
of this (which I openly admit may be just a figment of my imagination,
or just a tip of an iceberg) for some professional contributions
(good) and some additional exposure that won't do justice to their
software (bad), because these users will miss a big part of the
picture by not being able to tinker with the software in the
environment where they use the software.

>> It's very human to look only at the potential downside of an action
>> and conclude it's a bad action.

> And you believe yourself to be immune to that - right?

Last I looked, I was still human.  So no.  I try to use logic to
reason out such behaviors when I realize they might be in action.
But, as the saying goes, logic is a tool we use to justify our
intutions.  Or, logical reasoning is a tool to make the wrong
decisions with a greater amount of confidence ;-)

>> You can create the device using GPLv3 software in it.

> Not as long as I want to prevent the user from tampering with it, no.

<broken record>mumble ROM mumble</broken record>

> But do you really expect a vendor to put a device on the market where
> they also lock themselves out of upgrading it and releasing new
> software for it?

Depends on how badly they want to use the GPLed software.

Don't you guys think Linux is so technically superior that some
vendors might prefer to stick with it (should it move to GPLv3, or
tivoization be found to be already forbidden by GPLv2 in a US court or
elsewhere) even if this means going to ROM or respecting users'
freedoms?

Or are Linux advantages so thin and fragile (if they exist at all)
that you're just hoping nobody realizes there are better choices out
there, and you're desperate for vendors not to realize this?

> For a few select individuals that may be true. But for the majority of
> the population it won't mean a thing.

Agreed.  You're thinking of percentages (fewer percent hackers among
consumers of user products than among PC users).  I'm thinking all
hackers in PCs could become hackers of such devices as well, and then
some.

>> This is the upside that you left out from your analysis, and from
>> every other analysis that set out to "prove" that anti-tivoization is
>> bad that I've seen so far.

> I'm sorry, but I don't think it holds water.

Fair enough.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Sun, 17 Jun 2007, Alexandre Oliva wrote:
> > A third way of stating it is "software for software". No, the romans never 
> > said that, but I just did, to make it just more obvious that the whole 
> > point is that you are expected to answer IN KIND!
> 
> Yes.  And this was precisely what meant when I wrote "quid-pro-quo"
> above.

Ok, so we're on the same page.

"software for software".

> > 	"You are either terminally stupid, or you're not allowing yourself 
> > 	 to see an obvious argument, because it destroys your world-view".
> > 
> > The latter is very possible. It's a very human thing.
> 
> /me hands Linus a mirror

"I'm a damn handsome dude, ain't I?"

> Serious, what's so hard to understand about:

You're talking about something totally different.

Answer my argument:

 - I think Tivoization is *good*.

 - Your license stops something *good*.

Ergo:

 - Your license is *bad*.

> Wrong.  It enables copyright holders to decide whether forgiveness is
> appropriate, rather than forcing them to forgive.  Being forced to
> forgive deception is not tit-for-tat, and it's a losing strategy.

There is NOTHING TO FORGIVE!

Your whole idiotic argument misses the point:

	What Tivo did is *good* in my opinion!

Can't you get that through your skull?

They gave the software back! Be happy! They *followed* the rules. They 
*followed* the tit-for-tat.

			Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 17, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:

> On Sun, 17 Jun 2007, Alexandre Oliva wrote:

>> /me hands Linus a mirror

> "I'm a damn handsome dude, ain't I?"

Heh.  I beg to differ ;-)

>> Serious, what's so hard to understand about:

> You're talking about something totally different.

No, I'm not.  You can say tivoization is *good* however much you like.
This doesn't dispute in any way my claim that no tivoization would be
*better*, that you'd get contributions from the people that, because
of tivoization, don't feel compelled to develop and contribute,
because they can't use the fruits of their efforts in the device where
they would be most useful for them.

> 	What Tivo did is *good* in my opinion!

> Can't you get that through your skull?

No.  I disagree.  We can agree to disagree on that.

> They gave the software back! Be happy! They *followed* the rules. They 
> *followed* the tit-for-tat.

But they removed incentive for far more users to do so.  So you get
fewer contributions than you could without tivoization.

"Can't you get that through your skull?" :-)

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Sun, 17 Jun 2007, Alexandre Oliva wrote:
> > 	What Tivo did is *good* in my opinion!
> 
> > Can't you get that through your skull?
> 
> No.  I disagree.  We can agree to disagree on that.

Sorry, no we cannot. You seem to not accept that "in my opinion".

That's not somethign we can disagree on. My opinion is _mine_ to state. 
There's no room for disagreement. It's my opinion, and your "agreement" is 
not optional.

So when I say "I think what Tivo did is good", then you'd better just say, 
"ok, that's your opinion, and I respect you for it".

Otherwise you're a douche-bag and an idiot.

And once you realize that _I_ think that Tivo did is good, you have to 
accept that I think that the GPLv3 is the worse license in my opinion.

So stop blathering about anything else.

Just accept it. Just repeat after me: "Linus thinks that the GPLv3 is a 
bad license, and Linus is not confused".

Don't call me "confused". Don't bother talking about what _you_ think, or 
what the FSF thinks is the "spirit". Don't say that you cannot understand 
it. 

Because if you cannot understand it, the only thing it shows is _your_ 
lack of understanding.

		Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 18, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:

> On Sun, 17 Jun 2007, Alexandre Oliva wrote:
>> > 	What Tivo did is *good* in my opinion!

>> > Can't you get that through your skull?

>> No.  I disagree.  We can agree to disagree on that.

> Sorry, no we cannot. You seem to not accept that "in my opinion".

Sorry.  I stand corrected.  I didn't mean to disagree with the "in my
opinion".  I guess I was too distracted by the beauty of your kind
words ;-)  Sorry, it takes some effort to take focus away from that.


I agree that this is your opinion, and you're entitled to it, no
matter how much I disagree with this opinion.

I disagree that what TiVo did is good.  I think it is wrong on ethical
and moral grounds, and I think it is bad for the user, for the
community, and quite possibly even for TiVo itself.

We can agree to disagree as to our opinions, if you want.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Mon, 18 Jun 2007, Alexandre Oliva wrote:
> 
> We can agree to disagree as to our opinions, if you want.

That's all I ever asked for.

This whole thread started with me saying:

  I see the smiley, but I hate it how the FSF thinks others are morons and
  cannot read or think for themselves.

  Any time you disagree with the FSF, you "misunderstand" (insert
  condescending voice) the issue.

  _Please_ don't continue that idiocy. Disagreement and thinking that the
  FSF is controlling and putting its fingers where they don't belong is
  _not_ misunderstanding. It's just not "blind and unquestioning obedience".

so all I asked for in the first place was that you stop claiming that I 
had "misunderstood" anything.

That's really all I've always asked for:

 - I chose the GPLv2, and I understand it.

 - you don't have to agree with my choice, but you *do* have to accept it
   if you want to work on Linux. Because it's the only license that Linux 
   has ever been released under since early 1992.

So as long as you follow the GPLv2 (as a _legal_ license), I don't care if 
you like it or not. I don't care if you think you are a modern-day 
Napoleon, or if you are a demented squirrel. I don't care if you are an 
axe-murderer, or if you make sex toys with Linux. I don't care if your 
hardware is open or closed.

I care about one thing, and one thing only: I care that you respect my 
choice of license for the projects _I_ started. Nothing more.

And it doesn't matter one whit if *you* would have made a different 
choice. You are not me. You don't hold any power over me, and *your* 
choices are your own - not mine.

Choice of license is personal. Many people think that the BSD license is 
better than _any_ version of the GPL. Are they wrong? No, it's _their_ 
choice. Is it relevant for the kernel? No, their preference of license is 
simply irrelevant. They can choose to accept the license that the kernel 
is under, or go play somewhere else.

I think the GPLv2 is superior to the GPLv3.  That is simply not something 
you can argue against. You can just say "ok, it's your choice". You can 
ask me *why*, and I've told you at length, but in the end, it doesn't 
matter. 

And no, it's not because I'm "special", and I get to make all decisions. 
It's simply because I am _me_, and when it comes to my own opinions, I 
actually _do_ get to make all the decisions.

You can disagree, and choose to use the GPLv3. You just cannot do it for 
the *kernel*, because they kernel has always been under the GPLv2, and the 
GPLv3 is simply not compatible, and asks for things that the kernel 
license has never asked for.

But if you prefer the GPLv3, that's _your_ choice, and that choice can 
certainly guide you in the licensing of _your_ projects where _you_ are 
the copyright holder. And I will never complain.

		Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 18, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:

> I care about one thing, and one thing only: I care that you respect my 
> choice of license for the projects _I_ started. Nothing more.

I do.  Really.

Once the issue about the spirit of the GPL is (hopefully) settled with
all concerned about it, my job would have been done if it hadn't been
for my having got interested in this other issue:

> I think the GPLv2 is superior to the GPLv3.  That is simply not something 
> you can argue against. You can just say "ok, it's your choice". You can 
> ask me *why*, and I've told you at length, but in the end, it doesn't 
> matter. 

Let me explain why I don't see that you've told me at length why you
consider GPLv2 superior to GPLv3.

1. I asked you why GPLv2 is better, and you said it was because it
promoted giving back in kind.

2. I asked you what you didn't like about GPLv3, and you said it was
Tivoization.

3. Then I argued that, since Tivoization enables tivoizers to remove
some motivation for potential developers (= their customers) to
contribute, you trade the potential contributions of all those users
for the contributions of tivoizers, apparently assuming that all
tivoizers would simply move away from the community, taking their
future contributions away from your community, rather than moving to a
position in which you'd get not only the contributions from the
company itself, but also from all their users.

This last piece of the theorem that proves that GPLv2 is more aligned
with your stated goals than GPLv3 is the one that is missing, and so
far you've dodged that portion entirely.  That's the 'connecting the
dots' that I mentioned earlier.  You haven't even acknowledged its
existence, going back to points 1. and 2. as if they were enough, as
if 3. didn't show a contradiction between them.

Now, it may be that 3. is wrong, or that you think it is wrong.  But
you've never said so, or explained why you think so.  You've simply
disregarded that point entirely.

Do you understand now why I feel you haven't answered the 'why'?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Mon, 18 Jun 2007, Alexandre Oliva wrote:
> 
> 1. I asked you why GPLv2 is better, and you said it was because it
> promoted giving back in kind.

Where I explained that "in kind" was about *software*.

> 2. I asked you what you didn't like about GPLv3, and you said it was
> Tivoization.

Right. The GPLv3 asks you to give back *money*.

That's like the Microsoft license agreements. I don't like them either.

Oh, and replace "money" with "access to hardware", to make that thing 
technically correct. But the point is, that's what I don't like about the 
GPLv3.

> 3. Then I argued that, since Tivoization enables tivoizers to remove
> some motivation for potential developers (= their customers)

That's simply not my *reason* for doing "tit-for-tat". My basic reason for 
"tit-for-tat" was not about "lots of potential developers", but simply 
because I think it's the right choice for me!

Can you not understand that? I simply DO NOT LIKE TO CONTROL PEOPLE!

I just want software back. I think it is *wrong* for me to ask for 
anything else. It's literally my personal "moral choice": I think the 
hardware manufacturers need to make their _own_ choices when it comes to 
_their_ designs.

I feel that I have the moral right to ask for modifications to the kernel 
(because I started it), but I *personally* am very unhappy about asking 
people to also give their hardware access. That's *their* choice.

Is that really so hard to understand? I ask you to respect _my_ choice wrt 
license for my software, but the same way I expect others to respect my 
choices, I also myself need to respect *their* choices. 

So to me, it's the hardware manufacturers choice to to select the license 
for their hardware, exactly the same way it was *my* choice to select it 
for my software. I believe in basically *one* freedom: the freedom to make 
our own choices!

But if you actually want to discuss "number of developers" and their 
motications, I actually have another few arguments for you:

 - I just personally think your math is bogus. I think more people think 
   like I do, than people think like you and the FSF does.

But I don't even depend on that. Because:

 - I think that *technical*quality* is more important than *quantity*.

   And I think you have already proven a point: the GPLv3 seems to attract 
   people who make the wrong *technical* decisions.

Put another way: I'll much rather attract one Al Viro to the project, than 
a hundred rabid FSF followers.

See? Because I think that one Al Viro will make *more* of a difference 
than a hundred people who think that their *religion* is more important 
than making the technically correct choice is!

With the GPLv2, you need to give your software modifications back, but the 
GPLv2 never *ever* makes any technical limitations on the end result.

In the GPLv3 world, we have already discussed in this thread how you can 
follow the GPLv3 by making the TECHNICALLY INFERIOR choice of using a ROM 
instead of using a flash device.

Quite frankly, I don't *want* to attract develpers that are not 
technically "up to snuff". And if you think that making the technically 
worse decisions is the "rigth decision", then hey, you're clearly not in 
the same technical quality range as I am, or Al Viro is.

Am I elitist? HELL YES! I think some people are simply *better* engineers 
than other people. I've met my share of outstanding engineers, and I've 
met average engineers. 

I am firmly of the opinion that one of the signs of an outstanding 
engineer is making the right technical choices. The GPLv2 is ok with that. 
The GPLv3 is not. The GPLv3 makes *limits* what you can do from a 
technical angle, in a way that the GPLv2 does not.

The GPLv2 requires that you give source code out. But if you want to make 
your hardware in a way that it only runs signed versions, because of some 
reason like an FCC rule, or banking rule, or just because you damn well 
want, the GPLv2 doesn't stop that.

The GPLv3 doesn't stop it *either*, but the GPLv3 requries that you make 
the INFERIOR TECHNICAL CHOICE.

In other words: the GPLv3 is for people who care more about the opinions 
of the FSF than about the technology.

And why the hell should I trust people like that to make the right 
technical choices in *other* matters? They have already shown themselves 
to make bad technical choices.

> Do you understand now why I feel you haven't answered the 'why'?

Ok, so now I have. I have three *different* and independent answers for 
you:

 a) I don't personally feel like I have the "moral authority" to require
    hardware designers to give access to their hardware to me.

    I can tell them that I *like* open hardware more than closed hardware, 
    but they designed the hardware, and as a result I think it's *their* 
    choice.

    In contrast, I _do_ think I have the moral authority to ask for 
    modifications to the _software_ back. Because they didn't design it, 
    they just improved on it.

 b) I think you're simply wrong in your math. I think more people like the 
    middle-ground and not-frothing-at-the-mouth spirit of "open source" 
    over the religious dogma of "free software".

    I think Linux has pretty much proved my point. Look at Hurd, then look 
    at Linux. Am I *that* much better than the Hurd developers (yes, of 
    course I am, but let's assume not). Or is it just that my approach of 
    being more _pragmatic_ about things rather than talking about those 
    "four freedoms" all the time was just much easier for people to 
    accept?

 c) Even if you're not wrong in the math, I've seen the kind of people who 
    argue for it, and quite frankly, I think they are making bad technical 
    decisions. You arguing for a ROM over a flash is an excellent example. 

    You seem to never even have given a second _thought_ to the fact that 
    you actually advocated what is technically the inferior choice.

See? Those are three totally different reasons why I think the GPLv2 is 
the right license for me, and for the kernel.

			Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Mon, 18 Jun 2007, Linus Torvalds wrote:
> 
>  a) I don't personally feel like I have the "moral authority" to require
>     hardware designers to give access to their hardware to me.
> 
>     I can tell them that I *like* open hardware more than closed hardware, 
>     but they designed the hardware, and as a result I think it's *their* 
>     choice.
> 
>     In contrast, I _do_ think I have the moral authority to ask for 
>     modifications to the _software_ back. Because they didn't design it, 
>     they just improved on it.

Btw, this may explain the differences between rms/FSF and me.

Richard Stallman is out to "improve the world". He thinks he has not just 
the moral authority and right, but the _obligation_ to tell others how to 
live their lives.

So rms feels he has to "spread his message", the same way those annoying 
people come to your door on Saturday in order to "save you". rms literally 
is saving you from a sin.

Me, not so much. I'm not out to "improve the world". I'm not saving you 
from any sin. I just have this project I started, and I think that if you 
improve on the project, you should let me use your improvements.

And I don't ask you use that license anywhere else. For your other 
projects, you should follow *their* licenses. I'm just asking you to honor 
my choice of license for those projects I started, and I designed, and 
where I conned other people into helping me under the same rules.

				Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Al Viro]

On Mon, Jun 18, 2007 at 02:09:13PM -0700, Linus Torvalds wrote:

> Quite frankly, I don't *want* to attract develpers that are not 
> technically "up to snuff". And if you think that making the technically 
> worse decisions is the "rigth decision", then hey, you're clearly not in 
> the same technical quality range as I am, or Al Viro is.

You are confusing being generally fucked in head (aka being a true
believer) with being willing to go for technically worse decisions
*when* *it* *comes* *to* *making* *them*.  Which would be in gcc-related
work in this case.  I don't see any evidence of the latter...

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 18, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:

> On Mon, 18 Jun 2007, Alexandre Oliva wrote:
>> 
>> 1. I asked you why GPLv2 is better, and you said it was because it
>> promoted giving back in kind.

> Where I explained that "in kind" was about *software*.

Yes, we'd already established that.

>> 2. I asked you what you didn't like about GPLv3, and you said it was
>> Tivoization.

> Right. The GPLv3 asks you to give back *money*.

> That's like the Microsoft license agreements. I don't like them either.

> Oh, and replace "money" with "access to hardware", to make that thing 
> technically correct.

technically, it asks you to pass on (!= give back) access to the
software (not to the hardware that contains it).

>> 3. Then I argued that, since Tivoization enables tivoizers to remove
>> some motivation for potential developers (= their customers)

> That's simply not my *reason* for doing "tit-for-tat".

No.  The reason, again, is the portion you snipped out.

Could you try again?

> I just want software back. I think it is *wrong* for me to ask for 
> anything else. It's literally my personal "moral choice": I think the 
> hardware manufacturers need to make their _own_ choices when it comes to 
> _their_ designs.

It's comforting to see that you're not the pure-pragmatics-no-morals
character that some people (yourself included) try to paint you as.
Thank you for this.  I can relate with that.  I can easily respect
that, as much as I think that (poor attempt at humor follows, no
offense really intended) standing up for the freedoms of the poor
hardware manufacturers against the evil software developers who want
to control the ways they can use to control their customers serves the
common good or even your own stated interests.

> But if you actually want to discuss "number of developers" and their 
> motications, I actually have another few arguments for you:

>  - I just personally think your math is bogus. I think more people think 
>    like I do, than people think like you and the FSF does.

> But I don't even depend on that. Because:

>  - I think that *technical*quality* is more important than *quantity*.

This argument fails to make the point you're trying to make.  I wrote:

  you trade the potential contributions of all those users for the
  contributions of tivoizers, apparently assuming that all tivoizers
  would simply move away from the community, taking their future
  contributions away from your community, rather than moving to a
  position in which you'd get not only the contributions from the
  company itself, but also from all their users

and you say "oh, I don't care about quantity, I care about quality",
as if this somehow related with the above.

Just do the math.  Hypothetically, Linux goes GPLv3, without
permission to tivoize.  TiVo has to decide among:

- switching to another kernel, no further contributions from them

- sticking with old version, no further usable contributions from them

- switching to ROM, still the same contributions from TiVo

- no more tivoization, contributions from TiVo and users

So, you see, in no case do you get more contributors while at the same
time losing TiVo's quality contributions.

The argument is not about quality vs quantity, it's about getting
lots of additional contributions along with what's already in place,
VS ending up without some quality contributions you get today.

> With the GPLv2, you need to give your software modifications back, but the 
                      ^^^^  BZZT!
> GPLv2 never *ever* makes any technical limitations on the end result.

Actually, just think of how many times you've heard the argument "I
can't give you the source code for this driver/firmware/etc under the
GPLv2 because the law says so."

> In the GPLv3 world, we have already discussed in this thread how you can 
> follow the GPLv3 by making the TECHNICALLY INFERIOR choice of using a ROM 
> instead of using a flash device.

Yes.  This is one option that doesn't bring any benefits to anyone.
It maintains the status quo for users and the community, but it loses
the ability for the vendor to upgrade, fix or otherwise control the
users.  Bad for the vendor.

As another option, the vendor can respect users' freedoms, and then
everybody wins big.  That's the option that anti-tivoization provides
economic incentive for vendors to take.  Sure, they may still prefer
the alternative above, or stick with an older version (which has its
costs), or move to different software (which also has its costs), but
it's unreasonable to claim that I'm advocating for vendors to move to
ROM.

I'm saying they have this option.  I'm advocating for them to respect
users' freedom.  And if that's incompatible with their business model,
well, so what?  GPLv2 and Free Software in general are incompatible
with a number of business models too, and who's complaining?  Heck,
even using slave work-forces was part of legitimate business models at
some point in time.

> The GPLv2 requires that you give source code out.
            ^^^^^^^^ BZZT ;-)
> But if you want to make your hardware in a way that it only runs
> signed versions, because of some reason like an FCC rule, or banking
> rule, or just because you damn well want, the GPLv2 doesn't stop
> that.

And then, the user is stopped from making appropriate technical
decisions.

> The GPLv3 doesn't stop it *either*, but the GPLv3 requries that you
> make the INFERIOR TECHNICAL CHOICE.

This is a lie (by which I don't mean it's malicious).

What it requires is the vendor to decide between making the inferior
technical choice and respecting users' freedoms to make their own
technical choices.

> b) I think you're simply wrong in your math. I think more people
> like the middle-ground and not-frothing-at-the-mouth spirit of "open
> source" over the religious dogma of "free software".

It looks like the math you're talking about is in no way related with
anything I've argued about.  You seem to be thinking about the number
of people who claim to be on the "free software" or "open source"
sides, but I can't fathom in what way this is related with whether you
get more or less contributions from users as a consequence of users'
being permitted to tinker with the free software in their own devices.

>    I think Linux has pretty much proved my point. Look at Hurd, then
>    look at Linux. Am I *that* much better than the Hurd developers
>    (yes, of course I am, but let's assume not). Or is it just that
>    my approach of being more _pragmatic_ about things rather than
>    talking about those "four freedoms" all the time was just much
>    easier for people to accept?

I could argue that Hurd took a different approach that proved to be
far more difficult, and that the urgency for the development of a Free
Software kernel by the GNU project disappeared with the relicensing of
Linux under a Free Software license (thanks!).  But I guess you'll
just dismiss that on whatever reasons move you, and I don't really
care about these particular historical issues to spend time discussing
them.

> See? Those are three totally different reasons why I think the GPLv2 is 
> the right license for me, and for the kernel.

Ok, the only one that stands is the moral reason.  That's a good one,
but it contradicts the stated reasons as to why you prefer GPLv2 over
GPLv3, stating it was a purely pragmatic decision, based on getting
more software contributions back, based on an assumption, that so far
lacks any material evidence, that permission for tivoization somehow
gets you more of that than getting just as many contributions from the
former-tivoizer, plus whatever any of their users decide to
contribute.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Monday 18 June 2007 19:31:30 Alexandre Oliva wrote:
> On Jun 18, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:
<snip>
>
> > With the GPLv2, you need to give your software modifications back, but
> > the
>                                    ^^^^  BZZT!
> > GPLv2 never *ever* makes any technical limitations on the end result.
>
> Actually, just think of how many times you've heard the argument "I
> can't give you the source code for this driver/firmware/etc under the
> GPLv2 because the law says so."

Sorry to tell you this, but anyone that makes a modification to GPLv2 covered 
code and distributes that modification is bound by the license. If a law 
makes following the license illegal, then they can't use any rights granted 
by the license. They are breaking the law by refusing to follow the license.

<snip>
> > The GPLv2 requires that you give source code out.
>                  ^^^^^^^^ BZZT ;-)
> > But if you want to make your hardware in a way that it only runs
> > signed versions, because of some reason like an FCC rule, or banking
> > rule, or just because you damn well want, the GPLv2 doesn't stop
> > that.
>
> And then, the user is stopped from making appropriate technical
> decisions.

You marked the "requires" as an error. Technically it is. Practically, 
however, it is rare for a modification to not fall under the "distribution" 
part of the license, making the "release the source" requirement active 
almost all the time.

<snip>
> > b) I think you're simply wrong in your math. I think more people
> > like the middle-ground and not-frothing-at-the-mouth spirit of "open
> > source" over the religious dogma of "free software".
>
> It looks like the math you're talking about is in no way related with
> anything I've argued about.  You seem to be thinking about the number
> of people who claim to be on the "free software" or "open source"
> sides, but I can't fathom in what way this is related with whether you
> get more or less contributions from users as a consequence of users'
> being permitted to tinker with the free software in their own devices.

"More Developers" (either "Free Software" or "Open Source") == "More 
Contributions"

That equation is very simple to understand - claiming its wrong is impossible.
<snip>
> > See? Those are three totally different reasons why I think the GPLv2 is
> > the right license for me, and for the kernel.
>
> Ok, the only one that stands is the moral reason.  

Apparently because you can't admit that a good reason *IS* a good reason when 
it conflicts with your belief that the FSF is correct. (The same as 
the "Science can't be right because it conflicts with the bible" I hear from 
all kinds of Xtians these days)

DRH
PS: I know I've said I'm done with this conversation, but this is like a bad 
habit. I just couldn't help myself.

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 18, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> On Monday 18 June 2007 19:31:30 Alexandre Oliva wrote:
>> On Jun 18, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:

>> Actually, just think of how many times you've heard the argument "I
>> can't give you the source code for this driver/firmware/etc under the
>> GPLv2 because the law says so."

> Sorry to tell you this, but anyone that makes a modification to GPLv2 covered 
> code and distributes that modification is bound by the license.

Of course I know that.  I'm not the one making those arguments.
And then, not all of those pieces of code are indeed moficiations of
GPLv2-covered code, so your objection is off target.

>> > b) I think you're simply wrong in your math. I think more people
>> > like the middle-ground and not-frothing-at-the-mouth spirit of "open
>> > source" over the religious dogma of "free software".
>> 
>> It looks like the math you're talking about is in no way related with
>> anything I've argued about.  You seem to be thinking about the number
>> of people who claim to be on the "free software" or "open source"
>> sides, but I can't fathom in what way this is related with whether you
>> get more or less contributions from users as a consequence of users'
>> being permitted to tinker with the free software in their own devices.

> "More Developers" (either "Free Software" or "Open Source") == "More 
> Contributions"

> That equation is very simple to understand - claiming its wrong is impossible.

YES!  Thank you!  This is exactly the point I'm trying to make.

Now can you please explain this to Linus in terms that his brain won't
dismiss as "coming from a fundamentalist"?

> Apparently because you can't admit that a good reason *IS* a good reason when 
> it conflicts with your belief that the FSF is correct.

No, seriously.  Linus is disputing the equation above, dismissing my
various attempts to show it to him, whenever it appears in teh context
of tivoization, apparently because it doesn't match his moral belief
that tivoization ought to be permitted on his moral grounds.

> PS: I know I've said I'm done with this conversation, but this is like a bad 
> habit. I just couldn't help myself.

You've helped me a lot while at that.  Thanks!

I hope this helps others fundamentalist anti-fundamentalists :-) see
reason too.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Monday 18 June 2007 22:06:57 Alexandre Oliva wrote:
> On Jun 18, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> > On Monday 18 June 2007 19:31:30 Alexandre Oliva wrote:
> >> On Jun 18, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:
> >>
> >> Actually, just think of how many times you've heard the argument "I
> >> can't give you the source code for this driver/firmware/etc under the
> >> GPLv2 because the law says so."
> >
> > Sorry to tell you this, but anyone that makes a modification to GPLv2
> > covered code and distributes that modification is bound by the license.
>
> Of course I know that.  I'm not the one making those arguments.
> And then, not all of those pieces of code are indeed moficiations of
> GPLv2-covered code, so your objection is off target.

I had a parsing error with your statement. My mind made the jump to "they are 
doing it with GPLv2 code" - meaculpa.

> >> > b) I think you're simply wrong in your math. I think more people
> >> > like the middle-ground and not-frothing-at-the-mouth spirit of "open
> >> > source" over the religious dogma of "free software".
> >>
> >> It looks like the math you're talking about is in no way related with
> >> anything I've argued about.  You seem to be thinking about the number
> >> of people who claim to be on the "free software" or "open source"
> >> sides, but I can't fathom in what way this is related with whether you
> >> get more or less contributions from users as a consequence of users'
> >> being permitted to tinker with the free software in their own devices.
> >
> > "More Developers" (either "Free Software" or "Open Source") == "More
> > Contributions"
> >
> > That equation is very simple to understand - claiming its wrong is
> > impossible.
>
> YES!  Thank you!  This is exactly the point I'm trying to make.
>
> Now can you please explain this to Linus in terms that his brain won't
> dismiss as "coming from a fundamentalist"?

No need. Linus already understands the equation, and also the secondary fact 
that most home users are not developers. However, companies like TiVO do 
employ developers. This is why the equation works.

> > Apparently because you can't admit that a good reason *IS* a good reason
> > when it conflicts with your belief that the FSF is correct.
>
> No, seriously.  Linus is disputing the equation above, dismissing my
> various attempts to show it to him, whenever it appears in teh context
> of tivoization, apparently because it doesn't match his moral belief
> that tivoization ought to be permitted on his moral grounds.

Actually you are in error here. You are saying "More home users == More 
Developers" when the ratio of home users to developers isn't all that high. 
(small set of facts: "Hacker" == "Developer" (in most cases, where the term, 
as defined in the Jargon File, can actually be applied), "Home User" * 0.10 
(ie: 10%) == "Developer" (approximately, and the correlation may be 
lower). "TiVO" == "Developers" (note the plural - they do employ more than 
one person for development))

So "TiVO", even though they are walking all over the freedoms you love, means 
more *guaranteed* developers than the potential pool from the users of their 
boxes. (the pool of potential developers among the millions of TiVO users is 
actually miniscule, despite the size of the sample)

However, you do make a good argument. But when you look at the statistics[1] 
they don't hold water.

> > PS: I know I've said I'm done with this conversation, but this is like a
> > bad habit. I just couldn't help myself.
>
> You've helped me a lot while at that.  Thanks!
>
> I hope this helps others fundamentalist anti-fundamentalists :-) see
> reason too.

I love that phrase!

But seriously, all I did was stop trying to give fully reasoned counters 
(complete with examples) and state the simple truth.

DRH
[1] "There are three types of lies - lies, damn lies and statistics" - 
Attribution uncertain (Benjamin Disraeli, Mark Twain and Alfred Marshal are 
all said to have issued this famous quotation)

PS: I've beaten the addiction! This post was to clarify some things that were 
either misunderstood or stood a chance of being twisted to mean something 
other than what I intended. (not that anyone did the latter)

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 19, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> Actually you are in error here. You are saying "More home users == More 
> Developers" when the ratio of home users to developers isn't all that high. 
> (small set of facts: "Hacker" == "Developer" (in most cases, where the term, 
> as defined in the Jargon File, can actually be applied), "Home User" * 0.10 
> (ie: 10%) == "Developer" (approximately, and the correlation may be 
> lower). "TiVO" == "Developers" (note the plural - they do employ more than 
> one person for development))

As I wrote in another e-mail, why makes this proportion different for
the other conditions determined by the GPL?

I.e., what is it that makes this particular condition so allegedly
harmful for bringing in more developers and contributions, when
compared with the requirements on passing on source code, licensing
necessary patents, not suing other users over patent infringement in
the software, not invoking anti-circumvention laws, not entering
discriminatory agreements?

> So "TiVO", even though they are walking all over the freedoms you love, means 
> more *guaranteed* developers than the potential pool from the users of their 
> boxes. (the pool of potential developers among the millions of TiVO users is 
> actually miniscule, despite the size of the sample)

> However, you do make a good argument. But when you look at the statistics[1] 
> they don't hold water.

Err...  I have no idea of the actual user base of TiVo, but if it's
really in the millions, and your 10% figure above is right, this makes
for hundreds of thousands of hackers that could be scratching their
itches and improving Linux on TiVo boxes.

How many thousand employees does TiVo have working on Linux?

(I realize a full-time employee is a lot more than a Joe Random
Hacker, but still, I'm keeping a ratio of 100:1 to make up for that)

> PS: I've beaten the addiction!

Good for you!

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Mon, 18 Jun 2007, Alexandre Oliva wrote:
>
> > "More Developers" (either "Free Software" or "Open Source") == "More 
> > Contributions"
> 
> No, seriously.  Linus is disputing the equation above, dismissing my
> various attempts to show it to him, whenever it appears in teh context
> of tivoization, apparently because it doesn't match his moral belief
> that tivoization ought to be permitted on his moral grounds.

No. Linus is not AT ALL disputing the equation above.

But you are too f*cking stupid to admit that I *accepted* the

 - "More developers" == "More contributions" == good

equation, but I was claiming that your *other* part was totally broken.

You try to claim that the GPLv3 causes "More developers", and that, my 
idiotic penpal, is just crazy talk that you made up. 

But since you cannot follow a logical argument, and cannot make one up 
on your own, you instead make up some *other* argument, and try (like 
above) to try to say that I made that claim.

The GPLv2 is the one that allows more developers. 

The GPLv2 is the one that is acceptable to more people. 

Face it, the "open source" crowd is the *bigger* crowd. The FSF crowd is 
vocal and opinionated, but it's largely made up of people who _talk_ more 
than they actually _code_. 

Hot air doesn't make the world go round. Real code does.

Look at the kernel developers who claim that the GPLv2 is better. Not just 
me. Then look at the people who actually GET THINGS DONE. 

There's a big overlap there.

Now, look at the people who try to sell the GPLv3 as the best thing since 
sliced bread. How many of those are people who actually get things *done*?

I haven't really seen a single one. Last I did the statistic, I asked the 
top ~25-30 kernel developers about their opinion. NOT A SINGLE ONE 
preferred the GPLv3.

So I have actual *numbers* on my side. What do you have, except for a 
history of not actually understanding my arguments?

		Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Al Viro]

On Mon, Jun 18, 2007 at 08:46:44PM -0700, Linus Torvalds wrote:
 
> So I have actual *numbers* on my side. What do you have, except for a 
> history of not actually understanding my arguments?

Why do I suddenly have an image of Palin as Ximenez doing the answer?

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 19, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:

> The GPLv2 is the one that allows more developers. 

> The GPLv2 is the one that is acceptable to more people. 

Based on my understanding that the anti-tivoization provisions are
*the* objectionable issue about GPLv3 for those of you who dislike
GPLv3, this is circular reasoning:

  anti-tivoization is bad
  => we reject licenses with it
  => there are fewer developers willing to develop with such licenses
  => anti-tivoization is bad



> Face it, the "open source" crowd is the *bigger* crowd.

I really don't know about that.  I can believe it may be so in LKML.

> I haven't really seen a single one. Last I did the statistic, I asked the 
> top ~25-30 kernel developers about their opinion. NOT A SINGLE ONE 
> preferred the GPLv3.

Wow, that's a really big sample among all Free Software and Open
Source developers out there.  And not even a little bit biased at
that.

> So I have actual *numbers* on my side. What do you have, except for a 
> history of not actually understanding my arguments?

Which is worse, not understanding or repeatedly snipping out and
addressing unrelated points?


Let's please try again.

I'll try to keep it simple, since you can't seem to be able to grasp
the entire argument, and keep disregarding essential parts, disputing
unrelated points and jumping to the conclusions that you've disputed
the point I was trying to make.

I'll present it in parts, as an attempt to stop you from making this
mistake, that I'm sure is not intentional.

The first part is in this e-mail.


Dispute this:

non-tivoized hardware => users can scratch their itches => more
contributions from these users

tivoized hardware => users can't scratch their itches => fewer
contributions from these users


-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Tuesday 19 June 2007 01:51:19 Alexandre Oliva wrote:
> On Jun 19, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:
> > The GPLv2 is the one that allows more developers.
> >
> > The GPLv2 is the one that is acceptable to more people.
>
> Based on my understanding that the anti-tivoization provisions are
> *the* objectionable issue about GPLv3 for those of you who dislike
> GPLv3, this is circular reasoning:
>
>   anti-tivoization is bad
>   => we reject licenses with it
>   => there are fewer developers willing to develop with such licenses
>   => anti-tivoization is bad

The logic is close to:
   => License forbids X
   => developer has requirement for X in license, can't add to project
   => License forbidding X is bad

When it comes to TiVO the reason the developers "required" the "tivoization" 
was because the company itself demanded it. The reason: providers of the 
content their device works with demanded it.

> > Face it, the "open source" crowd is the *bigger* crowd.
>
> I really don't know about that.  I can believe it may be so in LKML.

Actually, it is. Every "Free Software" developer that I personally know could 
care less about the FSF's motives - until they impact them. Since they don't 
care what the FSF does, publishes, etc... they cannot be termed "Free 
Software" developers (using the definition of the term "Free Software" 
provided by the FSF). In fact, almost all of them will state either: "I work 
on Open Source software" or "I develop FOSS stuff".

> > I haven't really seen a single one. Last I did the statistic, I asked the
> > top ~25-30 kernel developers about their opinion. NOT A SINGLE ONE
> > preferred the GPLv3.
>
> Wow, that's a really big sample among all Free Software and Open
> Source developers out there.  And not even a little bit biased at
> that.

Yes, the sample could be considered "biased" - jst as a sample taken among the 
GCC developers could be considered "biased" towards the other end of the 
spectrum.

> > So I have actual *numbers* on my side. What do you have, except for a
> > history of not actually understanding my arguments?
>
> Which is worse, not understanding or repeatedly snipping out and
> addressing unrelated points?

It's time to quote a very ancient source: "Don't point out the speck in your 
neighbors eye when you cannot see the log in your own"

In other words - you've done the same and more.

>
> Let's please try again.
>
> I'll try to keep it simple, since you can't seem to be able to grasp
> the entire argument, and keep disregarding essential parts, disputing
> unrelated points and jumping to the conclusions that you've disputed
> the point I was trying to make.
>
> I'll present it in parts, as an attempt to stop you from making this
> mistake, that I'm sure is not intentional.
>
> The first part is in this e-mail.
>
>
> Dispute this:
>
> non-tivoized hardware => users can scratch their itches => more
> contributions from these users
>
> tivoized hardware => users can't scratch their itches => fewer
> contributions from these users

Linus doesn't have to. Statistically the number of people that will even think 
of modifying the code running on a "tivoized" device is minute - at most 5% 
of the users of such a device. Of those people the ones with the skill to 
actually do the work is an even smaller number - figure 2.5 to 3% of them. Of 
those  with the skill, probably about 10% of them are actually *good* enough 
at it for their changes to be useful. Of that number, figure that only 25%, 
at most, will contribute the changes back.

Apply that to a sample case:
"tivoized" device total users: 1,000,000
people that think about modifying: 50,000 (5%)
people with skill: 1500 (3%)
people who are good enough for the changes to be useful: 150 (10%)
those who will contribute them back: 38 (25%)

Now a normal companies software department is anywhere from 10 to 50 people. 
Large companies can employ more - IBM employs hundreds.

What you are arguing is that people should abandon a firm set of developers 
that is proven to be large for the potential at adding, at most, 38 
developers per million users of the device. If that number was more than 1000 
per device I'd agree. The numbers don't support your argument.

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 19, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> On Tuesday 19 June 2007 01:51:19 Alexandre Oliva wrote:
>> On Jun 19, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:
>> > The GPLv2 is the one that allows more developers.
>> >
>> > The GPLv2 is the one that is acceptable to more people.
>> 
>> Based on my understanding that the anti-tivoization provisions are
>> *the* objectionable issue about GPLv3 for those of you who dislike
>> GPLv3, this is circular reasoning:
>> 
>> anti-tivoization is bad
>> => we reject licenses with it
>> => there are fewer developers willing to develop with such licenses
>> => anti-tivoization is bad

> The logic is close to:

>    => License forbids X
>    => developer has requirement for X in license, can't add to project
>    => License forbidding X is bad

I'm not sure it was clear that '=>' was meant as logical implication.
Read it as "therefore".

It's actually funny that what your inference sequence (in spite of the
missing initial operand) rings so true about my impressions about some
of the reactions I'm getting here.

  GPLv3 forbids tivoization, therefore developer has requirement for
  tivoization in the license, therefore GPLv3 forbidding tivoization
  is bad.

:-)

>> > I haven't really seen a single one. Last I did the statistic, I asked the
>> > top ~25-30 kernel developers about their opinion. NOT A SINGLE ONE
>> > preferred the GPLv3.

>> Wow, that's a really big sample among all Free Software and Open
>> Source developers out there.  And not even a little bit biased at
>> that.

Sorry that I missed the <irony> markers.

> Yes, the sample could be considered "biased" - jst as a sample taken
> among the GCC developers could be considered "biased" towards the
> other end of the spectrum.

FWIW, I haven't taken such a sample, because I know my network of
contacts would likely make it statistically useless.  I'd not try to
make an argument based on that.

>> > So I have actual *numbers* on my side. What do you have, except for a
>> > history of not actually understanding my arguments?
>> 
>> Which is worse, not understanding or repeatedly snipping out and
>> addressing unrelated points?

> In other words - you've done the same and more.

I've honestly tried not to.  I believe Linus has, too.  Many of us
have talked past each other, a lot.

That was actually the point behind breaking up the argument in small
pieces.

If Linus hadn't got the whole argument, a number of times, before,
this might be described as dishonest, but since he did, and he can
refer back to those messages, he can know where I'm going.

>> Dispute this:

>> non-tivoized hardware => users can scratch their itches => more
>> contributions from these users

>> tivoized hardware => users can't scratch their itches => fewer
>> contributions from these users

> Linus doesn't have to.

Of course he doesn't.  But he will.  Because he's always right, and he
wants to show that.  That this is a bait and he knows it won't stop
him.  He knows there's no hook, because he knows where I'm going with
the argument.  But it's going to be interesting to watch.

> Statistically the number of people that will even think of modifying
> the code running on a "tivoized" device is minute

Wait a minute, these figures you made up are for the tivoized hardware
(no changes allowed to the GPLed software in it), or for the
non-tivoized hardware (changes allowed to the GPLed software in it)?

> those who will contribute them back: 38 (25%)

Regardless of what you meant, this is 38 developers *on top* of
however many the company pays to work on that, unless you're jumping
the gun and spoiling the multi-part argument.

> What you are arguing is that people should abandon

I'm not arguing any such thing.  Where's any such argument above?

At this point, I'm only comparing a tivoized device with a
non-tivoized device.  Nothing but it.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Tuesday 19 June 2007 02:44:32 Alexandre Oliva wrote:
> On Jun 19, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> > On Tuesday 19 June 2007 01:51:19 Alexandre Oliva wrote:
> >> On Jun 19, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:
> >> > The GPLv2 is the one that allows more developers.
> >> >
> >> > The GPLv2 is the one that is acceptable to more people.
> >>
> >> Based on my understanding that the anti-tivoization provisions are
> >> *the* objectionable issue about GPLv3 for those of you who dislike
> >> GPLv3, this is circular reasoning:
> >>
> >> anti-tivoization is bad
> >> => we reject licenses with it
> >> => there are fewer developers willing to develop with such licenses
> >> => anti-tivoization is bad
> >
> > The logic is close to:
> >
> >    => License forbids X
> >    => developer has requirement for X in license, can't add to project
> >    => License forbidding X is bad
>
> I'm not sure it was clear that '=>' was meant as logical implication.
> Read it as "therefore".
>
> It's actually funny that what your inference sequence (in spite of the
> missing initial operand) rings so true about my impressions about some
> of the reactions I'm getting here.
>
>   GPLv3 forbids tivoization, therefore developer has requirement for
>   tivoization in the license, therefore GPLv3 forbidding tivoization
>   is bad.
>
> :-)

However, my argument is straight logic, nothing "circular" about it.  :)
Replacing "X" in my logic path above with "tivoization" and "license" 
with "GPLv3", as you've done, does produce a valid chain of logic.  

> >> > I haven't really seen a single one. Last I did the statistic, I asked
> >> > the top ~25-30 kernel developers about their opinion. NOT A SINGLE ONE
> >> > preferred the GPLv3.
> >>
> >> Wow, that's a really big sample among all Free Software and Open
> >> Source developers out there.  And not even a little bit biased at
> >> that.
>
> Sorry that I missed the <irony> markers.
>
> > Yes, the sample could be considered "biased" - jst as a sample taken
> > among the GCC developers could be considered "biased" towards the
> > other end of the spectrum.
>
> FWIW, I haven't taken such a sample, because I know my network of
> contacts would likely make it statistically useless.  I'd not try to
> make an argument based on that.

FWIW the Linux Kernel shouldn't be as homogeneous a population as it is. I'd 
expect it with an FSF run project, because they require copyright assignment 
in order to participate, but with a project like Linux, where everyone 
maintains the copyright to their contributions, should be a hell of a lot 
less homogeneous than Linus' numbers make it seem.

<snip>
> > Statistically the number of people that will even think of modifying
> > the code running on a "tivoized" device is minute
>
> Wait a minute, these figures you made up are for the tivoized hardware
> (no changes allowed to the GPLed software in it), or for the
> non-tivoized hardware (changes allowed to the GPLed software in it)?

Actually, any generic "TiVO"-like hardware - whether it is tivoized or not. 
Admittedly the numbers are significantly different for PC's (and other types 
of general purpose computing devices).

> > those who will contribute them back: 38 (25%)
>
> Regardless of what you meant, this is 38 developers *on top* of
> however many the company pays to work on that, unless you're jumping
> the gun and spoiling the multi-part argument.

38ppm is a fairly small amount, regardless.

> > What you are arguing is that people should abandon
>
> I'm not arguing any such thing.  Where's any such argument above?
>
> At this point, I'm only comparing a tivoized device with a
> non-tivoized device.  Nothing but it.

You've been making the argument the entire time you've been arguing that 
the "anti-tivoization" language in the GPLv3 is necessary. I think I'd rather 
see a guaranteed increase of developers - even if it is only 10 - rather than 
hoping that the potential pool of 38 actually follows through. Wouldn't you?

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 19, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> On Tuesday 19 June 2007 02:44:32 Alexandre Oliva wrote:
>> GPLv3 forbids tivoization, therefore developer has requirement for
>> tivoization in the license, therefore GPLv3 forbidding tivoization
>> is bad.

> However, my argument is straight logic, nothing "circular" about it.  :)
> Replacing "X" in my logic path above with "tivoization" and "license" 
> with "GPLv3", as you've done, does produce a valid chain of logic.  

Yes.  Isn't it funny though that tivoization became necessary as a
consequence of GPLv3 forbidding it?

> FWIW the Linux Kernel shouldn't be as homogeneous a population as it
> is.

Nah.  Communities tend to form around similar values.  Linus started
the community.

>> Wait a minute, these figures you made up are for the tivoized hardware
>> (no changes allowed to the GPLed software in it), or for the
>> non-tivoized hardware (changes allowed to the GPLed software in it)?

> Actually, any generic "TiVO"-like hardware - whether it is tivoized or not. 

So your claim is that a user's possibility to scratch her own itches
makes no difference whatsoever as to their amount of contributions she
is likely to make?

Am I the only one who thinks this is utter nonsense?

>> > those who will contribute them back: 38 (25%)

>> Regardless of what you meant, this is 38 developers *on top* of
>> however many the company pays to work on that, unless you're jumping
>> the gun and spoiling the multi-part argument.

> 38ppm is a fairly small amount, regardless.

Yes.  And your estimates are way too low too, FWIW.  Any reason why
you changed your mind as to the 10% before?

>> > What you are arguing is that people should abandon

>> I'm not arguing any such thing.  Where's any such argument above?

>> At this point, I'm only comparing a tivoized device with a
>> non-tivoized device.  Nothing but it.

> You've been making the argument the entire time you've been arguing that 
> the "anti-tivoization" language in the GPLv3 is necessary.

And then I decided that, since the argument wasn't getting through, I
had to break it into pieces.

The piece I've presented so far has no abandonment whatsoever.  It's a
comparison between two different situations, to evaluate which of them
brings more contributions from users, regardless of the contributions
from the vendor, that are assumed to be the same, since there's no
material difference as far as the vendor is concerned (as in, vendor
has no reason to tivoize)

So your arguments bear zero relationship with the piece I've
proposed.  Can you see that?



> I think I'd rather see a guaranteed increase of developers - even if
> it is only 10 - rather than hoping that the potential pool of 38
> actually follows through. Wouldn't you?

Yes.  How does this relate with the piece of the argument I've
proposed so far, or the whole argument I've posted before?

Answer: It doesn't.  At all.  You're just showing you didn't
understand the argument.  Which shows why I have to explain it piece
by piece.  Which suggests you shouldn't try to jump to conclusions.



Once again, now with clearer starting conditions (not intended to
match TiVo in any way, BTW; don't get into that distraction)


Vendor doesn't care about tivoizing, their business works the same
either way.

Vendor's employees will contribute the same, one way or another, so
their contributions are out of the equation.

Users get source code in either case, and they can modify it and share
it.  They're in no way stopped from becoming part of the community.


Given these conditions:

In a tivoized device, users will be unable to scratch their itches.
This doesn't stop them from contributing to the project, but they may
lack self-interest motivation to contribute, because they won't be
able to use their modifications in the device they own.

In a non-tivoized device, users can scratch their itches.  They can
contribute just as much as they would in a tivoized device, but since
they can use the changes they make to make their own devices work
better for them, this works as a motivator for them to make changes,
and perhaps to contribute them.  Therefore, they will tend to
contribute more.


Can you point out any flaw in this reasoning, or can we admit it as
true?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Tuesday 19 June 2007 04:04:52 Alexandre Oliva wrote:
> On Jun 19, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> > On Tuesday 19 June 2007 02:44:32 Alexandre Oliva wrote:
> >> GPLv3 forbids tivoization, therefore developer has requirement for
> >> tivoization in the license, therefore GPLv3 forbidding tivoization
> >> is bad.
> >
> > However, my argument is straight logic, nothing "circular" about it.  :)
> > Replacing "X" in my logic path above with "tivoization" and "license"
> > with "GPLv3", as you've done, does produce a valid chain of logic.
>
> Yes.  Isn't it funny though that tivoization became necessary as a
> consequence of GPLv3 forbidding it?

-ELOGIC

It didn't become necessary as a result of the GPLv3 forbidding it. As I 
pointed out in text that was cut to keep the post short, there could be any 
number of reasons why "tivoization" is needed by the manufacturer. Other 
people have also pointed that out. This whole bit was to point out that you 
were inferring circular logic where none existed. 

<snip>
> >> Wait a minute, these figures you made up are for the tivoized hardware
> >> (no changes allowed to the GPLed software in it), or for the
> >> non-tivoized hardware (changes allowed to the GPLed software in it)?
> >
> > Actually, any generic "TiVO"-like hardware - whether it is tivoized or
> > not.
>
> So your claim is that a user's possibility to scratch her own itches
> makes no difference whatsoever as to their amount of contributions she
> is likely to make?

Exactly.

> Am I the only one who thinks this is utter nonsense?
>
> >> > those who will contribute them back: 38 (25%)
> >>
> >> Regardless of what you meant, this is 38 developers *on top* of
> >> however many the company pays to work on that, unless you're jumping
> >> the gun and spoiling the multi-part argument.
> >
> > 38ppm is a fairly small amount, regardless.
>
> Yes.  And your estimates are way too low too, FWIW.  Any reason why
> you changed your mind as to the 10% before?

That 10% was, IIRC, a reference to the potential number of "Hackers" that 
would own a TiVO. On thinking about it I realized that the number of hackers 
owning a TiVO would be vanishingly small because of "tivoization". So in this 
new set of numbers I dropped it entirely.

... crap I am tempted to respond to nastily has been cut ...

> > I think I'd rather see a guaranteed increase of developers - even if
> > it is only 10 - rather than hoping that the potential pool of 38
> > actually follows through. Wouldn't you?
>
> Yes.  How does this relate with the piece of the argument I've
> proposed so far, or the whole argument I've posted before?
>
> Answer: It doesn't.  At all.  You're just showing you didn't
> understand the argument.  Which shows why I have to explain it piece
> by piece.  Which suggests you shouldn't try to jump to conclusions.

Wrong. Nobody here needs a "piece by piece" explanation. So, in the belief 
that you were intelligent enough to understand that, I was providing proof 
that refutes your argument entirely.

With a situation as complex as what exists you can't split the argument into 
two and claim that, since "Argument A" is true in the "split" argument that 
it is true when the argument isn't split. This holds true for almost all 
real-world situations.

Now, I am not enjoying the discussion anymore. I've asked once before - remove 
me from the CC list.

DRH


-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 19, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> On Tuesday 19 June 2007 04:04:52 Alexandre Oliva wrote:
>> On Jun 19, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
>> > On Tuesday 19 June 2007 02:44:32 Alexandre Oliva wrote:
>> >> GPLv3 forbids tivoization, therefore developer has requirement for
>> >> tivoization in the license, therefore GPLv3 forbidding tivoization
>> >> is bad.
>> >
>> > However, my argument is straight logic, nothing "circular" about it.  :)
>> > Replacing "X" in my logic path above with "tivoization" and "license"
>> > with "GPLv3", as you've done, does produce a valid chain of logic.
>> 
>> Yes.  Isn't it funny though that tivoization became necessary as a
>> consequence of GPLv3 forbidding it?

> -ELOGIC

I see.  Try 'modprobe logic', it worked for me years ago ;-) :-D

> It didn't become necessary as a result of the GPLv3 forbidding it.

Which is why I said it was funny, because your inference chain stated
*exactly* (with an implied "for the developers") that it did.

Do you understand what an inference chain is?  A => B, as in A implies
B, which can also be read as A therefore B if A is known to hold.

> there could be any number of reasons why "tivoization" is needed by
> the manufacturer.

This claim is false.

Tivoization is when hardware manufacturer takes copyleft software and
blocks updates by the user of the hardware.

No single law so far has shown an example that even resembled to
mandate copyleft software, and no contract could possibly establish a
condition like this.

Therefore, this claim is false.

> This whole bit was to point out that you were inferring circular
> logic where none existed.

There *is* circular logic is in place.

The initial premise of this fallacy is that anti-tivoization is bad
for the project.

This is used to conclude that licenses with such provisions should be
rejected.

This is then used to conclude that there are fewer developers who
would develop under such licenses.

Which is then used to conclude that anti-tivozation is bad for the
project.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Tuesday 19 June 2007 13:06:17 Alexandre Oliva wrote:
> On Jun 19, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> > On Tuesday 19 June 2007 04:04:52 Alexandre Oliva wrote:
> >> On Jun 19, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> >> > On Tuesday 19 June 2007 02:44:32 Alexandre Oliva wrote:
> >> >> GPLv3 forbids tivoization, therefore developer has requirement for
> >> >> tivoization in the license, therefore GPLv3 forbidding tivoization
> >> >> is bad.
> >> >
> >> > However, my argument is straight logic, nothing "circular" about it. 
> >> > :) Replacing "X" in my logic path above with "tivoization" and
> >> > "license" with "GPLv3", as you've done, does produce a valid chain of
> >> > logic.
> >>
> >> Yes.  Isn't it funny though that tivoization became necessary as a
> >> consequence of GPLv3 forbidding it?
> >
> > -ELOGIC
>
> I see.  Try 'modprobe logic', it worked for me years ago ;-) :-D

Try using real logic rather than the logic of your religion.

> > It didn't become necessary as a result of the GPLv3 forbidding it.
>
> Which is why I said it was funny, because your inference chain stated
> *exactly* (with an implied "for the developers") that it did.
>
> Do you understand what an inference chain is?  A => B, as in A implies
> B, which can also be read as A therefore B if A is known to hold.

Okay, since you want it in a specific language rather than as a set of bullet 
points (which is what I used the => for)

Company X has requirement for restriction Y
  => License on product Z disallows restriction Y
  => Product Z loses Company X and the exposure use in their product gives
  => License on product Z is bad for the product

Understandable now?

> > there could be any number of reasons why "tivoization" is needed by
> > the manufacturer.
>
> This claim is false.
>
> Tivoization is when hardware manufacturer takes copyleft software and
> blocks updates by the user of the hardware.

By that definition you are correct.

> No single law so far has shown an example that even resembled to
> mandate copyleft software, and no contract could possibly establish a
> condition like this.

No argument here. What I was stating is that are legal (and other reasons) why 
a company might have to lock down their software in a process similar 
to "Tivoization".

> Therefore, this claim is false.

Only when you define a term as specifically as you have done 
for "Tivoization". I should, perhaps, have used a different term - it would 
then have been patently true. Though, at that point, you would likely have 
argued that it wasn't "tivoization"

> > This whole bit was to point out that you were inferring circular
> > logic where none existed.
>
> There *is* circular logic is in place.
>
> The initial premise of this fallacy is that anti-tivoization is bad
> for the project.
>
> This is used to conclude that licenses with such provisions should be
> rejected.
>
> This is then used to conclude that there are fewer developers who
> would develop under such licenses.
>
> Which is then used to conclude that anti-tivozation is bad for the
> project.

Your view of the logic is, in this case, flawed. It's more along the lines of 
what I've detailed above. Now, please, go away. You aren't doing 
your "religion" any good. In fact, you are damaging it - repeatedly.

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 19, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> Company X has requirement for restriction Y
>   => License on product Z disallows restriction Y
>   => Product Z loses Company X and the exposure use in their product gives
>   => License on product Z is bad for the product

> Understandable now?

Well, considering that I've made this claim myself as part of my
complete argument in a number of times I've presented it, yes, this is
understandable and correct.  This is indeed one of the cases.

That said, very few companies have a scrict *requirement* for keeping
the ability to modify the software on the customer's computer while
denying this ability to the customer.

So this case you're discussing is the least common case.  It could
nearly be dismissed, rather than being the dominating topic in the
discussion, as it's been so far.

> What I was stating is that are legal (and other reasons) why a
> company might have to lock down their software in a process similar
> to "Tivoization".

Ok.  Most of these can be addressed (with inconvenience) with ROM.
Others are business reasons, and for these, the increased cost and
inconvenience of the alternatives may shift them to an unlock
situation.

>> Therefore, this claim is false.

> Only when you define a term as specifically as you have done 
> for "Tivoization".

It's not my definition.  This was from Wikipedia.

> I should, perhaps, have used a different term - it would 
> then have been patently true.

Depends on what the different term was.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 19, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> On Tuesday 19 June 2007 04:04:52 Alexandre Oliva wrote:
>> So your claim is that a user's possibility to scratch her own itches
>> makes no difference whatsoever as to their amount of contributions she
>> is likely to make?

> Exactly.

Hmm, interesting...  Are you doing this just to make it more
interesting, or did you never take Open Source 101? ;-)  Or economy,
game theory, ecology, politics, for that matter?

Do you realize how many contradictions arise from this claim?

A few examples for you.  Let me know if you don't understand why these
contradict this claim, I'll be happy to explain it to you.

How many newbies have patches accepted that they didn't test?

At which point in becoming a Linux developer does one leave behind
this altruistic attitude and becomes moved by self interest only?

How many people you know got to know Linux in their TiVos, took the
Linux sources that TiVo distributes and built it for their own PC?

>> Yes.  And your estimates are way too low too, FWIW.  Any reason why
>> you changed your mind as to the 10% before?

> That 10% was, IIRC, a reference to the potential number of "Hackers" that 
> would own a TiVO. On thinking about it I realized that the number of hackers 
> owning a TiVO would be vanishingly small because of "tivoization".

Aah, ok.

So, you lowered the estimate for the case of a tivoized device, but
then claimed it was the same for an otherwise-identical non-tivoized
device.

Seriously, try 'modprobe logic', you'll like it ;-)

> Wrong. Nobody here needs a "piece by piece" explanation.

Then why do you keep making claims that are not related with either
the part of the argument I'm posing or the full argument I've already
presented?

> So, in the belief that you were intelligent enough to understand
> that, I was providing proof that refutes your argument entirely.

The only proof you provided was that you didn't understand the
argument.

> With a situation as complex as what exists you can't split the
> argument into two and claim that, since "Argument A" is true in the
> "split" argument that it is true when the argument isn't split. This
> holds true for almost all real-world situations.

Which proves you don't understand how logical proofs work.

Let me explain it to you.

First, you need to establish initial premises.  Whether or not they
resemble any similarity with anything else you're thinking of is
irrelevant.  They might even be known to be false, in proof by
contradiction.

Then, you apply logical inference rules to the initial premises, and
establish consequences of the initial premises.

One of these consequences may be what you are trying to prove.  Or,
you may come to a contradiction, and prove that initial premises are
self-contradictory.  Or you may come to no useful conclusion
whatsoever.

> Now, I am not enjoying the discussion anymore.

Understandable.

> I've asked once before - remove me from the CC list.

I'll try to remember to do that.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[david]

On Tue, 19 Jun 2007, Alexandre Oliva wrote:

>
> Once again, now with clearer starting conditions (not intended to
> match TiVo in any way, BTW; don't get into that distraction)
>
>
> Vendor doesn't care about tivoizing, their business works the same
> either way.
>
> Vendor's employees will contribute the same, one way or another, so
> their contributions are out of the equation.

no, this is what you are missing.

with tivoizing, the vendors employees are working on linux and 
contributing

without tivoizing and with GPLv3 license involved the vendors employees 
are working on a propriatary OS and are contributing nothing back

> Users get source code in either case, and they can modify it and share
> it.  They're in no way stopped from becoming part of the community.

not if the GPLv3 achieves it's objectives of forcing the vendor to stop 
useing opensource software.

>
> Given these conditions:
>
> In a tivoized device, users will be unable to scratch their itches.
> This doesn't stop them from contributing to the project, but they may
> lack self-interest motivation to contribute, because they won't be
> able to use their modifications in the device they own.
>
> In a non-tivoized device, users can scratch their itches.  They can
> contribute just as much as they would in a tivoized device, but since
> they can use the changes they make to make their own devices work
> better for them, this works as a motivator for them to make changes,
> and perhaps to contribute them.  Therefore, they will tend to
> contribute more.
>
>
> Can you point out any flaw in this reasoning, or can we admit it as
> true?

in a tivoized device, users have access to the source code and can 
understand how things work, incoporate improvements into other projects, 
etc.

In addition, users who bypass the lockdown restrictions can modify the 
software on that device.

in a non-tivoized device users have a black-box and have to reverse 
engineer everything becouse the vendor releases no source at all.

David Lang

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 19, 2007, david@lang.hm wrote:

> On Tue, 19 Jun 2007, Alexandre Oliva wrote:
>> 
>> Once again, now with clearer starting conditions (not intended to
>> match TiVo in any way, BTW; don't get into that distraction)
>> 
>> 
>> Vendor doesn't care about tivoizing, their business works the same
>> either way.
>> 
>> Vendor's employees will contribute the same, one way or another, so
>> their contributions are out of the equation.

> no, this is what you are missing.

No, this is just one of many possible situations.

You're just thinking of a different scenario, in which tivoization is
important for the tivoizer.  We'll get to that.

Please be patient.  Don't get the illusion that you're disputing or
weakening the argument with these statements.  Disputing premises of a
logical inference is pointless as far as disputing a larger argument
that this particular lemma would be a part of.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[david]

On Tue, 19 Jun 2007, Alexandre Oliva wrote:

> On Jun 19, 2007, david@lang.hm wrote:
>
>> On Tue, 19 Jun 2007, Alexandre Oliva wrote:
>>>
>>> Once again, now with clearer starting conditions (not intended to
>>> match TiVo in any way, BTW; don't get into that distraction)
>>>
>>>
>>> Vendor doesn't care about tivoizing, their business works the same
>>> either way.
>>>
>>> Vendor's employees will contribute the same, one way or another, so
>>> their contributions are out of the equation.
>
>> no, this is what you are missing.
>
> No, this is just one of many possible situations.
>
> You're just thinking of a different scenario, in which tivoization is
> important for the tivoizer.  We'll get to that.
>
> Please be patient.  Don't get the illusion that you're disputing or
> weakening the argument with these statements.  Disputing premises of a
> logical inference is pointless as far as disputing a larger argument
> that this particular lemma would be a part of.

if a company doesn't care about tivoizing then they won't do it, it takes 
time and money to tivoize some product and it will cause headaches for the 
company.

their reasons for wanting to tivoize a product may be faulty, but they 
think that the reasons are valid or they wouldn't go to the effort.

David Lang

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 19, 2007, david@lang.hm wrote:

> if a company doesn't care about tivoizing then they won't do it, it
> takes time and money to tivoize some product and it will cause
> headaches for the company.

> their reasons for wanting to tivoize a product may be faulty, but they
> think that the reasons are valid or they wouldn't go to the effort.

Absolutely right.  And we'll get to that.  Please just be patient.

In fact, how much the company cares about tivoizing is completely
irrelevant to that point.  I shouldn't even have included it, but I
did because I thought it would be useful as a boundary condition.

So just disregard that.

Is there agreement that, comparing tivoized and non-tivoized hardware,
we get'd more contributions if the hardware is not tivoized, because
users can scratch their own itches, than we would for tivoized
hardware?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[david]

On Tue, 19 Jun 2007, Alexandre Oliva wrote:

> Subject: Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3
> 
> On Jun 19, 2007, david@lang.hm wrote:
>
>> if a company doesn't care about tivoizing then they won't do it, it
>> takes time and money to tivoize some product and it will cause
>> headaches for the company.
>
>> their reasons for wanting to tivoize a product may be faulty, but they
>> think that the reasons are valid or they wouldn't go to the effort.
>
> Absolutely right.  And we'll get to that.  Please just be patient.
>
> In fact, how much the company cares about tivoizing is completely
> irrelevant to that point.  I shouldn't even have included it, but I
> did because I thought it would be useful as a boundary condition.
>
> So just disregard that.
>
> Is there agreement that, comparing tivoized and non-tivoized hardware,
> we get'd more contributions if the hardware is not tivoized, because
> users can scratch their own itches, than we would for tivoized
> hardware?

if you also make the assumption that the company won't use propriatary 
software instead then I think you would get agreement. but the 
disagrement is over this exact assumption. you assume that these companies 
will use non-tivoized products if you make it hard to use the software 
covered by the GPL, most other people are saying that they disagree and 
the result would be fewer companies useing software covered by GPL 
instead.

David Lang

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 19, 2007, david@lang.hm wrote:

> if you also make the assumption that the company won't use propriatary
> software instead then I think you would get agreement.

Ah, good point.  When I posed the one of the two cases of the inicial
scenario as "no tivoization", I meant Free Software without
constraints.

> but the disagrement is over this exact assumption. you assume that
> these companies will use non-tivoized products if you make it hard
> to use the software covered by the GPL, most other people are saying
> that they disagree and the result would be fewer companies useing
> software covered by GPL instead.

I understand that.  And what I'm saying is that, even if fewer such
companies use GPLed software, you may still be better off, out of
additional contributions you'll get from customers of companies that
switch from tivoization to unconstrained Free Software, because of the
additional costs of the alternatives.

And no, I can't prove it, but it's good that at least the argument is
no longer completely disregarded while something else is disputed.


Now that you guys at least understand what the argument is, you can
figure out the solution by yourselves.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[david]

On Tue, 19 Jun 2007, Alexandre Oliva wrote:

> On Jun 19, 2007, david@lang.hm wrote:
>
>> if you also make the assumption that the company won't use propriatary
>> software instead then I think you would get agreement.
>
> Ah, good point.  When I posed the one of the two cases of the inicial
> scenario as "no tivoization", I meant Free Software without
> constraints.
>
>> but the disagrement is over this exact assumption. you assume that
>> these companies will use non-tivoized products if you make it hard
>> to use the software covered by the GPL, most other people are saying
>> that they disagree and the result would be fewer companies useing
>> software covered by GPL instead.
>
> I understand that.  And what I'm saying is that, even if fewer such
> companies use GPLed software, you may still be better off, out of
> additional contributions you'll get from customers of companies that
> switch from tivoization to unconstrained Free Software, because of the
> additional costs of the alternatives.
>
> And no, I can't prove it, but it's good that at least the argument is
> no longer completely disregarded while something else is disputed.
>
>
> Now that you guys at least understand what the argument is, you can
> figure out the solution by yourselves.

good, if you are no longer going to claim that your opinion on this 
unprovable point is the Truth (not the capitol) hopefully you can accept 
that a lot of very smart people are convinced that you are wrong on this 
point and not just 'confused'

David Lang

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Lennart Sorensen]

On Tue, Jun 19, 2007 at 05:04:52AM -0300, Alexandre Oliva wrote:
> Yes.  How does this relate with the piece of the argument I've
> proposed so far, or the whole argument I've posted before?
> 
> Answer: It doesn't.  At all.  You're just showing you didn't
> understand the argument.  Which shows why I have to explain it piece
> by piece.  Which suggests you shouldn't try to jump to conclusions.
> 
> 
> 
> Once again, now with clearer starting conditions (not intended to
> match TiVo in any way, BTW; don't get into that distraction)
> 
> 
> Vendor doesn't care about tivoizing, their business works the same
> either way.

Not true.  A PVR that can record pay per view and encrypted digital
channels and other such things (as mandated by MPAA or whoever is
unfortunately in charge of such stupid thigns), is much easier to sell
to customers, and as a result they have to prevent things that the
content providers want prevented (never mind that such prevention is
futile because there is no DRM that can't be broken, only made more
difficult to break).  So their business does not work anywhere near as
well if they can't sell devices that do what their customers want them
to do (which is record stuff from TV, not play tetris, even though
tetris might be fun too).  Now if they don't do what the content
providers want, they have to sell an inferior PVR, while a competitor
that does what the content providers demand (in order to provide what
customers demand), then they won't sell very much.  They could go with
another non GPL os and software of course, and spend way more moeny on
it (or potentially use BSD or something) and deal with the problem that
way.  Either way GPL software gets no contributions under the GPL v3 if
the company wants to be competitive in that business.

> Vendor's employees will contribute the same, one way or another, so
> their contributions are out of the equation.

No, because they are no longer using GPL software.

> Users get source code in either case, and they can modify it and share
> it.  They're in no way stopped from becoming part of the community.

The users don't get any code now, because the vendor has no requirement
to release it, or may even have no permission to realse it.  And end
users are for the most part incapable of contributing anyhow.

> Given these conditions:
> 
> In a tivoized device, users will be unable to scratch their itches.
> This doesn't stop them from contributing to the project, but they may
> lack self-interest motivation to contribute, because they won't be
> able to use their modifications in the device they own.

I bought some hardware and built a box to run mythtv.  I can only record
stuff from regular analog cable, because I can't buy any hardware that
supports digital encrypted cable.  If I bought a locked down PVR
instead, then I would be able to record that, but I wouldn't be able to
play around.  To me playing around was more important than recording
encrypted channels, but to most customers, being able to record
everything is what is important.

> In a non-tivoized device, users can scratch their itches.  They can
> contribute just as much as they would in a tivoized device, but since
> they can use the changes they make to make their own devices work
> better for them, this works as a motivator for them to make changes,
> and perhaps to contribute them.  Therefore, they will tend to
> contribute more.

Their device can't physically do some of the thigns they wanted it to do
though, since those features disappeared when the tivoization was
removed due to licensing issues with the content makers.

> Can you point out any flaw in this reasoning, or can we admit it as
> true?

Certainly fails to be true.

Who is to blame?  The content makers for requiring lock down of devices
that can access their content?  Consumers for demanding to be able to
use that content and hence buying devices that can use it?  The company
that made a device that did what consumers demanded by implementing it
the way required by the content makers?

--
Len Sorensen

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 20, 2007, lsorense@csclub.uwaterloo.ca (Lennart Sorensen) wrote:

> On Tue, Jun 19, 2007 at 05:04:52AM -0300, Alexandre Oliva wrote:
>> Once again, now with clearer starting conditions (not intended to
>> match TiVo in any way, BTW; don't get into that distraction)
>> 
>> 
>> Vendor doesn't care about tivoizing, their business works the same
>> either way.

> Not true.  A PVR that can record pay per view and encrypted digital
> channels

You see the "not intended to match TiVo" above?

Do you see that it's pointless to dispute antecedents of a logical
inference rule, if you don't know what role it plays in the full
argument?

Consider that this could be a proof by contradiction to realize how
pointless your objection is, no matter how true the point you state
is.  It bears no relationship with the argument at hand, and you said
so yourself, by disputing the assumptions of the inference, rather
than its conclusions.

Assumptions that were not even used to arrive at the conclusions, BTW.

>> Can you point out any flaw in this reasoning, or can we admit it as
>> true?

> Certainly fails to be true.

Once you change the conditions to twist whatever else you want, then
you arrive at different conclusions.  What's the surprise here?

What you're doing is like, after getting an argument like this:

  assumptions:
    1+1 = 2
    2+1 = 3

  provable consequence:

    1+1+1 = 3

responding:

  no, no, that's wrong!  the right argument is:

  assumptions:

    3+4 = 7
    2+1 = 3

  provable consequence:

    2+1+4 = 7

  therefore the answer is 5, not 3!

You see how illogical this is?

It doesn't matter whether your argument is correct.  It just doesn't
dispute the proposition at hand.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[david]

On Tue, 19 Jun 2007, Daniel Hazelton wrote:

>> Dispute this:
>>
>> non-tivoized hardware => users can scratch their itches => more
>> contributions from these users
>>
>> tivoized hardware => users can't scratch their itches => fewer
>> contributions from these users
>
> Linus doesn't have to. Statistically the number of people that will even think
> of modifying the code running on a "tivoized" device is minute - at most 5%
> of the users of such a device. Of those people the ones with the skill to
> actually do the work is an even smaller number - figure 2.5 to 3% of them. Of
> those  with the skill, probably about 10% of them are actually *good* enough
> at it for their changes to be useful. Of that number, figure that only 25%,
> at most, will contribute the changes back.
>
> Apply that to a sample case:
> "tivoized" device total users: 1,000,000
> people that think about modifying: 50,000 (5%)
> people with skill: 1500 (3%)
> people who are good enough for the changes to be useful: 150 (10%)
> those who will contribute them back: 38 (25%)

based on my experiance looking at the software released for tivos, I think 
you are over-estimating these numbers. if there are more then a dozen 
people producing things that are good enough to be useful and releasing 
their results as opensource software I would be surprised.

and for all that the FSF is claiming that tivos can't being modified it's 
really not that hard to change.

David Lang

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 19, 2007, david@lang.hm wrote:

> based on my experiance looking at the software released for tivos, I
> think you are over-estimating these numbers. if there are more then a
> dozen people producing things that are good enough to be useful and
> releasing their results as opensource software I would be surprised.

Yup.

And how many more would there be should it not be tivoized?  More
hackers would buy the devices, a number of them with the explicit
intent and interest in modifying the software in it.

You're losing all that.

> and for all that the FSF is claiming that tivos can't being modified
> it's really not that hard to change.

But is it legal?

How many would contribute changes to a list where there are TiVo
people watching, which might expose these contributors to liabilities?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[david]

On Tue, 19 Jun 2007, Alexandre Oliva wrote:

> On Jun 19, 2007, david@lang.hm wrote:
>
>> based on my experiance looking at the software released for tivos, I
>> think you are over-estimating these numbers. if there are more then a
>> dozen people producing things that are good enough to be useful and
>> releasing their results as opensource software I would be surprised.
>
> Yup.
>
> And how many more would there be should it not be tivoized?  More
> hackers would buy the devices, a number of them with the explicit
> intent and interest in modifying the software in it.
>
> You're losing all that.

based on the knowledge shown by these users you aren't loosing much.

remember, not all tivo models are locked down, and most of those that are 
locked down can be unlocked pretty easily. it's only a couple models that 
have required a soldering iron.

the xbox hacks show that people will produce ways to bypass hardware 
restrictions that are easy for less techinical people to use (and people 
less technical then this aren't gong to be contributing code anyway)

as a result of watching the hacker groups I can safely say that the 
lockdown has not blocked many users. it has slowed modification of the 
hacks to new types of hardware, but not for very long.

>> and for all that the FSF is claiming that tivos can't being modified
>> it's really not that hard to change.
>
> But is it legal?

that depends on what lawyer you ask, and what they think of the 
applicability of the DMCA.

> How many would contribute changes to a list where there are TiVo
> people watching, which might expose these contributors to liabilities?

in the case of Tivo, there are many public boards that talk about hacking 
tivos, and tivo employees login and contribute to them. just about all of 
these boards ban specific topics. for example:

getting service without paying for it is banned on every board I 
personally look at

transferring video off of the tivo is banned on some boards, but not on 
others.

David Lang

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 19, 2007, david@lang.hm wrote:

> On Tue, 19 Jun 2007, Alexandre Oliva wrote:
>> You're losing all that.

> based on the knowledge shown by these users you aren't loosing much.

Remember, the sample is biased, the hackers who'd like to hack it are
less likely to buy it, and some might be only making private changes
to avoid legal hassles.

And then, it doesn't have to be all that much.  Since you're talking
specifically about TiVo customers, how much has TiVo effectly
contributed in terms of code?  (It was pointed out before, I know, but
it's important to keep this in perspective, lest people lose sight of
what's at stake)

> remember, not all tivo models are locked down,

Only the earliest that you can't find for sale any more, right?

> as a result of watching the hacker groups I can safely say that the
> lockdown has not blocked many users. it has slowed modification of the
> hacks to new types of hardware, but not for very long.

Well, then...  What's the point *for* tivoization, again?  To slow
down the contributions?  And that's good because...?

> in the case of Tivo, there are many public boards that talk about
> hacking tivos, and tivo employees login and contribute to them.

Nice!


Thanks for the info, this was very enlightening.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[david]

On Tue, 19 Jun 2007, Alexandre Oliva wrote:

> On Jun 19, 2007, david@lang.hm wrote:
>
>> On Tue, 19 Jun 2007, Alexandre Oliva wrote:
>>> You're losing all that.
>
>> based on the knowledge shown by these users you aren't loosing much.
>
> Remember, the sample is biased, the hackers who'd like to hack it are
> less likely to buy it, and some might be only making private changes
> to avoid legal hassles.
>
> And then, it doesn't have to be all that much.  Since you're talking
> specifically about TiVo customers, how much has TiVo effectly
> contributed in terms of code?  (It was pointed out before, I know, but
> it's important to keep this in perspective, lest people lose sight of
> what's at stake)

they contributed all changes they made to GPL code, exactly as required.

in addition they have provided the full build environments, not just 
listing the versions of the compilers, etc.

>> remember, not all tivo models are locked down,
>
> Only the earliest that you can't find for sale any more, right?
>
>> as a result of watching the hacker groups I can safely say that the
>> lockdown has not blocked many users. it has slowed modification of the
>> hacks to new types of hardware, but not for very long.
>
> Well, then...  What's the point *for* tivoization, again?  To slow
> down the contributions?  And that's good because...?

no, the point is that while tivoization is not nessasarily the best thing 
it's far better then the company useing propriatary code.

delayed contributions are better then no contributions.

David Lang

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Tuesday 19 June 2007 19:49:24 david@lang.hm wrote:
> On Tue, 19 Jun 2007, Alexandre Oliva wrote:
> > On Jun 19, 2007, david@lang.hm wrote:
> >> remember, not all tivo models are locked down,
> >
> > Only the earliest that you can't find for sale any more, right?
> >
> >> as a result of watching the hacker groups I can safely say that the
> >> lockdown has not blocked many users. it has slowed modification of the
> >> hacks to new types of hardware, but not for very long.
> >
> > Well, then...  What's the point *for* tivoization, again?  To slow
> > down the contributions?  And that's good because...?
>
> no, the point is that while tivoization is not nessasarily the best thing
> it's far better then the company useing propriatary code.
>
> delayed contributions are better then no contributions.
>
> David Lang

This logic has been proven already. Apple used KHTML and KJS as the backend 
systems for Safari. While Safari was in development they held onto all their 
changes and modifications. When they released Safari, they contributed it all 
back, making both things better. Fact: KHTML and KJS are also the core of 
Apples "WebKit" and "WebCore" technologies - both KHTML and KJS are Open 
Source projects, making up a part of KDE.

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Pekka Enberg]

Hi Alexandre,

On 6/19/07, Alexandre Oliva <aoliva@redhat.com> wrote:
> Dispute this:
>
> non-tivoized hardware => users can scratch their itches => more
> contributions from these users
>
> tivoized hardware => users can't scratch their itches => fewer
> contributions from these users

Maybe, but in what numbers? It's not like every user out there is able
to scratch their itches in the first place. Besides, people who want
to hack their hardware either work around the restrictions (like Linux
on xbox) or buy hardware that doesn't have any.

In any case, I hope we call quits on this thread. People have already
explained to you that there are certain devices (ATMs, medical
equipment) where you absolutely want "tivoization". Furthermore, Linus
has repeatedly explained to you why he (and bunch of other _kernel_
hackers -- this is a thread on LKML, remember) thinks it's stupid for
a software license to restrict hardware design choices. If you don't
see the point, fine, but lets stop wasting each others time, ok?

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 19, 2007, "Pekka Enberg" <penberg@cs.helsinki.fi> wrote:

> Hi Alexandre,
> On 6/19/07, Alexandre Oliva <aoliva@redhat.com> wrote:
>> Dispute this:
>> 
>> non-tivoized hardware => users can scratch their itches => more
>> contributions from these users
>> 
>> tivoized hardware => users can't scratch their itches => fewer
>> contributions from these users

> Maybe, but in what numbers?

We'll get to that.

> Furthermore, Linus has repeatedly explained to you why he (and bunch
> of other _kernel_ hackers -- this is a thread on LKML, remember)
> thinks it's stupid for a software license to restrict hardware
> design choices.

Most of these explanations carried the assumption that this would
lower the amount of contributions he'd get.  And this is precisely the
circular logic in their arguments, and this is why this argument is
intended to show the initial assumption is false.  But hey, maybe I'm
wrong, and it won't show that at all, and I'll be proved wrong, and
then there's going to be one more very vocal defensor of your ideals.


Now, if people want to keep on fooling themselves, I guess they can,
as long as they don't hurt others in the process.  If they try to fool
others, I feel it is my moral duty to intervene.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Pekka Enberg]

Hi Alexandre,

At some point in time, I wrote:
> > Furthermore, Linus has repeatedly explained to you why he (and bunch
> > of other _kernel_ hackers -- this is a thread on LKML, remember)
> > thinks it's stupid for a software license to restrict hardware
> > design choices.

On 6/19/07, Alexandre Oliva <aoliva@redhat.com> wrote:
> Most of these explanations carried the assumption that this would
> lower the amount of contributions he'd get.

Then may I suggest you go back and re-read those threads because
that's not at all what I am reading.

On 6/19/07, Alexandre Oliva <aoliva@redhat.com> wrote:
> Now, if people want to keep on fooling themselves, I guess they can,
> as long as they don't hurt others in the process.  If they try to fool
> others, I feel it is my moral duty to intervene.

Sure. I just wanted to point out that your "moral duty" really comes
off as "trolling" at this end. So could we please move this discussion
to maybe "moral-duties@lists.fsf.org" or some other more appropriate
list and stop flooding LKML?

                                      Pekka

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Mon, 18 Jun 2007, Alexandre Oliva wrote:
> 
> technically, it asks you to pass on (!= give back) access to the
> software (not to the hardware that contains it).

That "technically" is just another way of saying "if you look cross-eyed 
at it, and don't look too closely".

> No.  The reason, again, is the portion you snipped out.
> 
> Could you try again?

No.

Not worth my time. You have shown yourself unable to learn and understand. 

Just live with it. The GPLv2 is better for the kernel.

			Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Kevin Bowling]

> On 6/18/07, Alexandre Oliva <aoliva@redhat.com> wrote:
> > On Jun 18, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:

> > I just want software back. I think it is *wrong* for me to ask for
> > anything else. It's literally my personal "moral choice": I think the
> > hardware manufacturers need to make their _own_ choices when it comes to
> > _their_ designs.
>
> >  - I think that *technical*quality* is more important than *quantity*.
>
> This argument fails to make the point you're trying to make.

No, it has been countered many times and you are simply proving your ignorance.

>   you trade the potential contributions of all those users for the
>   contributions of tivoizers, apparently assuming that all tivoizers
>   would simply move away from the community, taking their future
>   contributions away from your community, rather than moving to a
>   position in which you'd get not only the contributions from the
>   company itself, but also from all their users
>
> and you say "oh, I don't care about quantity, I care about quality",
> as if this somehow related with the above.

Being strictly pragmatic - what makes you think TiVoland is some
fledgling grounds of /expert kernel developers/ that are otherwise
deprived of contributing, unless they can illegally modify their TiVo?

Under GPLv2, we have the kernel modifications and can include them in
our software.  If you don't agree with TiVo, purchase an open product.
 Download their kernel source and use it on your open product.  Pure
consumerism and capitalism at work.

The GPLv3 is a solution in search of a problem.  Worse, it creates
problems outside the grasp of your understanding.

> Just do the math.  Hypothetically, Linux goes GPLv3, without
> permission to tivoize.  TiVo has to decide among:
>
> - switching to another kernel, no further contributions from them
Bad for us, bad for users.
> - sticking with old version, no further usable contributions from them
Bad for us, bad for users.
> - switching to ROM, still the same contributions from TiVo
Bad for users.
> - no more tivoization, contributions from TiVo and users

Bad for us, bad for users.  Legitimate laws and practices require that
certain devices not be modified by end users.  Therefore TiVo fails
and contributions cease.

> So, you see, in no case do you get more contributors while at the same
> time losing TiVo's quality contributions.

No.  Outside of this FSF {academic,religious} diatribe, in the real
world, things aren't as you see them.  The fact is that these people
can get the code and contribute.  It just won't run on TiVo.  So don't
buy TiVo but use the code.  Problem solved, free software in action.

> > In the GPLv3 world, we have already discussed in this thread how you can
> > follow the GPLv3 by making the TECHNICALLY INFERIOR choice of using a ROM
> > instead of using a flash device.
>
> Yes.  This is one option that doesn't bring any benefits to anyone.
> It maintains the status quo for users and the community, but it loses
> the ability for the vendor to upgrade, fix or otherwise control the
> users.  Bad for the vendor.

And users.  Don't spin the facts.

You are advocating things which hurt the end user, which the "Spirit"
should seek to help.  The GPLv3 is here to stroke the FSF ego because
they don't like how somebody has legally found a way to use free
software in a way they don't agree with.  Sort of like dynamite being
used violently...

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 18, 2007, "Kevin Bowling" <lkml@kev009.com> wrote:

> Legitimate laws and practices require that certain devices not be
> modified by end users.  Therefore TiVo fails and contributions
> cease.

I've never denied this possibility.

But how about all the other devices that are being tivoized that do
NOT require this?

Are people just blind to this possibility?  Or does it really not
exist, and I'm the first who ever though of it?

>> Yes.  This is one option that doesn't bring any benefits to anyone.
>> It maintains the status quo for users and the community, but it loses
>> the ability for the vendor to upgrade, fix or otherwise control the
>> users.  Bad for the vendor.

> And users.  Don't spin the facts.

How is it good that the vendor can downgrade the software behind the
user's back?

Oh, yeah, right, it could upgrade it too.

> You are advocating things which hurt the end user,

Actually, no.  I'm advocating for respect for users' freedoms.
Whoever choose not to do that gets slightly hurt in the process, as an
incentive for respecting users' freedoms.  And yes, when the users'
freedoms are not respected (the cases you mentioned), that's bad for
the user, no doubt.  And bad for the community as well.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Jan Harkes]

On Mon, Jun 18, 2007 at 08:31:30PM -0300, Alexandre Oliva wrote:
> On Jun 18, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:
> > In the GPLv3 world, we have already discussed in this thread how you can 
> > follow the GPLv3 by making the TECHNICALLY INFERIOR choice of using a ROM 
> > instead of using a flash device.
> 
> Yes.  This is one option that doesn't bring any benefits to anyone.
> It maintains the status quo for users and the community, but it loses
> the ability for the vendor to upgrade, fix or otherwise control the
> users.  Bad for the vendor.

Not really, Tivo could simply sell you a box without any installed
software. The actual software is mailed to you on a credit card sized
ROM when you activate service. When they want to (or need to) update the
software they send out a new ROM card, maybe yearly as part of the
service subscription renewal.

The box could even be sold by third party vendors, I think they may even
have started off that way, my old Series 1 had a big Philips logo on it.
So now we make sure that this hardware refuses to boot any unsigned
code, but it wasn't shipped containing GPLv3 software, so it's license
terms simply does not apply.

The software is shipped on a ROM card which can no longer be modified by
the manufacturer or any third party, so it would seem to comply with the
GPLv3. I can even imagine that the hardware is really general purpose
but the ROM is encrypted so that only the BIOS/bootloader can unlock it.

So the GPLv3 seems to fall short on actually preventing tivoization. It
just requires an extra layer of indirection, ship hardware seperately
from software.

Jan

RE: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Schwartz]

> The box could even be sold by third party vendors, I think they may even
> have started off that way, my old Series 1 had a big Philips logo on it.
> So now we make sure that this hardware refuses to boot any unsigned
> code, but it wasn't shipped containing GPLv3 software, so it's license
> terms simply does not apply.
>
> The software is shipped on a ROM card which can no longer be modified by
> the manufacturer or any third party, so it would seem to comply with the
> GPLv3. I can even imagine that the hardware is really general purpose
> but the ROM is encrypted so that only the BIOS/bootloader can unlock it.
>
> So the GPLv3 seems to fall short on actually preventing tivoization. It
> just requires an extra layer of indirection, ship hardware seperately
> from software.
>
> Jan

If this flaw is still the latest GPLv3 drafts, it should be fixed. It's a
simple, technical error that can easily be rectified.

I don't have the latest draft handy, but my recollection is that it talks
about software that is "transferred along with" hardware. What it should say
is object code that contains activation logic for hardware.

That is, if I ship the Linux kernel (assuming for the moment the Linux
kernel was under GPLv3) with special activation logic to run on platforms X,
Y, and Z, then the source code should have to include how to make my changes
have that same special activation logic for those same platforms.

If the GPLv3 doesn't do that, it's broken.

Note that I'm not opining on whether this is a good thing or a bad thing.
But people who choose the GPLv3 because they want to prohibit tivoization of
their software should in fact get that prohibition.

DS

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 18, 2007, Jan Harkes <jaharkes@cs.cmu.edu> wrote:

> Not really, Tivo could simply sell you a box without any installed
> software.

Yes.

> The actual software is mailed to you on a credit card sized
> ROM when you activate service.

If that's a separate transaction, then yes, I believe it would not be
convered under the terms of GPLv3, but IANAL.  There might be some
catch about intent, and also about contractual obligations in the
hardware sale to provide the software (coupons anyone? :-) or some
such.


But then, that you sell specialized devices with say only MIT-licensed
software shouldn't stop you from selling CD-ROMs with GPLed software,
even if those CD-ROMs could possibly run on that device.


The GPLv3 won't remove every way in which people who want/need to stop
the user from making changes to the software could accomplishing this
(ROM).  It will just make this a bit more inconvenient, such that
vendors that have the option respect users' freedoms, and those that
find it too inconvenient respect the wishes of users who don't want
their software turned non-free.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Jan Harkes]

On Tue, Jun 19, 2007 at 02:40:59AM -0300, Alexandre Oliva wrote:
> > The actual software is mailed to you on a credit card sized
> > ROM when you activate service.
...
> The GPLv3 won't remove every way in which people who want/need to stop
> the user from making changes to the software could accomplishing this
> (ROM).  It will just make this a bit more inconvenient, such that
> vendors that have the option respect users' freedoms, and those that
> find it too inconvenient respect the wishes of users who don't want
> their software turned non-free.

I am trying to read that last sentence and it just doesn't seem to make
any sense.

Or are you saying that all that anti-tivoization language that adds
complex requirements which change depending on the market some device
happens to be sold in and which will most likely make GPLv3 software
unusable for various applications ranging from medical equipment to
financial transaction systems (and probably others) is there to just
make it a _bit_ more inconvenient for vendors to implement a tivo-like
scheme?

So what exactly is the point of all this then?

Jan

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 19, 2007, Jan Harkes <jaharkes@cs.cmu.edu> wrote:

> On Tue, Jun 19, 2007 at 02:40:59AM -0300, Alexandre Oliva wrote:
>> > The actual software is mailed to you on a credit card sized
>> > ROM when you activate service.
> ...
>> The GPLv3 won't remove every way in which people who want/need to stop
>> the user from making changes to the software could accomplishing this
>> (ROM).  It will just make this a bit more inconvenient, such that
>> vendors that have the option respect users' freedoms, and those that
>> find it too inconvenient respect the wishes of users who don't want
>> their software turned non-free.

> Or are you saying that all that anti-tivoization language that adds
> complex requirements which change depending on the market some device
> happens to be sold in

You allude to the definition of User Product, not geographies, I
suppose.

> and which will most likely make GPLv3 software unusable for various
> applications ranging from medical equipment to financial transaction
> systems (and probably others)

Not unusable, except perhaps for the one example about credit card
terminals presented so far.

> is there to just make it a _bit_ more inconvenient for vendors to
> implement a tivo-like scheme?

I'm not sure they find it to be "just a bit".

Point is to keep Free Software Free freedoms, and ROM doesn't make it
non-Free, so this provision is a means to ensure the compliance with
the wishes of users who want their software to not be used in ways
that make it non-Free.

As it so happens, this also places economic pressure on vendors who
tivoize, such that they either face more costly solutions, or enable
users to tinker with the software as well.  And if neither ROM nor
permission are an option for the vendor, well, too bad, the author
gets to decide how his software is to be used, right?

> So what exactly is the point of all this then?

Keeping Free Software Free.

(and, as a consequence that many of you may welcome, keeping
open-source software open source)

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Jan Harkes]

On Tue, Jun 19, 2007 at 06:20:24PM -0300, Alexandre Oliva wrote:
> On Jun 19, 2007, Jan Harkes <jaharkes@cs.cmu.edu> wrote:
> > and which will most likely make GPLv3 software unusable for various
> > applications ranging from medical equipment to financial transaction
> > systems (and probably others)
> 
> Not unusable, except perhaps for the one example about credit card
> terminals presented so far.
> 
> > is there to just make it a _bit_ more inconvenient for vendors to
> > implement a tivo-like scheme?
> 
> I'm not sure they find it to be "just a bit".
> 
> Point is to keep Free Software Free freedoms, and ROM doesn't make it
> non-Free, so this provision is a means to ensure the compliance with
> the wishes of users who want their software to not be used in ways
> that make it non-Free.

You keep referring to the four freedoms so I googled for them and found

    http://www.gnu.org/philosophy/free-sw.html

So which of the freedoms did Tivo take away?

    * The freedom to run the program, for any purpose (freedom 0).

    * The freedom to study how the program works, and adapt it to
      your needs (freedom 1). Access to the source code is a
      precondition for this.

    * The freedom to redistribute copies so you can help your neighbor
      (freedom 2).

    * The freedom to improve the program, and release your improvements
      to the public, so that the whole community benefits (freedom 3).
      Access to the source code is a precondition for this.

It doesn't seem to me they took away freedoms 1, 2 or 3. They released
the source to any free software components and we can study, modify,
redistribute, improve and release our improvements for the benefit of
the whole community.

btw. freedom 3 seems to be just repeating what we already got from
freedoms 1 and 2.

So the only one we could differ in opinion about is freedom 0. I would
say that they in no way are limiting my use of the Linux kernel (which
is the part I mostly care about) I can run the program for any purpose I
see fit. What if I want to run mythtv on my PC at home? Tivo has no
control whether or not I can do so even when my kernel contains any of
their modification or improvements, so I claim that I in fact retained
freedom 0.

Your position (and I hope I will get this right), is that they should
allow you to run the program for any purpose _on the hardware it is
installed on_. Interpreted that way, the whole modification part doesn't
even come into play, your freedom is taken away if you cannot even run
your own software on top of the already installed kernel.

This 'freedom 0' the way (I hope) you are interpreting it, could clearly
never have been protected by the GPLv2 since it explicitly does not
cover "Activities other than copying, distribution and modification"
(GPLv2, term 0)

However the GPLv3 does not seem to address this point either, the whole
discussion in section 5 about user products and installation information
completely misses the fact that none of the 'four freedoms' (which I
assume formed the foundation for the license) is about allowing a user
to install the program on some particular piece of hardware and that is
exactly where I think all this anti-tivoization language is going wrong.
It is clearly having a hard time pinning down the exact requirements for
something that was not well defined in the first place.

The real issue with Tivo isn't that they signed their kernel or their
initrd, but that they do not allow you to use that kernel for any
purpose, for instance run mythtv or a webserver on their unmodified
kernel. Maybe the GPLv3 shouldn't try to talk about license keys or
installation information and then heap on some exceptions for
non-consumer devices or rom-based implementations and then some further
legaleze patches to close the really obvious loopholes. Maybe it if it
actually addressed the fact that you want to be able to use the
distributed software for any purpose you see fit by allowing you to run
your own applications on the kernel they distributed.

I still believe they do allow me to use the program for any purpose as
they can not limit my use of the Linux kernel whether or not my copy
contains any of their contributions, so excuse my while I'll go enjoy
some more of my freedom 0.

Jan

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Dave Neuer]

On 6/19/07, Jan Harkes <jaharkes@cs.cmu.edu> wrote:
>
> You keep referring to the four freedoms so I googled for them and found
>
>     http://www.gnu.org/philosophy/free-sw.html
>
> So which of the freedoms did Tivo take away?
>
>     * The freedom to run the program, for any purpose (freedom 0).
>
>     * The freedom to study how the program works, and adapt it to
>       your needs (freedom 1). Access to the source code is a
>       precondition for this.
>
>     * The freedom to redistribute copies so you can help your neighbor
>       (freedom 2).
>
>     * The freedom to improve the program, and release your improvements
>       to the public, so that the whole community benefits (freedom 3).
>       Access to the source code is a precondition for this.
>
> It doesn't seem to me they took away freedoms 1, 2 or 3. They released
> the source to any free software components and we can study, modify,
> redistribute, improve and release our improvements for the benefit of
> the whole community.
>
> btw. freedom 3 seems to be just repeating what we already got from
> freedoms 1 and 2.
>
> So the only one we could differ in opinion about is freedom 0. I would
> say that they in no way are limiting my use of the Linux kernel (which
> is the part I mostly care about) I can run the program for any purpose I
> see fit. What if I want to run mythtv on my PC at home? Tivo has no
> control whether or not I can do so even when my kernel contains any of
> their modification or improvements, so I claim that I in fact retained
> freedom 0.

Much as I hate to extend the life of this execrable thread, since I
think Alexandre makes Sisyphus look like a hard-nosed pragmatist, it
seems pretty clear that TiVO impinges "[my] freedom to run the
program, for any purpose" if "any purpose" includes "make my TiVO do
what I want," and likewise to "adapt it to [my] needs" -- freedoms 0
and part of 1. It is just disingenuous to argue otherwise.

Dave

RE: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Schwartz]

> Much as I hate to extend the life of this execrable thread, since I
> think Alexandre makes Sisyphus look like a hard-nosed pragmatist, it
> seems pretty clear that TiVO impinges "[my] freedom to run the
> program, for any purpose" if "any purpose" includes "make my TiVO do
> what I want," and likewise to "adapt it to [my] needs" -- freedoms 0
> and part of 1. It is just disingenuous to argue otherwise.
>
> Dave

The freedom to "run the program, for any purpose" is just as much violated
by Microsoft when they make the Xbox. You can't run the Linux kernel on that
either, for the exact same reasons you can't run a modified Linux kernel on
the Tivo. It is manifestly obvious that the "freedom to run the program for
any purpose" must be limited to those pieces of hardware where you are the
one who decides what hardware runs on that software. (And GPL rights were
always applicable equally to all hardware in the universe, not specially to
some hardware and not others.)

DS

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 20, 2007, "David Schwartz" <davids@webmaster.com> wrote:

> The freedom to "run the program, for any purpose" is just as much
> violated by Microsoft when they make the Xbox.

That's correct.

> You can't run the Linux kernel on that either, for the exact same
> reasons you can't run a modified Linux kernel on the Tivo.

Technically speaking, yes.

Legally speaking, these are completely different situations.

First of all, Microsoft is not in any way involved in how you received
the copy of the kernel Linux you'd like to run on the Xbox.  If it is,
you might have a claim that they couldn't impose this restriction on
you.  IANAL.

Second, you're not a user of Linux on the Xbox in the first place, so
you don't have an inherent right to run the program there that the
hardware manufacturer is denying you.  It's not much different than
wanting to run the program on on computers you have remote access to,
on which you have no permission to install Linux, or on computers you
have but with incompatible hardware architectures.

The case of incompatible hardware architectures is actually a bit
different, since nobody is really stopping you from running the
program there, it would just take you porting work to get to run it.

While in both the remote-access case and the Xbox case someone *is*
indeed stopping you from installing and running the software on those
computers, this does not render the software non-free, because you're
not a user of the software there in the first place, and stating that,
in order for software to be free, you'd have to have the ability to
install it on any existing computer in the world would be nonsensical.

However, in the case of TiVo, you have received the software from TiVo
with the express purpose of running it on the hardware they sold you,
and you can't install or run modified versions precisely because they
don't want to let you.  They're denying you a freedom that you are
entitled to have, which renders that binary copy of Linux installed on
it non-Free, and at the same time they're imposing further
restrictions on your freedoms that the GPL is designed to respect and
defend, which violates the spirit, if not the letter, of the GPL.

So you see, these are two very different cases, the key difference
being the obligation that TiVo accepted towards the copyright holders
of Linux, when it distributed the software to you, of respecting your
freedoms, by not imposing restrictions on your exercise of the
freedoms in addition the restrictions that the license itself imposes.

> (And GPL rights were always applicable equally to all hardware in
> the universe, not specially to some hardware and not others.)

Correct.  Whoever distributed you the software entitled you to enjoy
the freedoms wherever you manage to run the software.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Jesper Juhl]

On 19/06/07, Alexandre Oliva <aoliva@redhat.com> wrote:
> On Jun 18, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:
> > In the GPLv3 world, we have already discussed in this thread how you can
> > follow the GPLv3 by making the TECHNICALLY INFERIOR choice of using a ROM
> > instead of using a flash device.
>
> Yes.  This is one option that doesn't bring any benefits to anyone.
> It maintains the status quo for users and the community, but it loses
> the ability for the vendor to upgrade, fix or otherwise control the
> users.  Bad for the vendor.
>
Also bad for the user since now the vendor can no longer supply
updated firmware to fix bugs or otherwise improve the device. Also,
the vendow has the problem that devices that get returned for repair
or similar can't easily be updated software wise. For the vendor not
to be able to update the box creates a lot of problems for both the
vendor and the end user.


> As another option, the vendor can respect users' freedoms, and then
> everybody wins big.  That's the option that anti-tivoization provides
> economic incentive for vendors to take.  Sure, they may still prefer
> the alternative above, or stick with an older version (which has its
> costs), or move to different software (which also has its costs), but
> it's unreasonable to claim that I'm advocating for vendors to move to
> ROM.
>
I am fairly confident that if too much software switches to GPLv3
we'll see a lot of businesses move to BSD or proprietary software.
That means *we* lose bigtime.

-- 
Jesper Juhl <jesper.juhl@gmail.com>
Don't top-post  http://www.catb.org/~esr/jargon/html/T/top-post.html
Plain text mails only, please      http://www.expita.com/nomime.html

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Tomas Neme]

I have been following this discussion for the last week or so, and
what I haven't been able to figure out is what the hell is the big
deal with TiVO doing whatever they want to with their stupid design.
They made a design, they build a machine, they sell it as is, and
provide source code for GPL'ed software... what's your problem?

In order to play backuped games in, say, a PS2 I need to modify it's
hardware. Gut it up, solder some cables to a chip that will bypass the
signature key reading in the CD/DVDs. It's legal to mod it (it's my
hardware, I paid for it) although it voids my warranty. It's also
legal to copy the games (where I live) if I own the original, but
we're no talking about DMR here. Not exactly

Now, I'm sure some alike modification can be made to "help" TiVO
bypassing the signature key check... and I'm sure it's legal. Again,
what is your problem?

Linus said it already, but TiVO has seen himself FORCED to protect its
software for the very same reason Microsoft had to uncanningly
overhead EVERYTHING in Vista: the Media (Fox, Universal, Etcetera)
companies forced them too. But this is just a comment.. the thing is
that I really can't see the problem

T

-- 
|_|0|_|
|_|_|0|
|0|0|0|

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Michael Poole]

Tomas Neme writes:

> I have been following this discussion for the last week or so, and
> what I haven't been able to figure out is what the hell is the big
> deal with TiVO doing whatever they want to with their stupid design.
> They made a design, they build a machine, they sell it as is, and
> provide source code for GPL'ed software... what's your problem?

It's simple: they don't provide _complete_ source code.  They keep the
source code for the part of their Linux kernel images that provides
the functionality "runs on Tivo DVRs".  The GPL requires that
distributors of binary versions provide complete source code, not just
the parts of source code that are convenient.

Michael Poole

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[H. Peter Anvin]

Michael Poole wrote:
> Tomas Neme writes:
> 
>> I have been following this discussion for the last week or so, and
>> what I haven't been able to figure out is what the hell is the big
>> deal with TiVO doing whatever they want to with their stupid design.
>> They made a design, they build a machine, they sell it as is, and
>> provide source code for GPL'ed software... what's your problem?
> 
> It's simple: they don't provide _complete_ source code.  They keep the
> source code for the part of their Linux kernel images that provides
> the functionality "runs on Tivo DVRs".  The GPL requires that
> distributors of binary versions provide complete source code, not just
> the parts of source code that are convenient.
> 

If true, that's a clear violation of the GPLv2, but it's very different
from preventing the loading of modified images on their hardware (which
appears to be permitted by v2.)

	-hpa

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[david]

On Wed, 20 Jun 2007, H. Peter Anvin wrote:

> Michael Poole wrote:
>> Tomas Neme writes:
>>
>>> I have been following this discussion for the last week or so, and
>>> what I haven't been able to figure out is what the hell is the big
>>> deal with TiVO doing whatever they want to with their stupid design.
>>> They made a design, they build a machine, they sell it as is, and
>>> provide source code for GPL'ed software... what's your problem?
>>
>> It's simple: they don't provide _complete_ source code.  They keep the
>> source code for the part of their Linux kernel images that provides
>> the functionality "runs on Tivo DVRs".  The GPL requires that
>> distributors of binary versions provide complete source code, not just
>> the parts of source code that are convenient.
>>
>
> If true, that's a clear violation of the GPLv2, but it's very different
> from preventing the loading of modified images on their hardware (which
> appears to be permitted by v2.)
>
> 	-hpa

this is very much NOT true. if you take the source the provide you can 
compile a kernel that will run on the tivo, the only thing you have to do 
(on some models) is to change the bios to skip the step that checks if the 
kernel has been tampered with.

in fact, people have ported the nessary changes all the way up to the 2.6 
kernels (the tivos run a 2.1 derived kernel). you can buy hard drives to 
plug into your tivo with newer kernels (and yes, you do get source from 
those folks as well)

by the way, the first hacks that disabled these checks were done by people 
who were interested in getting their tivo to boot faster (they found an 
expensive bios routing that wasn't nessasary to boot), they weren't 
interested in loading different software on the box. when they released 
the hack other people realized that they had opened the door for other 
changes.

David Lang

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Michael Poole]

david@lang.hm writes:

> this is very much NOT true. if you take the source the provide you can
> compile a kernel that will run on the tivo, the only thing you have to
> do (on some models) is to change the bios to skip the step that checks
> if the kernel has been tampered with.

If we are opining whether Tivo provided complete source code for their
Linux kernel images, the requirement to change non-GPLed software as a
condition to exercise GPL-protected rights speaks for itself.

Out of curiosity, what do you have to do on models besides those?  Are
newer models more or less restrictive in what they run?  If newer
models are more restrictive, I think that also speaks to whether Tivo
thinks it is conveying complete source code.

Michael Poole

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[david]

On Wed, 20 Jun 2007, Michael Poole wrote:

> david@lang.hm writes:
>
>> this is very much NOT true. if you take the source the provide you can
>> compile a kernel that will run on the tivo, the only thing you have to
>> do (on some models) is to change the bios to skip the step that checks
>> if the kernel has been tampered with.
>
> If we are opining whether Tivo provided complete source code for their
> Linux kernel images, the requirement to change non-GPLed software as a
> condition to exercise GPL-protected rights speaks for itself.

no, the GPL protected rights don't say anything about the hardware the 
system runs on.

you are saying that the GPL now controls what the BIOS software is allowed 
to do or not allowed to do.

that's a seperate body of code that is in no way derived from the linux 
kernel (even the anti-tampering functions would work equally well with 
other Operating systems and are in no way linux specific). it's no even 
loaded on the same media (the BIOS is in flash/rom on the botherboard, the 
OS is on the hard drive)

and note that the software that is checked to make sure that it hasn't 
been changed includes much more then the kernel. it checks the kernel and 
the initrd.

> Out of curiosity, what do you have to do on models besides those?  Are
> newer models more or less restrictive in what they run?  If newer
> models are more restrictive, I think that also speaks to whether Tivo
> thinks it is conveying complete source code.

newer models do tend to be more restrictive, but they also tend to connect 
to more propriatary networks (satellite or cable)

David Lang

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Michael Poole]

david@lang.hm writes:

> On Wed, 20 Jun 2007, Michael Poole wrote:
>
>> david@lang.hm writes:
>>
>>> this is very much NOT true. if you take the source the provide you can
>>> compile a kernel that will run on the tivo, the only thing you have to
>>> do (on some models) is to change the bios to skip the step that checks
>>> if the kernel has been tampered with.
>>
>> If we are opining whether Tivo provided complete source code for their
>> Linux kernel images, the requirement to change non-GPLed software as a
>> condition to exercise GPL-protected rights speaks for itself.
>
> no, the GPL protected rights don't say anything about the hardware the
> system runs on.
>
> you are saying that the GPL now controls what the BIOS software is
> allowed to do or not allowed to do.

Please retract that claim.  I have said no such thing, and have
avoided saying anything that I thought might be misconstrued in that
direction.

To be absolutely clear: My complaints with Tivo as a hardware or BIOS
vendor are moral and pragmatic, not legal.  My complaint with Tivo as
a distributor of Linux is what hinges on legal issues.

> that's a seperate body of code that is in no way derived from the
> linux kernel (even the anti-tampering functions would work equally
> well with other Operating systems and are in no way linux
> specific). it's no even loaded on the same media (the BIOS is in
> flash/rom on the botherboard, the OS is on the hard drive)
>
> and note that the software that is checked to make sure that it hasn't
> been changed includes much more then the kernel. it checks the kernel
> and the initrd.

Not legally relevant.

>> Out of curiosity, what do you have to do on models besides those?  Are
>> newer models more or less restrictive in what they run?  If newer
>> models are more restrictive, I think that also speaks to whether Tivo
>> thinks it is conveying complete source code.
>
> newer models do tend to be more restrictive, but they also tend to
> connect to more propriatary networks (satellite or cable)

What they connect to is also not relevant.  That imples that because a
vendor has been issued or licensed patents, they are not obliged to
follow the GPL -- that the vendor has other obligations that supercede
the GPL's license claims.  GPL section 7 addresses that situation.

Michael Poole

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[david]

On Wed, 20 Jun 2007, Michael Poole wrote:

> david@lang.hm writes:
>
>> On Wed, 20 Jun 2007, Michael Poole wrote:
>>
>>> david@lang.hm writes:
>>>
>>>> this is very much NOT true. if you take the source the provide you can
>>>> compile a kernel that will run on the tivo, the only thing you have to
>>>> do (on some models) is to change the bios to skip the step that checks
>>>> if the kernel has been tampered with.
>>>
>>> If we are opining whether Tivo provided complete source code for their
>>> Linux kernel images, the requirement to change non-GPLed software as a
>>> condition to exercise GPL-protected rights speaks for itself.
>>
>> no, the GPL protected rights don't say anything about the hardware the
>> system runs on.
>>
>> you are saying that the GPL now controls what the BIOS software is
>> allowed to do or not allowed to do.
>
> Please retract that claim.  I have said no such thing, and have
> avoided saying anything that I thought might be misconstrued in that
> direction.
>
> To be absolutely clear: My complaints with Tivo as a hardware or BIOS
> vendor are moral and pragmatic, not legal.  My complaint with Tivo as
> a distributor of Linux is what hinges on legal issues.

but if the GPL doesn't control the BIOS how in the world are you saying 
that the fact that the GPL covers the kernel makes what the BIOS does 
wrong (even if the kernel was covered by GPLv3)?

>> that's a seperate body of code that is in no way derived from the
>> linux kernel (even the anti-tampering functions would work equally
>> well with other Operating systems and are in no way linux
>> specific). it's no even loaded on the same media (the BIOS is in
>> flash/rom on the botherboard, the OS is on the hard drive)
>>
>> and note that the software that is checked to make sure that it hasn't
>> been changed includes much more then the kernel. it checks the kernel
>> and the initrd.
>
> Not legally relevant.

I disagree. it's very relevant if your argument is that becouse the 
checksum if a checksum of the kernel that the license for the kernel 
somehow controlls what can be done with it.

>>> Out of curiosity, what do you have to do on models besides those?  Are
>>> newer models more or less restrictive in what they run?  If newer
>>> models are more restrictive, I think that also speaks to whether Tivo
>>> thinks it is conveying complete source code.
>>
>> newer models do tend to be more restrictive, but they also tend to
>> connect to more propriatary networks (satellite or cable)
>
> What they connect to is also not relevant.  That imples that because a
> vendor has been issued or licensed patents, they are not obliged to
> follow the GPL -- that the vendor has other obligations that supercede
> the GPL's license claims.  GPL section 7 addresses that situation.

you are arguing that the fact that later models are more locked down is 
'proof' that tivo knows that it's doing the wrong thing. I'm pointing out 
an alternate reason, the people who control the networks that the newer 
models are connecting to are imposing additional restrictions on tivo that 
cause them to need to be locked down more.

David Lang

> Michael Poole
>

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Michael Poole]

david@lang.hm writes:

> On Wed, 20 Jun 2007, Michael Poole wrote:
>
>> Please retract that claim.  I have said no such thing, and have
>> avoided saying anything that I thought might be misconstrued in that
>> direction.
>>
>> To be absolutely clear: My complaints with Tivo as a hardware or BIOS
>> vendor are moral and pragmatic, not legal.  My complaint with Tivo as
>> a distributor of Linux is what hinges on legal issues.
>
> but if the GPL doesn't control the BIOS how in the world are you
> saying that the fact that the GPL covers the kernel makes what the
> BIOS does wrong (even if the kernel was covered by GPLv3)?

I do not say that the BIOS is doing anything (legally) wrong.  The
wrong act is distributing the binary kernel image without distributing
complete source code for it.

>>> that's a seperate body of code that is in no way derived from the
>>> linux kernel (even the anti-tampering functions would work equally
>>> well with other Operating systems and are in no way linux
>>> specific). it's no even loaded on the same media (the BIOS is in
>>> flash/rom on the botherboard, the OS is on the hard drive)
>>>
>>> and note that the software that is checked to make sure that it hasn't
>>> been changed includes much more then the kernel. it checks the kernel
>>> and the initrd.
>>
>> Not legally relevant.
>
> I disagree. it's very relevant if your argument is that becouse the
> checksum if a checksum of the kernel that the license for the kernel
> somehow controlls what can be done with it.

To the extent that it is relevant, it strengthens the argument against
Tivo: they are tying together many works of authorship, including some
GPLed works, in a way that makes them effectively inseparable.  This
is beyond "mere aggregation" on a distribution medium, and tends to
implicate *all* parts of the whole as GPL encumbered.

Michael Poole

RE: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Schwartz]

> I do not say that the BIOS is doing anything (legally) wrong.  The
> wrong act is distributing the binary kernel image without distributing
> complete source code for it.

Why are you not complaining that Linus does not distribute the keys he uses
to sign kernel source distributions? If a digital signature is part of the
distribution, why is the key used to produce that signature not part of the
distribution?

If you can cite some legal reason there is a difference, I would be quite
impressed.

In any event, the argument is obvious nonsense. The signature is merely
aggregated with the kernel. Cooperation, dependent function, and convergent
design can't break mere aggregation or you get ridiculous results. (For
example, a device shipped with the Linux kernel and some applications would
have to GPL all the applications.)

DS

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Michael Poole]

David Schwartz writes:

>> I do not say that the BIOS is doing anything (legally) wrong.  The
>> wrong act is distributing the binary kernel image without distributing
>> complete source code for it.
>
> Why are you not complaining that Linus does not distribute the keys he uses
> to sign kernel source distributions? If a digital signature is part of the
> distribution, why is the key used to produce that signature not part of the
> distribution?
>
> If you can cite some legal reason there is a difference, I would be quite
> impressed.
>
> In any event, the argument is obvious nonsense. The signature is merely
> aggregated with the kernel. Cooperation, dependent function, and convergent
> design can't break mere aggregation or you get ridiculous results. (For
> example, a device shipped with the Linux kernel and some applications would
> have to GPL all the applications.)

Do you make it a habit to pose ranty questions to people while neither
attributing their text nor cc'ing them?  Especially when you claim the
person's argument is "obvious nonsense", it seems quite rude.

(Since you have dismissed my argument as nonsense before hearing my
response, I will not bother answering your question.  Since you are
acting like a troll, I will dismiss you as one.  Most of this list has
already dismissed your rather unique -- I would even say frivolous --
idea of how far "mere aggregation" goes: I, for one, have better
things to do than explain why a C file is not a "mere aggregation" of
the functions it contains.)

Michael Poole

RE: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Schwartz]

> Most of this list has
> already dismissed your rather unique -- I would even say frivolous --
> idea of how far "mere aggregation" goes: I, for one, have better
> things to do than explain why a C file is not a "mere aggregation" of
> the functions it contains.)
>
> Michael Poole

Of course it's not mere aggregation. The functions in a C file are
creatively combined. How many times do I have to say that the opposite of
"mere aggregation" is creative combination?

It is not unique, it is part of the definition of a "derivative work".

DS

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Michael Poole]

David Schwartz writes:

>> Most of this list has
>> already dismissed your rather unique -- I would even say frivolous --
>> idea of how far "mere aggregation" goes: I, for one, have better
>> things to do than explain why a C file is not a "mere aggregation" of
>> the functions it contains.)
>>
>> Michael Poole
>
> Of course it's not mere aggregation. The functions in a C file are
> creatively combined. How many times do I have to say that the opposite of
> "mere aggregation" is creative combination?
>
> It is not unique, it is part of the definition of a "derivative work".

By "creative combination" do you mean what US copyright law refers to
as compilations (or their subset collective works)?

Compilations can be creative combinations while still being mere
aggregation under the GPL.  For example, if applications are selected
to run with a Linux kernel, and they are distributed together, the
collection is a creative selection -- and this seems to be one of the
cases evoked by the GPL's reference to "mere aggregation".  See also
practically every Linux distribution on the planet.

Compilations also can be creative combinations and *more* than mere
aggregation: for example, Linux with respect to its subsystems, or any
case where a larger work is derivative of one of its components.

However, compilations (even to the extent they are creative
combinations) are not necessarily derivative works of their elements.
For more details, see
http://www.copyright.gov/circs/circ14.html#compilations

Michael Poole

RE: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Schwartz]

> By "creative combination" do you mean what US copyright law refers to
> as compilations (or their subset collective works)?

Not only. By "creative combination" I mean either a compilation or a
derivative work. I was a bit unclear about that because I wasn't really
addressing compilation rights at the time.

For example, if you adapt a FreeBSD driver to work on Linux, you may be
creatively combining aspects of the driver with code from the kernel. The
result is one that you have copyright interest in because it is a derivative
work but probably not a compilation copyright. The choice of one driver and
one OS, where the goal is to make the driver work on the OS, probably is not
sufficiently creative to justify a compilation copyright. However, this is
clearly not mere aggregation if significant changes are needed to make the
driver work with Linux.

> Compilations can be creative combinations while still being mere
> aggregation under the GPL.  For example, if applications are selected
> to run with a Linux kernel, and they are distributed together, the
> collection is a creative selection -- and this seems to be one of the
> cases evoked by the GPL's reference to "mere aggregation".  See also
> practically every Linux distribution on the planet.

You are quite correct. In this case, the GPL may not require you to license
the compilation copyright. I believe this is so even if all the works are
covered by the GPL. Arguably, that's a defect in the GPL because it means
there might be situations in which you might receive a CD that contains only
GPL'd software and not be able to redistribute it due to a compilation
copyright.

I honestly have no position on whether "mere aggregation" should include
aggregating works where there is sufficient creative input to justify a
compilation copyright on the result. I think either position can be argued.
I think the intent of the GPL was probably that mere aggregation not include
compilation rights because that leads to strange results. I don't know of
any evidence that compilation rights were considered when the GPL was
written. If so, the deliberate lack of mention might weigh in the balance.

> Compilations also can be creative combinations and *more* than mere
> aggregation: for example, Linux with respect to its subsystems, or any
> case where a larger work is derivative of one of its components.

Of course. If I write a Linux kernel module, it might be a derivative work
because it contains significant portions of the Linux kernel source code.
This is true before anyone compiles it or links it.

When I say linking cannot create a derivative work, I mean assuming the work
was not derivative in the first place. I am also further assuming there is
insufficient creativity in the choice of which works to link to justify a
compilation copyright.

> However, compilations (even to the extent they are creative
> combinations) are not necessarily derivative works of their elements.
> For more details, see
> http://www.copyright.gov/circs/circ14.html#compilations

Because compilation copyrights don't really affect the Tivo and GPLv2/GPLv3
issue, I tend to ignore them when discussing that subject. If you think I'm
wrong and there is some relationship between them, please let me know. I
admit I may not have given that possibility enough thought.

DS

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Michael Poole]

David Schwartz writes:

>> However, compilations (even to the extent they are creative
>> combinations) are not necessarily derivative works of their elements.
>> For more details, see
>> http://www.copyright.gov/circs/circ14.html#compilations
>
> Because compilation copyrights don't really affect the Tivo and GPLv2/GPLv3
> issue, I tend to ignore them when discussing that subject. If you think I'm
> wrong and there is some relationship between them, please let me know. I
> admit I may not have given that possibility enough thought.

I believe compilation copyrights do bear on GPL-licensed software, by
virtue of the GPL's sentence "[...] rather, the intent is to exercise
the right to control the distribution of derivative _or collective_
works based on the Program." (emphasis added).

There is a lot of grey and/or arguable area about what constitutes a
GPL-encumbered collective work versus mere aggregation.  Although I
disagree, I understand and respect that some believe that the kernel
plus a digital signature over it is "mere aggregation".  I would like
to focus the discussion on that question, though, rather than whether
the GPL is worded to control the rights to compilations-in-general
that include GPLed works.

Michael Poole

RE: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Schwartz]

> I believe compilation copyrights do bear on GPL-licensed software, by
> virtue of the GPL's sentence "[...] rather, the intent is to exercise
> the right to control the distribution of derivative _or collective_
> works based on the Program." (emphasis added).

Ahh, good. So there's no problem with the GPL. I already thought it the most
sensible reading that it included collective works (and it's clear that it
can legally do so), but that makes it clear.

I didn't mean that compilation copyrights have no relevance to GPL issues in
general, just none to the issue of GPLv2 versus GPLv3 and Tivoization. Are
you going to argue that there's a compilation copyright justified by
combining a kernel binary with a signature for that binary? That seems
untenable to me.

> There is a lot of grey and/or arguable area about what constitutes a
> GPL-encumbered collective work versus mere aggregation.

I think it's technically/legally clear what the standards are, but certainly
arguable whether particular works meet that standard. If the choice of works
to combine is sufficiently creative (above and beyond any choices dictated
by functional considerations), it's a GPL-encumbered collective work.

I don't think it's arguable that a signature shipped along with a binary is
a collective work. In any event, if that were true, I think we should be
able to agree that Linus would be required to release his kernel signing
keys.

> Although I
> disagree, I understand and respect that some believe that the kernel
> plus a digital signature over it is "mere aggregation".

It clearly is. What else could it be? The digital signature is a separate
item, a pure number for which there is no copyright interest, that is simply
appended to the kernel. It does not contain significant protected elements
of the kernel. No creative process is used to generate it or attach it. The
decision to append is dictated by purely functional considerations (nobody
creatively picks which signature to bundle with which kernel).

> I would like
> to focus the discussion on that question, though, rather than whether
> the GPL is worded to control the rights to compilations-in-general
> that include GPLed works.

If the kernel plus a digital signature over it is not mere aggregation, then
Linus is violating the GPL by shipping kernel plus digital signatures but
not including the "source code" to produce the signatures. If the
ridiculousness of that is not sufficiently obvious, I'm not sure what else
to say. Where is the outcry that Linus is keeping his signing key secret,
failing to include the source code that he used to build the Linux kernel
distibution?

The past few times this has come up, every possible irrelevent side issue
was raised. For example, that the signature in the case of Linux is not
functional. These considerations do not have anything to do with whether the
combination is mere aggregation or not.

Ironically, this case is even clearer than linking. The two works are quite
literally stapled together. There is no intermingling at all.

DS

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Michael Poole]

David Schwartz writes:

>> There is a lot of grey and/or arguable area about what constitutes a
>> GPL-encumbered collective work versus mere aggregation.
>
> I think it's technically/legally clear what the standards are, but certainly
> arguable whether particular works meet that standard. If the choice of works
> to combine is sufficiently creative (above and beyond any choices dictated
> by functional considerations), it's a GPL-encumbered collective work.
>
> I don't think it's arguable that a signature shipped along with a binary is
> a collective work. In any event, if that were true, I think we should be
> able to agree that Linus would be required to release his kernel signing
> keys.

The distinction between GPL-covered works and "mere aggregation" is
not a function only of legal classifications.  If it were, the GPL
would be worded differently than it is -- and have different effects
than most people believe it does.

Michael Poole

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[david]

On Wed, 20 Jun 2007, Michael Poole wrote:

> David Schwartz writes:
>
>>> However, compilations (even to the extent they are creative
>>> combinations) are not necessarily derivative works of their elements.
>>> For more details, see
>>> http://www.copyright.gov/circs/circ14.html#compilations
>>
>> Because compilation copyrights don't really affect the Tivo and GPLv2/GPLv3
>> issue, I tend to ignore them when discussing that subject. If you think I'm
>> wrong and there is some relationship between them, please let me know. I
>> admit I may not have given that possibility enough thought.
>
> I believe compilation copyrights do bear on GPL-licensed software, by
> virtue of the GPL's sentence "[...] rather, the intent is to exercise
> the right to control the distribution of derivative _or collective_
> works based on the Program." (emphasis added).
>
> There is a lot of grey and/or arguable area about what constitutes a
> GPL-encumbered collective work versus mere aggregation.  Although I
> disagree, I understand and respect that some believe that the kernel
> plus a digital signature over it is "mere aggregation".  I would like
> to focus the discussion on that question, though, rather than whether
> the GPL is worded to control the rights to compilations-in-general
> that include GPLed works.

if the GPL can excercise control over compilations, then if Oracle were to 
ship a Oracle Linux live CD that contained the Oracle Database in the 
filesystem image, ready to run. then the GPL would be able to control the 
Oracle Database code.

if the GPL can't do this then it can't control the checksum either.

again, it's not just the kernel that's part of the checksum on a tivo, the 
checksum is over the kernel + initial filesystem, much of which contains 
code not covered by the gPL)

David Lang

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Michael Poole]

david@lang.hm writes:

> if the GPL can excercise control over compilations, then if Oracle
> were to ship a Oracle Linux live CD that contained the Oracle Database
> in the filesystem image, ready to run. then the GPL would be able to
> control the Oracle Database code.

By copyright law, it could.  By its language, it does not.

> if the GPL can't do this then it can't control the checksum either.
>
> again, it's not just the kernel that's part of the checksum on a tivo,
> the checksum is over the kernel + initial filesystem, much of which
> contains code not covered by the gPL)

Again, did you miss where I pointed out that this makes it *worse* for
Tivo, because they are tying together -- and making inseparable -- a
combination that would otherwise be "mere aggregation"?

Michael Poole

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[david]

On Wed, 20 Jun 2007, Michael Poole wrote:

> david@lang.hm writes:
>
>> if the GPL can excercise control over compilations, then if Oracle
>> were to ship a Oracle Linux live CD that contained the Oracle Database
>> in the filesystem image, ready to run. then the GPL would be able to
>> control the Oracle Database code.
>
> By copyright law, it could.  By its language, it does not.

many people (including many lawyers will disagree that it could by 
copyright law

>> if the GPL can't do this then it can't control the checksum either.
>>
>> again, it's not just the kernel that's part of the checksum on a tivo,
>> the checksum is over the kernel + initial filesystem, much of which
>> contains code not covered by the gPL)
>
> Again, did you miss where I pointed out that this makes it *worse* for
> Tivo, because they are tying together -- and making inseparable -- a
> combination that would otherwise be "mere aggregation"?

and it makes most distro CD's illegal since they contain code under 
different incompatible licenses and they make a checksum across the entire 
CD image.

David Lang

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Michael Poole]

david@lang.hm writes:

> On Wed, 20 Jun 2007, Michael Poole wrote:
>
>> david@lang.hm writes:
>>
>>> if the GPL can excercise control over compilations, then if Oracle
>>> were to ship a Oracle Linux live CD that contained the Oracle Database
>>> in the filesystem image, ready to run. then the GPL would be able to
>>> control the Oracle Database code.
>>
>> By copyright law, it could.  By its language, it does not.
>
> many people (including many lawyers will disagree that it could by
> copyright law

On what grounds would Oracle have a license to ship that part of
Linux?  Unless you are the sole copyright owner, you have no right to
copy a given piece of software _at all_ without a license.  The GPL is
a remarkably giving license in terms of how little it requires.

(This potentially wide scope is one of the major reasons that the GPL
mentions "mere aggregation" and that the Debian Free Software
Guidelines' include guideline #9.)

>>> if the GPL can't do this then it can't control the checksum either.
>>>
>>> again, it's not just the kernel that's part of the checksum on a tivo,
>>> the checksum is over the kernel + initial filesystem, much of which
>>> contains code not covered by the gPL)
>>
>> Again, did you miss where I pointed out that this makes it *worse* for
>> Tivo, because they are tying together -- and making inseparable -- a
>> combination that would otherwise be "mere aggregation"?
>
> and it makes most distro CD's illegal since they contain code under
> different incompatible licenses and they make a checksum across the
> entire CD image.

When distributions generate checksums (as opposed to signatures) over
images, they do provide all of the inputs.  When most distributions
provide signatures, the signatures do not function as statements or
instructions to computers -- they function as statements to users.
Tivo's digital signatures differ in both of these respects.

Michael Poole

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Lennart Sorensen]

On Wed, Jun 20, 2007 at 04:07:57PM -0400, Michael Poole wrote:
> I do not say that the BIOS is doing anything (legally) wrong.  The
> wrong act is distributing the binary kernel image without distributing
> complete source code for it.

So how about this idea then:

Tivo builds a kernel for their box, and release all the sources for how
to build exactly that kernel.

Tivo builds a bios image for their box, and encodes into it the checksum
of the kernel, or at least parts of it that they want to ensure are
present and unmodified.

Everytime the device boots, it checks the kernel image, and it the
checksums match, it loads and runs the kernel, and otherwise it checks
if there is a new bios image with a proper signature, updates itself and
reboots and tries again.

Preventing people from doing things with their own hardware certainly
seems morally wrong, but legally, I don't see any way to prevent it.

I suppose you could say in the license: You may not use this code in any
way if you do what the RIPP/MPAA/etc want you to do.

--
Len Sorensen

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[david]

On Thu, 21 Jun 2007, Lennart Sorensen wrote:

> 
> On Wed, Jun 20, 2007 at 04:07:57PM -0400, Michael Poole wrote:
>> I do not say that the BIOS is doing anything (legally) wrong.  The
>> wrong act is distributing the binary kernel image without distributing
>> complete source code for it.
>
> So how about this idea then:
>
> Tivo builds a kernel for their box, and release all the sources for how
> to build exactly that kernel.
>
> Tivo builds a bios image for their box, and encodes into it the checksum
> of the kernel, or at least parts of it that they want to ensure are
> present and unmodified.
>
> Everytime the device boots, it checks the kernel image, and it the
> checksums match, it loads and runs the kernel, and otherwise it checks
> if there is a new bios image with a proper signature, updates itself and
> reboots and tries again.

the bios doesn't have enough capability to talk to the outside world for 
updates.

what tivo actually does is very similar to this

they encode into the bios the ability to check a checksum/signature for 
the kernel+boot filesystem and if they don't match look to see if there is 
another kernel+boot filesystem available

then software on the boot filesystem checks to see if the rest of the 
system has been tampered with before it mounts /

> Preventing people from doing things with their own hardware certainly
> seems morally wrong, but legally, I don't see any way to prevent it.
>
> I suppose you could say in the license: You may not use this code in any
> way if you do what the RIPP/MPAA/etc want you to do.

the GPLv3 is trying to do this.

David Lang

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Lennart Sorensen]

On Thu, Jun 21, 2007 at 10:26:04AM -0700, david@lang.hm wrote:
> the bios doesn't have enough capability to talk to the outside world for 
> updates.

Of course, although perhaps it could.  More likely my thought was that
the service when it decides to download an update, would include the
updated bios image and put it on the boot drive where the existing bios
can find it.  No signature needs to be added to the boot drive or
kernel, just checksums in the bios image.

> what tivo actually does is very similar to this
> 
> they encode into the bios the ability to check a checksum/signature for 
> the kernel+boot filesystem and if they don't match look to see if there is 
> another kernel+boot filesystem available
> 
> then software on the boot filesystem checks to see if the rest of the 
> system has been tampered with before it mounts /
> 
> the GPLv3 is trying to do this.

Perhaps they should just explicitly say that then.

--
Len Sorensen

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[david]

On Thu, 21 Jun 2007, Lennart Sorensen wrote:

> On Thu, Jun 21, 2007 at 10:26:04AM -0700, david@lang.hm wrote:
>> the bios doesn't have enough capability to talk to the outside world for
>> updates.
>
> Of course, although perhaps it could.  More likely my thought was that
> the service when it decides to download an update, would include the
> updated bios image and put it on the boot drive where the existing bios
> can find it.  No signature needs to be added to the boot drive or
> kernel, just checksums in the bios image.
>
>> what tivo actually does is very similar to this
>>
>> they encode into the bios the ability to check a checksum/signature for
>> the kernel+boot filesystem and if they don't match look to see if there is
>> another kernel+boot filesystem available
>>
>> then software on the boot filesystem checks to see if the rest of the
>> system has been tampered with before it mounts /

you snippede the bit about not knowing how to stop it

>> the GPLv3 is trying to do this.
>
> Perhaps they should just explicitly say that then.

they call the section the anti-tivoization, how much more explicit can 
they get?

David Lang

by the way, just in case anyone is misunderstanding me. I don't believe 
for a moment that all these anti-tamper features actually work in the real 
world (the PS3 hacking kits are proof of the lengths people will go to to 
make the 'hard' hardware-level hacking trivial to do) but the approach 
needs to be at secure modulo hardware tampering or software bugs.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Lennart Sorensen]

On Thu, Jun 21, 2007 at 10:51:06AM -0700, david@lang.hm wrote:
> you snippede the bit about not knowing how to stop it

I did?  As far as I can tell I quoted it all.  What did I miss?

> they call the section the anti-tivoization, how much more explicit can 
> they get?

They could be as explicit as: 
You can't use this code if you cooporate with anyone that requires
DRM systems.

All their attempts to define user devices and such is just going to
screw up and miss some things they wanted covered, and disallow things
they didn't intend to disallow (assuming there is any such thing).

> by the way, just in case anyone is misunderstanding me. I don't believe 
> for a moment that all these anti-tamper features actually work in the real 
> world (the PS3 hacking kits are proof of the lengths people will go to to 
> make the 'hard' hardware-level hacking trivial to do) but the approach 
> needs to be at secure modulo hardware tampering or software bugs.

DRM is completely pointless.  It only stops casual end users from doing
things.  It doesn't stop anyone with any technical clue from doing
things.  I keep hoping one day the people in charge at the big media
companies will understand this, and stop asking for people to implement
it.  Of course in the mean time there are companies perfectly willing to
claim to have unbreakable DRM for sale, while knowing full well (if they
are competent) that it is a lie.  So as long as the people in charge at
big media are clueless about technology, and as long as there are
companies willing to lie to them for money, then we will probably
continue to have DRM crap to deal with.

I don't think the GPLv3 is the place to try to remove DRM.  What the FSF
should be doing is try to educate the people who are advocating the use
of DRM about the fact that it can't ever work.  You can make more and
more stupid laws about how people can't remove the DRM, but people who
break copyright obviously already are breaking the law, so what is the
point in having more lows for them to break.  That is where this problem
should be fought, not in the GPLv3.  The GPLv3 is never going to solve
the problem, only educating people can do that.

--
Len Sorensen

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[david]

On Thu, 21 Jun 2007, Lennart Sorensen wrote:

> On Thu, Jun 21, 2007 at 10:51:06AM -0700, david@lang.hm wrote:
>> you snippede the bit about not knowing how to stop it
>
> I did?  As far as I can tell I quoted it all.  What did I miss?
>
>> they call the section the anti-tivoization, how much more explicit can
>> they get?
>
> They could be as explicit as:
> You can't use this code if you cooporate with anyone that requires
> DRM systems.

I think their earlier versions did say this.

> All their attempts to define user devices and such is just going to
> screw up and miss some things they wanted covered, and disallow things
> they didn't intend to disallow (assuming there is any such thing).
>
>> by the way, just in case anyone is misunderstanding me. I don't believe
>> for a moment that all these anti-tamper features actually work in the real
>> world (the PS3 hacking kits are proof of the lengths people will go to to
>> make the 'hard' hardware-level hacking trivial to do) but the approach
>> needs to be at secure modulo hardware tampering or software bugs.
>
> DRM is completely pointless.  It only stops casual end users from doing
> things.  It doesn't stop anyone with any technical clue from doing
> things.  I keep hoping one day the people in charge at the big media
> companies will understand this, and stop asking for people to implement
> it.  Of course in the mean time there are companies perfectly willing to
> claim to have unbreakable DRM for sale, while knowing full well (if they
> are competent) that it is a lie.  So as long as the people in charge at
> big media are clueless about technology, and as long as there are
> companies willing to lie to them for money, then we will probably
> continue to have DRM crap to deal with.

DRM does have some legitimate uses, for example redhat installations not 
installing unsigned software is a form of DRM

> I don't think the GPLv3 is the place to try to remove DRM.  What the FSF
> should be doing is try to educate the people who are advocating the use
> of DRM about the fact that it can't ever work.  You can make more and
> more stupid laws about how people can't remove the DRM, but people who
> break copyright obviously already are breaking the law, so what is the
> point in having more lows for them to break.  That is where this problem
> should be fought, not in the GPLv3.  The GPLv3 is never going to solve
> the problem, only educating people can do that.

this is exactly what most of the people who are arguing against this 
provision are saying.

David Lang

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 21, 2007, david@lang.hm wrote:

> On Thu, 21 Jun 2007, Lennart Sorensen wrote:

>> You can't use this code if you cooporate with anyone that requires
>> DRM systems.

> I think their earlier versions did say this.

Show me a GPLv3 draft that did it?

Start here, section 3:
http://gplv3.fsf.org/gpl-draft-2006-01-16.html

> DRM does have some legitimate uses, for example redhat installations
> not installing unsigned software is a form of DRM

Doh.  Then chmod og-r is DRM too.  And stronger DRM while at that,
since the user denied permission to read the file cannot take it back,
whereas the verification of unsigned software is just a warning, that
you can often bypass by telling the software to go ahead and install
it regardless of signatures.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Tomas Neme]

> It's simple: they don't provide _complete_ source code.  They keep the
> source code for the part of their Linux kernel images that provides
> the functionality "runs on Tivo DVRs".  The GPL requires that

http://en.wikipedia.org/wiki/Tivoization does not agree that this is
the problem but rather "TiVo circumvented this goal by making their
products run programs only if the program's digital signature matches
those authorised by the manufacturer of the TiVo."

I'm downloading the sources now.. if they compile, then you're lying
to me, right? Moreover, if I compile them as is, and I can run them on
a TiVo (let's say upgrading the machine's kernel) then you're even
more so..

if this is all true, then I read GPLv2 and it tells me (clipping and ** mine):

0.  This License applies to any *program* or other work [...]. The
"Program", below, refers to any such program or work, and a "work
based on the Program" means either the Program or any derivative work
under copyright law: that is to say, *a work containing the Program or
a portion of it, either verbatim or with modifications* [...]

2.  You may modify your copy or copies of the Program or any portion
of it, thus forming a work based on the Program, and copy and
distribute such modifications [...]

3.  You may copy and distribute the Program (or a work based on it,
under Section 2) in object code or executable form under the terms of
Sections 1 and 2 above provided that you also do one of the following:

    a) Accompany it with the complete corresponding machine-readable
source code, which must be distributed under the terms of Sections 1
and 2 above on a medium customarily used for software interchange;
[...]

The source code for a work means the preferred form of the work for
making modifications to it. For an executable work, complete source
code means all the source code for all modules it contains, plus any
associated interface definition files, plus the scripts used to
control compilation and installation of the executable. However, as a
special exception, the source code distributed
*need*not*include*anything that is *normally*distributed (in *either*
source or *binary* form) with the major components (compiler, kernel,
and so on) of the operating system on which the executable runs,
unless that component itself accompanies the executable.

11.  BECAUSE THE PROGRAM IS LICENSED FREE OF CHARGE, THERE IS NO
WARRANTY FOR THE PROGRAM, TO THE EXTENT PERMITTED BY APPLICABLE LAW.
EXCEPT WHEN OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR
OTHER PARTIES PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY OF ANY
KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE
IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
PURPOSE. THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE
PROGRAM IS WITH YOU. SHOULD THE PROGRAM PROVE DEFECTIVE, YOU ASSUME
THE COST OF ALL NECESSARY SERVICING, REPAIR OR CORRECTION.

so, it very clearly says that the program and the modifications come
with NO WARRANTY of working, either expressed or implied, including
the warranty of fitness for a particular purpose (for example being
able to run in your TiVo box). Other than that the licence says you
can do whatever you want as long as you give out the source code.

Let me put it this way: if TiVo gave microsoft the keys to run in
TiVo, and Microsoft made a Windows Kernel for TiVo and it run, they
would not be needing to give you any source codes because the box
itself isn't GPL'ed and even if it was, there's no need to GLP stuff
that work on a system based on The Program (say Adobe finally ports
Photoshop into linux) but is not a modification or aggregation.
Whatever checks the kernel's key is outside of the kernel itself, and
probably not a modification of any GPL'ed software

plus, what I marked in point 3. says that they don't need to give you
anything that is normally distributed with the system the program runs
on. They don't need to give you the source or the design of their
hardware, or any embebbed software

so.. you're free to modify the kernel, try to figure out what TiVo
does for authentication and modify your binaries so they look what
TiVo wants them to.. add some garbage bits in the end, or something.

One more thing: even if the program they use to make their binaries
TiVo-key-check compliant was GPL'ed, they don't need to give it to
you, because the kernel does not use them. *IT* uses the compiled
kernel as input data (and produce signed kernel binaries as output),
not the other way around

Tomás

-- 
|_|0|_|
|_|_|0|
|0|0|0|

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Michael Poole]

Tomas Neme writes:

>> It's simple: they don't provide _complete_ source code.  They keep the
>> source code for the part of their Linux kernel images that provides
>> the functionality "runs on Tivo DVRs".  The GPL requires that
>
> http://en.wikipedia.org/wiki/Tivoization does not agree that this is
> the problem but rather "TiVo circumvented this goal by making their
> products run programs only if the program's digital signature matches
> those authorised by the manufacturer of the TiVo."
>
> I'm downloading the sources now.. if they compile, then you're lying
> to me, right? Moreover, if I compile them as is, and I can run them on
> a TiVo (let's say upgrading the machine's kernel) then you're even
> more so..

Whether the sources you download from Tivo compile says very little.
It certainly does not mean I am wrong.  I could give you binary for
version A and sources for version B -- and the sources would compile.

The kernel you build from the source code that Tivo distributes must
be accepted by Tivo's hardware without making other modifications (to
Tivo's hardware or bootloader).  If that is possible, I will retract
what I said.  If it is not possible, they are omitting part of the
program's source code:

    A "computer program" is a set of statements or instructions to be
    used directly or indirectly in a computer in order to bring about
    a certain result.
    -- US Code, Title 17, Section 101

Michael Poole

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[david]

On Wed, 20 Jun 2007, Michael Poole wrote:

> Tomas Neme writes:
>
>>> It's simple: they don't provide _complete_ source code.  They keep the
>>> source code for the part of their Linux kernel images that provides
>>> the functionality "runs on Tivo DVRs".  The GPL requires that
>>
>> http://en.wikipedia.org/wiki/Tivoization does not agree that this is
>> the problem but rather "TiVo circumvented this goal by making their
>> products run programs only if the program's digital signature matches
>> those authorised by the manufacturer of the TiVo."
>>
>> I'm downloading the sources now.. if they compile, then you're lying
>> to me, right? Moreover, if I compile them as is, and I can run them on
>> a TiVo (let's say upgrading the machine's kernel) then you're even
>> more so..
>
> Whether the sources you download from Tivo compile says very little.
> It certainly does not mean I am wrong.  I could give you binary for
> version A and sources for version B -- and the sources would compile.
>
> The kernel you build from the source code that Tivo distributes must
> be accepted by Tivo's hardware without making other modifications (to
> Tivo's hardware or bootloader).  If that is possible, I will retract
> what I said.  If it is not possible, they are omitting part of the
> program's source code:

no, saying that the result must be acceptable to other software (in this 
case the software running in the BIOS) is not part of the source code.

David Lang

>    A "computer program" is a set of statements or instructions to be
>    used directly or indirectly in a computer in order to bring about
>    a certain result.
>    -- US Code, Title 17, Section 101
>
> Michael Poole
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
>

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Michael Poole]

david@lang.hm writes:

> no, saying that the result must be acceptable to other software (in
> this case the software running in the BIOS) is not part of the source
> code.

Why not?  The digital signature is a statement (which translates
roughly to "Tivo approves this") to be used in a computer in order to
bring about a certain result.  That result is making it boot on the
PVR.  Source code simply means the original forms or inputs used to
generate machine-readable statements.

Michael Poole

>>    A "computer program" is a set of statements or instructions to be
>>    used directly or indirectly in a computer in order to bring about
>>    a certain result.
>>    -- US Code, Title 17, Section 101

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[david]

On Wed, 20 Jun 2007, Michael Poole wrote:

> david@lang.hm writes:
>
>> no, saying that the result must be acceptable to other software (in
>> this case the software running in the BIOS) is not part of the source
>> code.
>
> Why not?  The digital signature is a statement (which translates
> roughly to "Tivo approves this") to be used in a computer in order to
> bring about a certain result.  That result is making it boot on the
> PVR.  Source code simply means the original forms or inputs used to
> generate machine-readable statements.

but the signature isn't part of the kernel, and the code that checks the 
signature is completely independant. and finally the PVR functions are not 
part of the kernel (and not under the GPL in any case)

if your argument was true then Oracle releasing a database appliance would 
require Oracle to give you the source to their database since it's part of 
'bringing about a certain result' namely operating as a database server.

if your argument was true then releasing a GPL package for windows would 
require that the windows kernel source be released, after it it's 
nessasary for 'brining about a certain result' namely letting your code 
run.

these are both nonsense results.

David Lang

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Michael Poole]

david@lang.hm writes:

> On Wed, 20 Jun 2007, Michael Poole wrote:
>
>> david@lang.hm writes:
>>
>>> no, saying that the result must be acceptable to other software (in
>>> this case the software running in the BIOS) is not part of the source
>>> code.
>>
>> Why not?  The digital signature is a statement (which translates
>> roughly to "Tivo approves this") to be used in a computer in order to
>> bring about a certain result.  That result is making it boot on the
>> PVR.  Source code simply means the original forms or inputs used to
>> generate machine-readable statements.
>
> but the signature isn't part of the kernel, and the code that checks
> the signature is completely independant. and finally the PVR functions
> are not part of the kernel (and not under the GPL in any case)
>
> if your argument was true then Oracle releasing a database appliance
> would require Oracle to give you the source to their database since
> it's part of 'bringing about a certain result' namely operating as a
> database server.

>From the kernel's COPYING file:

   NOTE! This copyright does *not* cover user programs that use kernel
 services by normal system calls - this is merely considered normal use
 of the kernel, and does *not* fall under the heading of "derived work".

See also the portion below.

> if your argument was true then releasing a GPL package for windows
> would require that the windows kernel source be released, after it
> it's nessasary for 'brining about a certain result' namely letting
> your code run.

>From section 3 of the GPL:

  However, as a special exception, the source code distributed need
  not include anything that is normally distributed (in either source
  or binary form) with the major components (compiler, kernel, and so
  on) of the operating system on which the executable runs, unless
  that component itself accompanies the executable.

> these are both nonsense results.

.. which is why they are recognized to be different.

Michael Poole

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[david]

On Wed, 20 Jun 2007, Michael Poole wrote:

> david@lang.hm writes:
>
>> On Wed, 20 Jun 2007, Michael Poole wrote:
>>
>>> david@lang.hm writes:
>>>
>>>> no, saying that the result must be acceptable to other software (in
>>>> this case the software running in the BIOS) is not part of the source
>>>> code.
>>>
>>> Why not?  The digital signature is a statement (which translates
>>> roughly to "Tivo approves this") to be used in a computer in order to
>>> bring about a certain result.  That result is making it boot on the
>>> PVR.  Source code simply means the original forms or inputs used to
>>> generate machine-readable statements.
>>
>> but the signature isn't part of the kernel, and the code that checks
>> the signature is completely independant. and finally the PVR functions
>> are not part of the kernel (and not under the GPL in any case)
>>
>> if your argument was true then Oracle releasing a database appliance
>> would require Oracle to give you the source to their database since
>> it's part of 'bringing about a certain result' namely operating as a
>> database server.
>
> From the kernel's COPYING file:
>
>   NOTE! This copyright does *not* cover user programs that use kernel
> services by normal system calls - this is merely considered normal use
> of the kernel, and does *not* fall under the heading of "derived work".
>
> See also the portion below.

the PVR software does exactly this, so the functionality as a PVR has 
nothing to do with the kernel by your own arguments.

>> if your argument was true then releasing a GPL package for windows
>> would require that the windows kernel source be released, after it
>> it's nessasary for 'brining about a certain result' namely letting
>> your code run.
>
> From section 3 of the GPL:
>
>  However, as a special exception, the source code distributed need
>  not include anything that is normally distributed (in either source
>  or binary form) with the major components (compiler, kernel, and so
>  on) of the operating system on which the executable runs, unless
>  that component itself accompanies the executable.

but then how can the kernel impose any restriction on the BIOS? (and 
remember, you are the one arguing that it somehow does)

David Lang

>> these are both nonsense results.
>
> .. which is why they are recognized to be different.
>
> Michael Poole
>

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Dave Neuer]

On 6/20/07, david@lang.hm <david@lang.hm> wrote:
>
> but the signature isn't part of the kernel

But some would argue that it's part of the source with which the
binary is derived (only a court could meaningfully decide if they're
right).

> and the code that checks the
> signature is completely independant.

Irrelevant.

> and finally the PVR functions are not
> part of the kernel

Booting is.

Dave

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 20, 2007, david@lang.hm wrote:

> but the signature isn't part of the kernel, and the code that checks
> the signature is completely independant.

Well, then remove or otherwise mangle the signature in the disk of
your TiVo DVR and see at what point the boot-up process halts.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[david]

On Thu, 21 Jun 2007, Alexandre Oliva wrote:

> On Jun 20, 2007, david@lang.hm wrote:
>
>> but the signature isn't part of the kernel, and the code that checks
>> the signature is completely independant.
>
> Well, then remove or otherwise mangle the signature in the disk of
> your TiVo DVR and see at what point the boot-up process halts.

I have actually done exactly that. what happens is that the bootloader 
detects a problem and switches to the other active partition ane reboots. 
ir neither of the boot partitions meet the requirements the cycle will 
continue forever.

David Lang

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 21, 2007, david@lang.hm wrote:

> On Thu, 21 Jun 2007, Alexandre Oliva wrote:
>> On Jun 20, 2007, david@lang.hm wrote:
>> 
>>> but the signature isn't part of the kernel, and the code that checks
>>> the signature is completely independant.
>> 
>> Well, then remove or otherwise mangle the signature in the disk of
>> your TiVo DVR and see at what point the boot-up process halts.

> I have actually done exactly that. what happens is that the bootloader
> detects a problem and switches to the other active partition ane
> reboots. ir neither of the boot partitions meet the requirements the
> cycle will continue forever.

Oh, too bad.  That must be a bug in the boot loader, right? :-)


BTW, since you got a TiVo...  I'm writing an article on Tivoization,
could you (or anyone else) please help me get information as to what
other GPLed programs or libraries TiVo includes in their devices?

Thanks in advance,

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[david]

On Thu, 21 Jun 2007, Alexandre Oliva wrote:

> On Jun 21, 2007, david@lang.hm wrote:
>
>> On Thu, 21 Jun 2007, Alexandre Oliva wrote:
>>> On Jun 20, 2007, david@lang.hm wrote:
>>>
>>>> but the signature isn't part of the kernel, and the code that checks
>>>> the signature is completely independant.
>>>
>>> Well, then remove or otherwise mangle the signature in the disk of
>>> your TiVo DVR and see at what point the boot-up process halts.
>
>> I have actually done exactly that. what happens is that the bootloader
>> detects a problem and switches to the other active partition ane
>> reboots. ir neither of the boot partitions meet the requirements the
>> cycle will continue forever.
>
> Oh, too bad.  That must be a bug in the boot loader, right? :-)
>
>
> BTW, since you got a TiVo...  I'm writing an article on Tivoization,
> could you (or anyone else) please help me get information as to what
> other GPLed programs or libraries TiVo includes in their devices?
>
> Thanks in advance,

frankly, I haven't checked the licenses on the software. I'd suggest going 
to www.tivo.com/linux and download all the source for all the different 
versions there.

by the way, it looks like there is one wireless driver that they ship in 
some releases but don't provide the source for. but if you plan on going 
after them for that you better go after everyone who ships binary kernel 
modules.

David Lang

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 21, 2007, david@lang.hm wrote:

> frankly, I haven't checked the licenses on the software. I'd suggest
> going to www.tivo.com/linux and download all the source for all the
> different versions there.

Yeah, thanks, I remembered someone had posted that URL the second
after a hit Send :-(

I've already got cmd.tar.gz and I'm looking at it now.

Thanks again,

> by the way, it looks like there is one wireless driver that they ship
> in some releases but don't provide the source for. but if you plan on
> going after them for that you better go after everyone who ships
> binary kernel modules.

Only copyright holders of Linux can go after them on matters of kernel
drivers.  Or is this driver derived from any software copyrighted by
myself?  Or did you mean the FSF, with whom I'm not associated in any
way other than ideologically?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[david]

On Thu, 21 Jun 2007, Alexandre Oliva wrote:

> On Jun 21, 2007, david@lang.hm wrote:
>
>> frankly, I haven't checked the licenses on the software. I'd suggest
>> going to www.tivo.com/linux and download all the source for all the
>> different versions there.
>
> Yeah, thanks, I remembered someone had posted that URL the second
> after a hit Send :-(
>
> I've already got cmd.tar.gz and I'm looking at it now.
>
> Thanks again,
>
>> by the way, it looks like there is one wireless driver that they ship
>> in some releases but don't provide the source for. but if you plan on
>> going after them for that you better go after everyone who ships
>> binary kernel modules.
>
> Only copyright holders of Linux can go after them on matters of kernel
> drivers.  Or is this driver derived from any software copyrighted by
> myself?  Or did you mean the FSF, with whom I'm not associated in any
> way other than ideologically?

from the page listed above

NOTE: 4.0.1a and later releases contain Atmel WLAN drivers in addition to 
the other network adapter drivers that are posted here. These WLAN drivers 
were received by TiVo directly from Atmel and are under the terms of a 
formal non-disclosure agreement, so we cannot publish them under the terms 
of the GPL version 2. Atmel has since released a different version of the 
drivers under the GPL version 2; however, that release does not change our 
prior obligation with Atmel. At this time, we have no plans to use Atmel's 
publicly released drivers in our development.

when I talked about 'going after tivo' I wasn't just refering to doing so 
with lawsuits. the way some people in this thread have acted I could see 
people trying to use this as 'the smoking gun' that proves that the evil 
tivo people didn't release everything they were supposed to.

modules are a grey area, some are clearly derived from the kernel and some 
are just as clearly not derived from the kernel, most are not as clear. I 
don't know the situation for this particular module, but that has nothing 
to do with this topic.

David Lang

RE: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Schwartz]

> The kernel you build from the source code that Tivo distributes must
> be accepted by Tivo's hardware without making other modifications (to
> Tivo's hardware or bootloader).  If that is possible, I will retract
> what I said.  If it is not possible, they are omitting part of the
> program's source code:
>
>     A "computer program" is a set of statements or instructions to be
>     used directly or indirectly in a computer in order to bring about
>     a certain result.
>     -- US Code, Title 17, Section 101

A key is a number. A signature is a number. They are neither statements nor
instructions. The argument that GPLv2 prohibits Tivoization is really and
truly absurd. It has neither a legal nor a moral leg to stand on.

DS

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 20, 2007, "David Schwartz" <davids@webmaster.com> wrote:

> A key is a number. A signature is a number.

And a program is a number.

http://asdf.org/~fatphil/maths/illegal.html

Your point?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alan Cox]

On Wed, 20 Jun 2007 12:55:10 -0700
"David Schwartz" <davids@webmaster.com> wrote:

> 
> > The kernel you build from the source code that Tivo distributes must
> > be accepted by Tivo's hardware without making other modifications (to
> > Tivo's hardware or bootloader).  If that is possible, I will retract
> > what I said.  If it is not possible, they are omitting part of the
> > program's source code:
> >
> >     A "computer program" is a set of statements or instructions to be
> >     used directly or indirectly in a computer in order to bring about
> >     a certain result.
> >     -- US Code, Title 17, Section 101
> 
> A key is a number. A signature is a number. They are neither statements nor
> instructions. The argument that GPLv2 prohibits Tivoization is really and
> truly absurd. It has neither a legal nor a moral leg to stand on.

A computer program is a number too.

Alan

RE: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Schwartz]

> On Wed, 20 Jun 2007 12:55:10 -0700

> "David Schwartz" <davids@webmaster.com> wrote:

> > A key is a number. A signature is a number. They are neither
> > statements nor
> > instructions. The argument that GPLv2 prohibits Tivoization is
> > really and
> > truly absurd. It has neither a legal nor a moral leg to stand on.

> A computer program is a number too.

No, it's not. It can be expressed as a number, but it is not a number.

Keys are purely numbers, they are nothing else. Signatures are pure
primitive facts encoded as numbers (authority X blessed object Y).

A computer program is a set of instructions to accomplish a particular
result. It can be expressed as a number, but that doesn't mean it is a
number.

It might be true in principle to develop a scheme whereby every physical
object uniquely corresponds to an extremely large number. That doesn't turn
physical objects into numbers.

DS

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Stephen Clark]

David Schwartz wrote:

>>On Wed, 20 Jun 2007 12:55:10 -0700
>>    
>>
>
>  
>
>>"David Schwartz" <davids@webmaster.com> wrote:
>>    
>>
>
>  
>
>>>A key is a number. A signature is a number. They are neither
>>>statements nor
>>>instructions. The argument that GPLv2 prohibits Tivoization is
>>>really and
>>>truly absurd. It has neither a legal nor a moral leg to stand on.
>>>      
>>>
>
>  
>
>>A computer program is a number too.
>>    
>>
>
>No, it's not. It can be expressed as a number, but it is not a number.
>  
>
??? can be expressed as a number, but it is not a number ???
sure its a number.

>Keys are purely numbers, they are nothing else. Signatures are pure
>primitive facts encoded as numbers (authority X blessed object Y).
>
>A computer program is a set of instructions to accomplish a particular
>result. It can be expressed as a number, but that doesn't mean it is a
>number.
>
>It might be true in principle to develop a scheme whereby every physical
>object uniquely corresponds to an extremely large number. That doesn't turn
>physical objects into numbers.
>
>DS
>
>
>-
>To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
>the body of a message to majordomo@vger.kernel.org
>More majordomo info at  http://vger.kernel.org/majordomo-info.html
>Please read the FAQ at  http://www.tux.org/lkml/
>
>  
>


-- 

"They that give up essential liberty to obtain temporary safety, 
deserve neither liberty nor safety."  (Ben Franklin)

"The course of history shows that as a government grows, liberty 
decreases."  (Thomas Jefferson)

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Kyle Moffett]

On Jun 21, 2007, at 15:19:35, Stephen Clark wrote:
> David Schwartz wrote:
>>> On Wed, 20 Jun 2007 12:55:10 -0700 "David Schwartz"  
>>> <davids@webmaster.com> wrote:
>>>> A key is a number. A signature is a number. They are neither  
>>>> statements nor instructions. The argument that GPLv2 prohibits  
>>>> Tivoization is really and truly absurd. It has neither a legal  
>>>> nor a moral leg to stand on.
>>>
>>> A computer program is a number too.
>>
>> No, it's not. It can be expressed as a number, but it is not a  
>> number.
>>
> ??? can be expressed as a number, but it is not a number ??? sure  
> its a number.
>
>> Keys are purely numbers, they are nothing else. Signatures are  
>> pure primitive facts encoded as numbers (authority X blessed  
>> object Y).
>>
>> A computer program is a set of instructions to accomplish a  
>> particular result. It can be expressed as a number, but that  
>> doesn't mean it is a number.
>>
>> It might be true in principle to develop a scheme whereby every  
>> physical object uniquely corresponds to an extremely large number.  
>> That doesn't turn physical objects into numbers.

Both of you lose this argument.  All irrational numbers, for example,  
"break" every copyright that could possibly exist.  For example, you  
can find any arbitrary sequence of Base-N digits when you express PI  
in base-N form.  I can simultaneously express both the laws of  
physics (not copyrightable) and the latest episode of the TV show  
"Numbers" (thoroughly copyrighted) as numbers.  In fact, we do both  
all the time (you can express both the latest equations for  
theoretical physics and a TV show as bits (IE: numbers) on an HDD.   
Ergo "$FOO is a number" says *NOTHING* about whether or not copyright  
applies to $FOO.  In case you haven't noticed, the whole damn point  
of math is that you can express *EVERYTHING* as numbers, albeit maybe  
horribly unbelievably complex ones.

Now, back to actual legal issues: Since most copyright laws  
explicitly prevent copyrighting of pure math, the only actual  
protection you have for some collection of so-called numbers is  
whether or not the numbers *REPRESENT* something which may be  
copyrighted.  Furthermore, copyright has _always_ been independent of  
representation; a person owns copyright on a book regardless of  
whether it's hardback, softcover, digital, memorized, etc.  The  
person who owns the copyright on a book is able to prevent someone  
who has memorized the book from giving public recitals of said book,  
and the neuron-linkage-based storage the brain uses is about as far  
as you can possibly get from twiddling magnetic bits on a disk drive  
or dumping carbon-based inks on a page made of plant cellulose.

Cheers,
Kyle Moffett

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 21, 2007, "David Schwartz" <davids@webmaster.com> wrote:

>> On Wed, 20 Jun 2007 12:55:10 -0700
>> "David Schwartz" <davids@webmaster.com> wrote:

>> > A key is a number. A signature is a number. They are neither
>> > statements nor
>> > instructions. The argument that GPLv2 prohibits Tivoization is
>> > really and
>> > truly absurd. It has neither a legal nor a moral leg to stand on.

>> A computer program is a number too.

> No, it's not. It can be expressed as a number, but it is not a number.

By this logic, then a key is a key, and a signature is a signature.
They can be expressed as numbers, sure.

> A computer program is a set of instructions to accomplish a particular
> result. It can be expressed as a number, but that doesn't mean it is a
> number.

A key is an input to a cryptographical algorithm, and a signature is
an output.  I could try to come up with more creative definitions, but
you get the idea already.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Tomas Neme]

>     A "computer program" is a set of statements or instructions to be
>     used directly or indirectly in a computer in order to bring about
>     a certain result.
>     -- US Code, Title 17, Section 101

so?

Not GPL related, but casino machine software that needs to be approved
by the casino regulation office in Argentina need to provide source,
compiling instructions AND binaries, and the binaries must pass a diff
check. This is impossible without a hacked compiler since the
timestamps WILL differ.

Just an example that legality doesn't always comply with itself, and
even less make sense.

T

-- 
|_|0|_|
|_|_|0|
|0|0|0|

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Tomas Neme]

> Just an example that legality doesn't always comply with itself, and
> even less make sense.

plus, and I repeat myself.. the program comes with no warranties whatsoever.

and if your complains are purely moral, see it this way: if TiVo
didn't sign their kernel, digital cable providers wouldn't give them
their hash keys, and they wouldn't be able to show HD signals,
rendering them useless, and making them go bankrupt... so they'd go
BSD, because they ARE a company after all, and they are after The
Moneys..

no?

-- 
|_|0|_|
|_|_|0|
|0|0|0|

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Michael Poole]

Tomas Neme writes:

>>     A "computer program" is a set of statements or instructions to be
>>     used directly or indirectly in a computer in order to bring about
>>     a certain result.
>>     -- US Code, Title 17, Section 101
>
> so?

People keep arguing that the signature is somehow not part of the
kernel or not subject to copyright law.  I suspect they do not realize
how broad the legal definition of "computer program" is.

> Not GPL related, but casino machine software that needs to be approved
> by the casino regulation office in Argentina need to provide source,
> compiling instructions AND binaries, and the binaries must pass a diff
> check. This is impossible without a hacked compiler since the
> timestamps WILL differ.
>
> Just an example that legality doesn't always comply with itself, and
> even less make sense.

This discussion is about copyright and the GPL, not legal quirks.  I
am aware of a large number of silly results reached by law.

> plus, and I repeat myself.. the program comes with no warranties whatsoever.
>
> and if your complains are purely moral, see it this way: if TiVo
> didn't sign their kernel, digital cable providers wouldn't give them
> their hash keys, and they wouldn't be able to show HD signals,
> rendering them useless, and making them go bankrupt... so they'd go
> BSD, because they ARE a company after all, and they are after The
> Moneys..
>
> no?

The GPL does not guarantee anyone a viable business model.  Following
it is not conditional on profitability.  It is only conditional on
exercising rights that are granted by the GPL.

Michael Poole

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Tim Post]

On Wed, 2007-06-20 at 16:25 -0400, Michael Poole wrote:
> Tomas Neme writes:
> 
> >>     A "computer program" is a set of statements or instructions to be
> >>     used directly or indirectly in a computer in order to bring about
> >>     a certain result.
> >>     -- US Code, Title 17, Section 101
> >
> > so?
> 
> People keep arguing that the signature is somehow not part of the
> kernel or not subject to copyright law.  I suspect they do not realize
> how broad the legal definition of "computer program" is.


"A computer program is a set of statements or instructions used by a
computer."

I have no idea what logic was employed that made continuing beyond that
point desirable. 

However, the GPL was designed so you couldn't be forced to pay to use
the software released under it, my guess is if it weren't for such a
broad definition the GPL would look a bit different than it does. 

When laws give you ground that gets you free stuff, typically its unwise
to suggest that they take some back ;)

Best,
--Tim

RE: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Schwartz]

> Tomas Neme writes:

> > I have been following this discussion for the last week or so, and
> > what I haven't been able to figure out is what the hell is the big
> > deal with TiVO doing whatever they want to with their stupid design.
> > They made a design, they build a machine, they sell it as is, and
> > provide source code for GPL'ed software... what's your problem?

> It's simple: they don't provide _complete_ source code.  They keep the
> source code for the part of their Linux kernel images that provides
> the functionality "runs on Tivo DVRs".  The GPL requires that
> distributors of binary versions provide complete source code, not just
> the parts of source code that are convenient.
>
> Michael Poole

That leads to lots of obvious nonsense unless you fix it with all kinds of
made up ad-hoc changes just to get the result you want. Why doesn't Linus
have to release the keys he uses to sign the Linux kernel source
distributions? That provides the functionality "can be proven to be
authorized by Linus". What you call "runs on Tivo DVRs", I call "can be
proven to be authorized by Tivo to run on Tivo DVRs".

The problem is that your description of the functionality as "runs on Tivo
DVRs" is an ad-hoc choice. You could describe that functionality any number
of other ways, and this is the only one that supports your argument. (Which
would be wrong anyway since functionality has nothing to do with whether
something is part of the source code or not. Copyright is not functional in
operation.)

Tivo's choice is an authorization decision. It is similar to you not having
root access to a Linux box. Sorry, you can't run a modified kernel on that
machine, but you can still modify the kernel and run it on any hardware
where authorization decisions don't stop you from doing so. The GPL was
never about such authorization decisions.

Suppose I make a machine that automatically accepts any kernel signed by
Linus and configures it, compiles it, and installs it. Does this change
Linus' signature into "works with my machine's kernel autoinstall"
functionality? The signature is proof Linus 'blessed' the release. Others
who trust Linus can use this functionally to make authorization decisions.
However, your releases are not blessed by Linus and there's no reason you
should be able to give them this 'blessed by Linus' functionality.

There are other legal reasons why this can't be right (most of them have
been stated at least three times in this thread), but I think the
commonsense argument is the most persuasive. Far from being part of the
source code, the signature is proof of source and authorization. These are
fact that don't apply to your modified release.

If I only want to run kernels signed by Linus, you have no right to trick me
into running a modified kernel. The Tivo only wants to run kernels signed by
Tivo.

Your argument is equivalent to saying that I have the right not just to run
modified Linux kernels on my own hardware but to compel others to run my
modifications even when authorization decisions say they won't run my
changes. (By bogusly calling authorization decisions 'functionality'.)

DS

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Dave Neuer]

On 6/20/07, David Schwartz <davids@webmaster.com> wrote:
>
> > Tomas Neme writes:
>
> > > I have been following this discussion for the last week or so, and
> > > what I haven't been able to figure out is what the hell is the big
> > > deal with TiVO doing whatever they want to with their stupid design.
> > > They made a design, they build a machine, they sell it as is, and
> > > provide source code for GPL'ed software... what's your problem?
>
> > It's simple: they don't provide _complete_ source code.  They keep the
> > source code for the part of their Linux kernel images that provides
> > the functionality "runs on Tivo DVRs".  The GPL requires that
> > distributors of binary versions provide complete source code, not just
> > the parts of source code that are convenient.
> >
> > Michael Poole
>
> That leads to lots of obvious nonsense unless you fix it with all kinds of
> made up ad-hoc changes just to get the result you want. Why doesn't Linus
> have to release the keys he uses to sign the Linux kernel source
> distributions? That provides the functionality "can be proven to be
> authorized by Linus". What you call "runs on Tivo DVRs", I call "can be
> proven to be authorized by Tivo to run on Tivo DVRs".

This argument is the obvious nonsense. "Runs on TiVO" is a property of
the software that TiVO distributes -- such an important property that
it would be nonsensical for them to distribute it with their hardware.
But they do distribute it, and only the GPL allows them to.

Linus' key is not required to use the software Linus distributes under
the GPL, by contrast.

>
> Tivo's choice is an authorization decision. It is similar to you not having
> root access to a Linux box. Sorry, you can't run a modified kernel on that
> machine, but you can still modify the kernel and run it on any hardware
> where authorization decisions don't stop you from doing so. The GPL was
> never about such authorization decisions.

Says judge Schwartz. Oops. That's right, you're not a judge in any
legal jurisdiction, nor an author of the GPL.

Dave

RE: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Schwartz]

> This argument is the obvious nonsense. "Runs on TiVO" is a property of
> the software that TiVO distributes -- such an important property that
> it would be nonsensical for them to distribute it with their hardware.
> But they do distribute it, and only the GPL allows them to.

Why does the importance of the property matter to the validity of the
argument?

> Linus' key is not required to use the software Linus distributes under
> the GPL, by contrast.

Why does whether or not the key is required to use the software matter? It
may be impossible to use a Linux kernel on a particular piece of hardware
without the BIOS, that doesn't mean the BIOS source code is part of the
kernel source code even if the kernel is shipped for that hardware.

> > Tivo's choice is an authorization decision. It is similar to
> > you not having
> > root access to a Linux box. Sorry, you can't run a modified
> > kernel on that
> > machine, but you can still modify the kernel and run it on any hardware
> > where authorization decisions don't stop you from doing so. The GPL was
> > never about such authorization decisions.

> Says judge Schwartz. Oops. That's right, you're not a judge in any
> legal jurisdiction, nor an author of the GPL.

Nice argument. I'm wrong because people can disagree with me.

DS

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Dave Neuer]

On 6/20/07, David Schwartz <davids@webmaster.com> wrote:
>
> > This argument is the obvious nonsense. "Runs on TiVO" is a property of
> > the software that TiVO distributes -- such an important property that
> > it would be nonsensical for them to distribute it with their hardware.
> > But they do distribute it, and only the GPL allows them to.
>
> Why does the importance of the property matter to the validity of the
> argument?

>From a legal standpoint, perhaps you're right, it doesn't matter what
the function is. From a moral standpoint it should be obvious to you
that "runs on TiVO" is TiVO's sole motivation to distribute the
software at all, it is "the software" and arguing that they have an
equivalent obligation WRT it as to some incidental thing like Linus'
signing key is just preposterous.

> > > Tivo's choice is an authorization decision. It is similar to
> > > you not having
> > > root access to a Linux box. Sorry, you can't run a modified
> > > kernel on that
> > > machine, but you can still modify the kernel and run it on any hardware
> > > where authorization decisions don't stop you from doing so. The GPL was
> > > never about such authorization decisions.
>
> > Says judge Schwartz. Oops. That's right, you're not a judge in any
> > legal jurisdiction, nor an author of the GPL.
>
> Nice argument. I'm wrong because people can disagree with me.

No, in this case you are wrong because absent authority to decide the
meaning from a dispositive legal standpoint (the law says the license
means this) or knowledge of the intent of the author of the GPL (I the
author intended it to mean this), your statement that the GPL was
"never about" "such decisions" is meaningless, AFAICT.

>
> DS

Dave

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 20, 2007, "Jesper Juhl" <jesper.juhl@gmail.com> wrote:

> On 19/06/07, Alexandre Oliva <aoliva@redhat.com> wrote:
>> On Jun 18, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:
>> > In the GPLv3 world, we have already discussed in this thread how you can
>> > follow the GPLv3 by making the TECHNICALLY INFERIOR choice of using a ROM
>> > instead of using a flash device.
>> 
>> Yes.  This is one option that doesn't bring any benefits to anyone.
>> It maintains the status quo for users and the community, but it loses
>> the ability for the vendor to upgrade, fix or otherwise control the
>> users.  Bad for the vendor.

> Also bad for the user

We already know the vendor doesn't care about the user, so why should
we take this into account when analyzing the reasoning of the vendor?

The rest of your paragraph was covered in what I wrote, BTW.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[david]

On Wed, 20 Jun 2007, Alexandre Oliva wrote:

> On Jun 20, 2007, "Jesper Juhl" <jesper.juhl@gmail.com> wrote:
>
>> On 19/06/07, Alexandre Oliva <aoliva@redhat.com> wrote:
>>> On Jun 18, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:
>>>> In the GPLv3 world, we have already discussed in this thread how you can
>>>> follow the GPLv3 by making the TECHNICALLY INFERIOR choice of using a ROM
>>>> instead of using a flash device.
>>>
>>> Yes.  This is one option that doesn't bring any benefits to anyone.
>>> It maintains the status quo for users and the community, but it loses
>>> the ability for the vendor to upgrade, fix or otherwise control the
>>> users.  Bad for the vendor.
>
>> Also bad for the user
>
> We already know the vendor doesn't care about the user, so why should
> we take this into account when analyzing the reasoning of the vendor?

no, we don't know this. you attribute the reason for the lockdown to be 
anti-user. others view it as being pro-user becouse it lets the user get 
functionality that they wouldn't have access to otherwise.

David Lang

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 20, 2007, david@lang.hm wrote:

> On Wed, 20 Jun 2007, Alexandre Oliva wrote:
>> We already know the vendor doesn't care about the user, so why should
>> we take this into account when analyzing the reasoning of the vendor?

> no, we don't know this. you attribute the reason for the lockdown to
> be anti-user. others view it as being pro-user becouse it lets the
> user get functionality that they wouldn't have access to otherwise.

Yeah, yeah, now please put your hands to the back such that I can
place your handcuffs, such that you can't use your hands to kill
someone.  Well, you might still be able to use your hands for this
purpose if you're really clever, or you can kill people by other
means.  But at least I'll be able to claim that I did my part.

:-/

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Andrew McKay]

> No, I'm not.  You can say tivoization is *good* however much you like.
> This doesn't dispute in any way my claim that no tivoization would be
> *better*, that you'd get contributions from the people that, because
> of tivoization, don't feel compelled to develop and contribute,
> because they can't use the fruits of their efforts in the device where
> they would be most useful for them.
> 

I know I'm jumping in to this conversation in the middle and a few days behind, 
but that's one of the biggest fallacies in the open source communities.  First 
of all more system developers contribute to the community than end users do. 
Secondly GPLv3 will cause companies like TIVO, router companies, security 
companies to not adopt Linux as an operating system, because they can't secure 
their system.  Placing code in a ROM so they can't upgrade their own systems is 
an absolute joke, and not a viable option.  In the end, GPLV3 will cause less 
contribution to the Linux Kernel community. Systems developers won't be choosing 
Linux and directly contributing to Kernel development while developing their 
product.  Fewer products on the market will use linux, so even end users, who 
really don't contribute a lot, won't be contributing either.

I'm not going to address whether GPLv3 changed the spirit of GPLv2, but saying 
that licensing the Linux Kernel under GPLv3 will result in more contributions is 
absolute BS.

Andrew McKay
Iders Inc.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alan Cox]

> Secondly GPLv3 will cause companies like TIVO, router companies, security 
> companies to not adopt Linux as an operating system, because they can't secure 
> their system.  Placing code in a ROM so they can't upgrade their own systems is 

You've made an important mistake. You said "their system". Now its "our
code" and "whoever bought the units' hardware" so it isn't their anything.

You've made a second mistake I think by assuming that vendor held keys
"improve" security and must be vendor held and secret for it to work. In
fact vendor owned key systems that cannot be changed usually reduce
security.

There are very very good reasons for having vendor owned secret keys.
There are also very very good reasons for being able to rekey or disable
the key on your box.

Ask people whose product vendor went bankrupt. With the ability to
override/replace the keys they could have maintained their system
securely instead they could make no updates and the boxes were left
insecure.

RPM for example intentionally follows this approach. RPM will check keys
and the keys for different vendors/projects will not be released. However
you can add keys, remove keys or even tell rpm "forget the key checking".

A silly example that rather makes the point is the X-Box. Using a bug in
a game you can get into the X-box system and run Linux instead. As the
owner of an X-box you can't fix the bug in the game because you don't
know the signing key, but if you could change or add keys you could stop
the "problem" occurring.

The Xbox is perhaps an oddity in that the insecurity is widely preferred
by the owners but the situation applies in cases where the owner would
prefer the reverse were true.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Andrew McKay]

Alan Cox wrote:
>> Secondly GPLv3 will cause companies like TIVO, router companies, security 
>> companies to not adopt Linux as an operating system, because they can't secure 
>> their system.  Placing code in a ROM so they can't upgrade their own systems is 
> 
> You've made an important mistake. You said "their system". Now its "our
> code" and "whoever bought the units' hardware" so it isn't their anything.


Yes, the hardware belongs to the user, and the software belongs to the Linux 
community.   However I think I wasn't 100% clear, I also mean keeping companies 
networks and content secured.  Credit card companies insuring the software 
hasn't been modified to skim cards (not that it's the only way to skim a card), 
  or Tivo making sure that their content providers are protected.  Lets look at 
the credit card example.  Sure the user could modify the system and boot their 
own kernel, but it doesn't have to play nice with Mastercard's network anymore. 
  Or better yet, would actually report that a certain business's card reader had 
been tampered with.

> 
> You've made a second mistake I think by assuming that vendor held keys
> "improve" security and must be vendor held and secret for it to work. In
> fact vendor owned key systems that cannot be changed usually reduce
> security.
> 
> There are very very good reasons for having vendor owned secret keys.
> There are also very very good reasons for being able to rekey or disable
> the key on your box.
> 
> Ask people whose product vendor went bankrupt. With the ability to
> override/replace the keys they could have maintained their system
> securely instead they could make no updates and the boxes were left
> insecure.

I do see what you're saying here, and I can see how this is a problem.  However 
what is the solution?  Sure having the system open for users to replace software 
and tinker is great.  It's how I got into engineering.  I can also appreciate 
the ability for the end user to fix and continue to use a system long after a 
vendor goes out of business.  However, I don't see how this would ever require a 
company like Tivo or Mastercard to have their networks play nice with a unit 
that has been modified by the end user, potentially opening up some serious 
security holes.  From what I understand this would still violate GPLv3 because 
the system could no longer preform the task it was designed to do with modified 
code, but maybe I have misunderstood.

Andrew McKay

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 20, 2007, Andrew McKay <amckay@iders.ca> wrote:

> However, I don't see how this would ever require a company like Tivo
> or Mastercard to have their networks play nice with a unit that has
> been modified by the end user, potentially opening up some serious
> security holes.

Which is why the GPLv3 doesn't make the requirement that you stated.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Tomas Neme]

> > However, I don't see how this would ever require a company like Tivo
> > or Mastercard to have their networks play nice with a unit that has
> > been modified by the end user, potentially opening up some serious
> > security holes.
>
> Which is why the GPLv3 doesn't make the requirement that you stated.

Why, if you let user-compiled kernels to run in a TiVo, it might be
modified so the TiVo can be used to pirate-copy protected content,
which is a serious security hole. TiVo would need to read, approve of,
and sign any modified kernels the users intend to use on their
hardware. If GPLv3 allows for this, it'd be doing exactly that

Tomás

-- 
|_|0|_|
|_|_|0|
|0|0|0|

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Dave Neuer]

On 6/20/07, Tomas Neme <lacrymology@gmail.com> wrote:
>
> Why, if you let user-compiled kernels to run in a TiVo, it might be
> modified so the TiVo can be used to pirate-copy protected content,

1) It may be far more likely that in the majority of cases it will be
modified with the intent to allow functionality which has no bearing
on copyrighted entertainment copyright or which is permitted under the
Fair Use doctrine. Neither you, nor I, nor anyone else on the list
knows whether that is the case.

2) There are far easier ways to pirate copyrighted entertainment
content (like buy the discs, professional duplicating hardware, and
just dup them) which I would wager is what actual pirates do 99% of
the time

> which is a serious security hole.

If you are a content company it's a security hole, if you are a TiVO
owner, it's a feature.

Dave

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Dave Neuer]

On 6/20/07, Tomas Neme <lacrymology@gmail.com> wrote:
>
> I'm about this far to Linus'izing my wording and calling you stupid,
> hypocrite, or bullshitter

Knock yourself out, it will no doubt lend much moral and logic weight
to your rhetoric.

Dave

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Tomas Neme]

On 6/20/07, Dave Neuer <mr.fred.smoothie@pobox.com> wrote:
> On 6/20/07, Tomas Neme <lacrymology@gmail.com> wrote:
> >
> > I'm about this far to Linus'izing my wording and calling you stupid,
> > hypocrite, or bullshitter
>
> Knock yourself out, it will no doubt lend much moral and logic weight
> to your rhetoric.

I might not have the best rhetoric, but I still hold my point about
the credit card. Ask yourself: are you going to complain about Firefox
(GPL'ed) not passing information unencrypted because it stops
potential users (crackers ARE users) from doing what they want to with
it? What's a security issue and what's not is a matter of legality and
it's each part's duty to enforce legality in every way they can (I'm
not saying that I agree, I'm an anarchist, but it's just how it goes).

The content providers do it by not allowing DVRs to work if they're
not secure, and DVRs are secure by doing whatever is legally possible
to avoid crackers from bypassing security measures. On the other hand
is legal for you to bypass those security measures as long as you
don't make illegal use of those bypasses.

The kernel TiVo distributes works on TiVo boxes, The kernel modified
by you, is no longer the kernel TiVo distributes, and therefore the
key that the original kernel had no longer applies to it. Try running
your TiVo kernel on a PC, I think you won't be able to without a lot
of modification.. and then again, once you do the proper modifying,
you will be able to use it on your multimedia computer, and use all of
the wonderful things you DIE to be able to modify the TiVo kernel
for.. If you modify your TiVo kernel and say make it so it doesn't
have an IDE controller module anymore, you won't be able to run it on
your TiVo either.. at least not in any useful way, and would you be
complaining?

And someone said this already, if the signature is created via a known
algorithm, and only the key isn't provided, saying that that's not
GPLv2 compliant is like saying that I can't publish investigation work
that was produced sharing via a secured network unless I also publish
the SSH key I used through investigation, or the original value of the
srand() if the investigation relied on random number generation,
because the exact same results won't be reproducible.

T

-- 
|_|0|_|
|_|_|0|
|0|0|0|

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 20, 2007, "Tomas Neme" <lacrymology@gmail.com> wrote:

>> > However, I don't see how this would ever require a company like Tivo
>> > or Mastercard to have their networks play nice with a unit that has
>> > been modified by the end user, potentially opening up some serious
>> > security holes.
>> 
>> Which is why the GPLv3 doesn't make the requirement that you stated.

> Why, if you let user-compiled kernels to run in a TiVo, it might be
> modified so the TiVo can be used to pirate-copy protected content,

And then the user who uses such features in ways not permitted by the
copyright holders are committing a crime.  They can be prosecuted by
the copyright holders and convicted of the crime.

That TiVo can somehow become liable for this just shows how broken the
legal system in the US is.  It's like making a knife manufacturer
liable for a killing using a knife they made, just because the knife
didn't have technical measures intended to prevent the knife from
being used to kill people.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Lennart Sorensen]

On Thu, Jun 21, 2007 at 01:23:01AM -0300, Alexandre Oliva wrote:
> And then the user who uses such features in ways not permitted by the
> copyright holders are committing a crime.  They can be prosecuted by
> the copyright holders and convicted of the crime.

Well we already clearly know the content providers in the US don't trust
anyone else, and certainly don't think just using copyright laws to sue
people who violate copyright is enough.  No they want to put all sorts
of things in place to try to prevent it from even being possible to
violate copyright in the first place.  They don't believe in innocent
until proven guilty at all.

> That TiVo can somehow become liable for this just shows how broken the
> legal system in the US is.  It's like making a knife manufacturer
> liable for a killing using a knife they made, just because the knife
> didn't have technical measures intended to prevent the knife from
> being used to kill people.

So much for "Land of the free". :(

--
Len Sorensen

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alan Cox]

> So much for "Land of the free". :(

That was always just a typo. Its the Land of the Fee

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bernd Petrovitsch]

On Wed, 2007-06-20 at 18:14 -0300, Tomas Neme wrote:
[....]
> Why, if you let user-compiled kernels to run in a TiVo, it might be
> modified so the TiVo can be used to pirate-copy protected content,

Or it might be modified to fix a bug - either a technical one or a legal
one as described below.

> which is a serious security hole. TiVo would need to read, approve of,

"Pirate copying" is forbidden anyways in almost every jurisdiction
AFAIK. Perhaps we should disallow cars on the streets since one could
drive too fast with them.

And it is not a security hole for the owner of the hardware (I consider
secret keys somewhere else a much greater security threat) to the Linux
community or a lot of other entities. Probably just music industry
thinks like above.

And there are legislations were it is *legal* to make private copies
(for sure as long as you don't pass them on and somewhere even giving
away for nothing is legal). So I actually have a *right* (which also
can't be killed by contracts) to copy that movie for my private use. And
up to now it is actually legal in .at to do (more or less) everything to
get this right (and that may include hacking the device).

> and sign any modified kernels the users intend to use on their
> hardware. If GPLv3 allows for this, it'd be doing exactly that

	Bernd
-- 
Firmix Software GmbH                   http://www.firmix.at/
mobil: +43 664 4416156                 fax: +43 1 7890849-55
          Embedded Linux Development and Services

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Manu Abraham]

Bernd Petrovitsch wrote:
> On Wed, 2007-06-20 at 18:14 -0300, Tomas Neme wrote:
> [....]
>> Why, if you let user-compiled kernels to run in a TiVo, it might be
>> modified so the TiVo can be used to pirate-copy protected content,
> 
> Or it might be modified to fix a bug - either a technical one or a legal
> one as described below.
> 
>> which is a serious security hole. TiVo would need to read, approve of,
> 
> "Pirate copying" is forbidden anyways in almost every jurisdiction
> AFAIK. Perhaps we should disallow cars on the streets since one could
> drive too fast with them.
> 

Pirate copying should not be a reason to keep things closed as there are
better methods to keep things open, yet provide a _not_ free service.
But that would be upto the vendor how/what they wish to do rather than
we talking about it. Which would be of no use.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[david]

On Wed, 20 Jun 2007, Alexandre Oliva wrote:

> Subject: Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3
> 
> On Jun 20, 2007, Andrew McKay <amckay@iders.ca> wrote:
>
>> However, I don't see how this would ever require a company like Tivo
>> or Mastercard to have their networks play nice with a unit that has
>> been modified by the end user, potentially opening up some serious
>> security holes.
>
> Which is why the GPLv3 doesn't make the requirement that you stated.

so if the BIOS checked the checksum of the boot image and if it found it 
wasn't correct would disable the video input hardware but let you boot the 
system otherwise it would be acceptable to you and the GPLv3?

somehow I doubt it, but that's what it would take to prevent modified 
software from interacting with their networks (remembering that these 
networks are the cable and satellite networks in some cases)

it also seems that if this was the case it would be a trivial work-around 
for the GPLv3 if it was acceptable.

David Lang

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Andrew McKay]

david@lang.hm wrote:
> On Wed, 20 Jun 2007, Alexandre Oliva wrote:
> 
>> Subject: Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3
>>
>> On Jun 20, 2007, Andrew McKay <amckay@iders.ca> wrote:
>>
>>> However, I don't see how this would ever require a company like Tivo
>>> or Mastercard to have their networks play nice with a unit that has
>>> been modified by the end user, potentially opening up some serious
>>> security holes.
>>
>> Which is why the GPLv3 doesn't make the requirement that you stated.
> 
> so if the BIOS checked the checksum of the boot image and if it found it 
> wasn't correct would disable the video input hardware but let you boot 
> the system otherwise it would be acceptable to you and the GPLv3?
> 
> somehow I doubt it, but that's what it would take to prevent modified 
> software from interacting with their networks (remembering that these 
> networks are the cable and satellite networks in some cases)
> 
> it also seems that if this was the case it would be a trivial 
> work-around for the GPLv3 if it was acceptable.
> 

That is exactly where I was going with that.  A trivial work around in the Tivo 
case.  I guess GPLv4 will have to close up that hole?

Andrew McKay

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 20, 2007, david@lang.hm wrote:

> On Wed, 20 Jun 2007, Alexandre Oliva wrote:
>> Subject: Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3
>> 
>> On Jun 20, 2007, Andrew McKay <amckay@iders.ca> wrote:
>> 
>>> However, I don't see how this would ever require a company like Tivo
>>> or Mastercard to have their networks play nice with a unit that has
>>> been modified by the end user, potentially opening up some serious
>>> security holes.
>> 
>> Which is why the GPLv3 doesn't make the requirement that you stated.

> so if the BIOS checked the checksum of the boot image and if it found
> it wasn't correct would disable the video input hardware but let you
> boot the system otherwise it would be acceptable to you and the GPLv3?

I don't think so, but IANAL.  What do you think?  Here's what I
think to be the relevant passages.

  [...] The information must suffice to ensure that the continued
  functioning of the modified object code is in no case prevented or
  interfered with solely because modification has been made.

  [...]

  The requirement to provide Installation Information does not include
  a requirement to continue to provide support service, warranty, or
  updates for a work that has been modified or installed by the
  recipient, or for the User Product in which it has been modified or
  installed.  Network access may be denied when the modification
  itself materially and adversely affects the operation of the network
  or violates the rules and protocols for communication across the
  network.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[david]

On Thu, 21 Jun 2007, Alexandre Oliva wrote:

> On Jun 20, 2007, david@lang.hm wrote:
>
>> On Wed, 20 Jun 2007, Alexandre Oliva wrote:
>>> Subject: Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3
>>>
>>> On Jun 20, 2007, Andrew McKay <amckay@iders.ca> wrote:
>>>
>>>> However, I don't see how this would ever require a company like Tivo
>>>> or Mastercard to have their networks play nice with a unit that has
>>>> been modified by the end user, potentially opening up some serious
>>>> security holes.
>>>
>>> Which is why the GPLv3 doesn't make the requirement that you stated.
>
>> so if the BIOS checked the checksum of the boot image and if it found
>> it wasn't correct would disable the video input hardware but let you
>> boot the system otherwise it would be acceptable to you and the GPLv3?
>
> I don't think so, but IANAL.  What do you think?  Here's what I
> think to be the relevant passages.
>
>  [...] The information must suffice to ensure that the continued
>  functioning of the modified object code is in no case prevented or
>  interfered with solely because modification has been made.
>
>  [...]
>
>  The requirement to provide Installation Information does not include
>  a requirement to continue to provide support service, warranty, or
>  updates for a work that has been modified or installed by the
>  recipient, or for the User Product in which it has been modified or
>  installed.  Network access may be denied when the modification
>  itself materially and adversely affects the operation of the network
>  or violates the rules and protocols for communication across the
>  network.

Ok, so if refusing to run software that's tampered with isn't acceptable, 
and disabling the hardware that would be needed to talk on the network 
isn't acceptable. how exactly can they prevent a system that's been 
tampered with from accessing their network? (something even you say they 
have a right to do)

asking a device that's running software that you haven't verified to give 
you a checksum of itself isn't going to work becouse the software can just 
lie to you.

you claim they have this right, but then claim to prohibit every possible 
method of them excercising that right.

pick one side or the other, you don't get both.

David Lang

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 21, 2007, david@lang.hm wrote:

> how exactly can they prevent a system that's been tampered with from
> accessing their network?

By denying access to their servers?  By not granting whatever is
needed to initiate network sessions?

And note, "it's been tampered with" is not necessarily enough of a
reason to cut someone off, it has to meet these requirements:

>   when the modification itself materially and adversely affects the
>   operation of the network or violates the rules and protocols for
>   communication across the network.

> (something even you say they have a right to do)

as long as this right is not used by the software distributor to
impose restrictions on the user's ability to adapt the software to
their own needs.  The GPLv3 paragraph above makes a fair concession in
this regard, don't you agree?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[david]

On Thu, 21 Jun 2007, Alexandre Oliva wrote:

> On Jun 21, 2007, david@lang.hm wrote:
>
>> how exactly can they prevent a system that's been tampered with from
>> accessing their network?
>
> By denying access to their servers?  By not granting whatever is
> needed to initiate network sessions?
>
> And note, "it's been tampered with" is not necessarily enough of a
> reason to cut someone off, it has to meet these requirements:

how can the server tell if it's been tampered with?

>>   when the modification itself materially and adversely affects the
>>   operation of the network or violates the rules and protocols for
>>   communication across the network.
>
>> (something even you say they have a right to do)
>
> as long as this right is not used by the software distributor to
> impose restrictions on the user's ability to adapt the software to
> their own needs.  The GPLv3 paragraph above makes a fair concession in
> this regard, don't you agree?

no, one of the rules for the network is that the software must be 
certified, you are requireing the device to permit the software to be 
changed to an uncertified version.(to store credit card numbers and send 
them to a third party for example)

David Lang

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 21, 2007, david@lang.hm wrote:

> no, one of the rules for the network is that the software must be
> certified,

In this case you might have grounds to enforce this restriction of the
network on the network controller itself, I suppose.

Not that you should disable the network controller entirely (this
would render the computer useless for many other purpose unrelated
with blocking connections to your network).

You could instead arrange for the network controller to send some
signal that enables the network to recognize that the device is
running certified software.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[david]

On Thu, 21 Jun 2007, Alexandre Oliva wrote:

> On Jun 21, 2007, david@lang.hm wrote:
>
>> no, one of the rules for the network is that the software must be
>> certified,
>
> In this case you might have grounds to enforce this restriction of the
> network on the network controller itself, I suppose.

how would the network controller know if the software has been modified?

> Not that you should disable the network controller entirely (this
> would render the computer useless for many other purpose unrelated
> with blocking connections to your network).
>
> You could instead arrange for the network controller to send some
> signal that enables the network to recognize that the device is
> running certified software.

what sort of signal can the network controller send that couldn't be 
forged by the OS?

how would you do this where the device is a receiver on the netwoek (such 
as a satellite receiver)

David Lang

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Manu Abraham]

david@lang.hm wrote:

> what sort of signal can the network controller send that couldn't be
> forged by the OS?
> 
> how would you do this where the device is a receiver on the netwoek
> (such as a satellite receiver)

just for the question on the HOWTO (not on anything else)

You can easily have scrambled streams with regards to DVB, eg: using
EN50221. Some (like NDS for example in _some_ instances) use proprietary
schemes, but there are standards based methods also. eg: Common
Scrambling Algorithm (CSA)

But in any case, this can be broken down too .. :-)

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 21, 2007, david@lang.hm wrote:

> On Thu, 21 Jun 2007, Alexandre Oliva wrote:
>> On Jun 21, 2007, david@lang.hm wrote:
>> 
>>> no, one of the rules for the network is that the software must be
>>> certified,
>> 
>> In this case you might have grounds to enforce this restriction of the
>> network on the network controller itself, I suppose.

> how would the network controller know if the software has been modified?

The loader could check that and set a flag in the controller.

> what sort of signal can the network controller send that couldn't be
> forged by the OS?

Whatever the network controller designer created to enable it to do
so.

> how would you do this where the device is a receiver on the netwoek
> (such as a satellite receiver)

If it's input-only, then you can't possibly harm the operation of the
network by only listening in, can you?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[david]

On Thu, 21 Jun 2007, Alexandre Oliva wrote:

> On Jun 21, 2007, david@lang.hm wrote:
>
>> On Thu, 21 Jun 2007, Alexandre Oliva wrote:
>>> On Jun 21, 2007, david@lang.hm wrote:
>>>
>>>> no, one of the rules for the network is that the software must be
>>>> certified,
>>>
>>> In this case you might have grounds to enforce this restriction of the
>>> network on the network controller itself, I suppose.
>
>> how would the network controller know if the software has been modified?
>
> The loader could check that and set a flag in the controller.
>
>> what sort of signal can the network controller send that couldn't be
>> forged by the OS?
>
> Whatever the network controller designer created to enable it to do
> so.
>
>> how would you do this where the device is a receiver on the netwoek
>> (such as a satellite receiver)
>
> If it's input-only, then you can't possibly harm the operation of the
> network by only listening in, can you?

Ok, so you consider any anti-piracy measures to be something that GPLv3 
should prohibit.

thanks for finally takeing a position.

David Lang

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 21, 2007, david@lang.hm wrote:

> On Thu, 21 Jun 2007, Alexandre Oliva wrote:
>> If it's input-only, then you can't possibly harm the operation of the
>> network by only listening in, can you?

> Ok, so you consider any anti-piracy measures to be something that
> GPLv3 should prohibit.

In general, I object to the use of my code in ways that don't permit
users to run it for any purpose, study it, adapt it to suit their own
needs, modify the code and distribute it, modified or not.

If this means my code can't be used to implement DRM, that's a good
thing.

All anti-piracy measures I've ever known deprive users of legitimate
rights too.  So, yes, I don't agree with these measures.

I believe in punishing the guilty, not in punishing innocent
bystanders just such that the guilty have to find work arounds to
become guilty.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[david]

On Thu, 21 Jun 2007, Alexandre Oliva wrote:

> On Jun 21, 2007, david@lang.hm wrote:
>
>> On Thu, 21 Jun 2007, Alexandre Oliva wrote:
>>> If it's input-only, then you can't possibly harm the operation of the
>>> network by only listening in, can you?
>
>> Ok, so you consider any anti-piracy measures to be something that
>> GPLv3 should prohibit.
>
> In general, I object to the use of my code in ways that don't permit
> users to run it for any purpose, study it, adapt it to suit their own
> needs, modify the code and distribute it, modified or not.
>
> If this means my code can't be used to implement DRM, that's a good
> thing.
>
> All anti-piracy measures I've ever known deprive users of legitimate
> rights too.  So, yes, I don't agree with these measures.
>
> I believe in punishing the guilty, not in punishing innocent
> bystanders just such that the guilty have to find work arounds to
> become guilty.

this is your right with your code. please stop browbeating people who 
disagree with you.

David Lang

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 21, 2007, david@lang.hm wrote:

> this is your right with your code. please stop browbeating people who
> disagree with you.

For the record, GPLv2 is already meant to accomplish this.  I don't
understand why people who disagree with this stance chose GPLv2.
Isn't "no further restrictions" clear enough?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[david]

On Thu, 21 Jun 2007, Alexandre Oliva wrote:

> On Jun 21, 2007, david@lang.hm wrote:
>
>> this is your right with your code. please stop browbeating people who
>> disagree with you.
>
> For the record, GPLv2 is already meant to accomplish this.  I don't
> understand why people who disagree with this stance chose GPLv2.
> Isn't "no further restrictions" clear enough?

everyone else is reading this as 'no further license restrictions' not 'no 
hardware restrictions' becouse GPLv2 explicitly says that it has nothing 
to do with running the software, only with distributing it.

David Lang

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 21, 2007, david@lang.hm wrote:

> On Thu, 21 Jun 2007, Alexandre Oliva wrote:
>> On Jun 21, 2007, david@lang.hm wrote:
>> 
>>> this is your right with your code. please stop browbeating people who
>>> disagree with you.
>> 
>> For the record, GPLv2 is already meant to accomplish this.  I don't
>> understand why people who disagree with this stance chose GPLv2.
>> Isn't "no further restrictions" clear enough?

> everyone else is reading this as 'no further license restrictions'

I didn't see anyone else add "license" where you did.  "No further
restrictions on the rights granted herein" is very powerful and
extensive, and that's how it was meant to be.

> not no hardware restrictions' becouse GPLv2 explicitly says that it
> has nothing to do with running the software, only with distributing
> it.

It also says that running the software is not restricted, and since
copyright law in the US doesn't regulate execution, receiving the
software does grant the recipient the right to run the software.  So
the distributor can't impose restrictions on it.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

RE: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Schwartz]

On Jun 21, 2007, aoliva@redhat.com wrote:

> On Jun 21, 2007, david@lang.hm wrote:

> >> For the record, GPLv2 is already meant to accomplish this.  I don't
> >> understand why people who disagree with this stance chose GPLv2.
> >> Isn't "no further restrictions" clear enough?

> > everyone else is reading this as 'no further license restrictions'

> I didn't see anyone else add "license" where you did.  "No further
> restrictions on the rights granted herein" is very powerful and
> extensive, and that's how it was meant to be.

I agree. For example, a patent might impose a further restriction. The GPL
was clearly meant to foreclose that. There are any number of other ways of
nasty, subtle ways to impose further restrictions, and the GPLv2 meant to
foreclose all of them.

> > not no hardware restrictions' becouse GPLv2 explicitly says that it
> > has nothing to do with running the software, only with distributing
> > it.

> It also says that running the software is not restricted, and since
> copyright law in the US doesn't regulate execution, receiving the
> software does grant the recipient the right to run the software.  So
> the distributor can't impose restrictions on it.

Right.

The response to this argument is that it is mind-bogglingly obvious that the
GPL doesn't mean that no *authorization* decisions can stand in your way. It
didn't mean that I couldn't keep the root password to my server secret even
though that denies you the "right" to modify the Linux kernel running on it.

Some entity has to decide what software runs on any particular piece of
hardware, and it was never the intent of the GPL to specify or limit who
that person was. This has been discussed many times over many years, longer
before Tivoization was even thought of, and it was agreed that the GPL
didn't foreclose authorization obstacles to software modification.

Why doesn't Linux allow a non-root user to install a module or change which
kernel the system runs? Doesn't that limit the GPL rights of all non-root
users to modify the GPL'd kernel software on that machine? OF COURSE NOT.

DS

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Tomas Neme]

> > as long as this right is not used by the software distributor to
> > impose restrictions on the user's ability to adapt the software to
> > their own needs.  The GPLv3 paragraph above makes a fair concession in
> > this regard, don't you agree?
>
> no, one of the rules for the network is that the software must be
> certified, you are requireing the device to permit the software to be
> changed to an uncertified version.(to store credit card numbers and send
> them to a third party for example)

Also another way of doing this is having every network ask the kernel
for its key, and checking it. If it doesn't match a certified key,
then not allowing you to access the network.

Besides the fact that this would be a very costly approach, having
every network needing to update their certified keys list every time
TiVo and every other DVR vendor updates their kernels, it would also
prevent any form of modified software to give you any of the TiVo's
expected functionality: it would load, you would be able to play pong
on it, but not watch or record TV, and they can't be blamed for it,
because if the kernel's been tampered with, it might have been made so
it saves the video unencrypted on the Harddrive, and it certainly *is*
the network's right to stop you from doing so. So what the fuck do you
want from them?

T

-- 
|_|0|_|
|_|_|0|
|0|0|0|

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[david]

On Thu, 21 Jun 2007, Tomas Neme wrote:

>> >  as long as this right is not used by the software distributor to
>> >  impose restrictions on the user's ability to adapt the software to
>> >  their own needs.  The GPLv3 paragraph above makes a fair concession in
>> >  this regard, don't you agree?
>>
>>  no, one of the rules for the network is that the software must be
>>  certified, you are requireing the device to permit the software to be
>>  changed to an uncertified version.(to store credit card numbers and send
>>  them to a third party for example)
>
> Also another way of doing this is having every network ask the kernel
> for its key, and checking it. If it doesn't match a certified key,
> then not allowing you to access the network.

no, this doesn't work becouse if the software has been altered you don't 
know if the key it's giving you matches that software. it could be giving 
you the key from the unmodified software.

David Lang

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Andrew McKay]

david@lang.hm wrote:
> On Thu, 21 Jun 2007, Alexandre Oliva wrote:
> 
>> On Jun 21, 2007, david@lang.hm wrote:
>>
>>> how exactly can they prevent a system that's been tampered with from
>>> accessing their network?
>>
>> By denying access to their servers?  By not granting whatever is
>> needed to initiate network sessions?
>>
>> And note, "it's been tampered with" is not necessarily enough of a
>> reason to cut someone off, it has to meet these requirements:
> 
> how can the server tell if it's been tampered with?
> 

I agree with this statement.  Imagine a proprietary private network where a 
device has been modified to run in an invisible promiscuous mode.  The device 
looks as if it isn't doing anything wrong, but is forwarding the network out 
another interface.  The only way to prevent that type of attack is to not allow 
unauthorized signed Kernels onto that network.

Andrew McKay

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 21, 2007, Andrew McKay <amckay@iders.ca> wrote:

> david@lang.hm wrote:
>> how can the server tell if it's been tampered with?

> I agree with this statement.

Err...  That's a question, not a statement ;-)

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Andrew McKay]

Alexandre Oliva wrote:
> On Jun 21, 2007, Andrew McKay <amckay@iders.ca> wrote:
> 
>> david@lang.hm wrote:
>>> how can the server tell if it's been tampered with?
> 
>> I agree with this statement.
> 
> Err...  That's a question, not a statement ;-)
> 

Sorry, that's what happens when one types before getting a cup of coffee in the 
morning.

Andrew

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Tim Post]

On Wed, 2007-06-20 at 22:30 -0700, david@lang.hm wrote:

> 
> asking a device that's running software that you haven't verified to give 
> you a checksum of itself isn't going to work becouse the software can just 
> lie to you.
> 

I don't think there is any way I _could_ make a device if it had to be
tamper proof and use free software if that was the case. 

I'd need to make some kind of proprietary network connection back to my
company that used its own network device. I could not trust it if the
free kernel could touch it, if I wanted to allow a modified in place
kernel.

If I hope for that device to use the internet to talk to me (i.e. just a
secondary nic), I'd have to write my own kernel to power this second
network device that was capable of encrypting and validating traffic
over tcp-ip. Or I have to pull my own copper to every location where my
device is used.

So either way, I'm writing my own kernel if I want to do that, because I
could not POSSIBLY allow the kernel talking to my private connection to
the device to be modified. 

What a nasty, vicious cycle that would be. Yikes!
 

--Tim

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Andrew McKay]

Alexandre Oliva wrote:
> On Jun 20, 2007, Andrew McKay <amckay@iders.ca> wrote:
> 
>> However, I don't see how this would ever require a company like Tivo
>> or Mastercard to have their networks play nice with a unit that has
>> been modified by the end user, potentially opening up some serious
>> security holes.
> 
> Which is why the GPLv3 doesn't make the requirement that you stated.
> 

So if it's not a requirement of the GPLv3, then Tivo could deny content based 
signing the binary image of the Linux kernel and using that signature as 
authentication on their network (or their content providers network).  A 
modified Tivo box would not be able to preform it's original task of being a PVR 
at that point, at least with the content provider's signal.  Seems pretty 
pointless to me.  Seems like almost the same thing as not allowing an unsigned 
Linux kernel to boot on the system.  Though it would still be possible to get 
the Tivo box to play tetris or something like that.

Andrew McKay

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alan Cox]

> > You've made an important mistake. You said "their system". Now its "our
> > code" and "whoever bought the units' hardware" so it isn't their anything.
> 
> Yes, the hardware belongs to the user, and the software belongs to the Linux 
> community.   However I think I wasn't 100% clear, I also mean keeping companies 
> networks and content secured.  Credit card companies insuring the software 
> hasn't been modified to skim cards (not that it's the only way to skim a card), 

If credit card companies are doing this they are failing badly and it
clearly isn't working. Also lets be clear about this - I don't need any
credit card company network access to skim older cards, and the newer
ones have been broken by various non software schemes.

>   or Tivo making sure that their content providers are protected.  Lets look at 
> the credit card example.  Sure the user could modify the system and boot their 
> own kernel, but it doesn't have to play nice with Mastercard's network anymore. 
>   Or better yet, would actually report that a certain business's card reader had 
> been tampered with.

That to me is a fair comment. I should IMHO be able to load my code and
my keys on my Tivo. And Walt Disney in return probably should be quite
free not to trust my keys. You need some fairly strong competition law
enforcement to make all that work right in the marketplace but as a
philosophical basis it seems fine. In practice it is likely to lead to
serious monopoly abuse problems and all sorts of ugly tying of goods that
you don't want in a free market. Given the completely ineffectual way the
US enforces its anti-monopoly law, and the slowness of the EU at it the
results might well be bad - but for other reasons.

Alan

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Andrew McKay]

Alan Cox wrote:
>>> You've made an important mistake. You said "their system". Now its "our
>>> code" and "whoever bought the units' hardware" so it isn't their anything.
>> Yes, the hardware belongs to the user, and the software belongs to the Linux 
>> community.   However I think I wasn't 100% clear, I also mean keeping companies 
>> networks and content secured.  Credit card companies insuring the software 
>> hasn't been modified to skim cards (not that it's the only way to skim a card), 
> 
> If credit card companies are doing this they are failing badly and it
> clearly isn't working. Also lets be clear about this - I don't need any
> credit card company network access to skim older cards, and the newer
> ones have been broken by various non software schemes.

Agreed.  Credit card companies are failing very badly at preventing skimming. 
They definitely need to rethink their model of how the credit card system should 
work.

> 
>>   or Tivo making sure that their content providers are protected.  Lets look at 
>> the credit card example.  Sure the user could modify the system and boot their 
>> own kernel, but it doesn't have to play nice with Mastercard's network anymore. 
>>   Or better yet, would actually report that a certain business's card reader had 
>> been tampered with.
> 
> That to me is a fair comment. I should IMHO be able to load my code and
> my keys on my Tivo. And Walt Disney in return probably should be quite
> free not to trust my keys. You need some fairly strong competition law
> enforcement to make all that work right in the marketplace but as a
> philosophical basis it seems fine. In practice it is likely to lead to
> serious monopoly abuse problems and all sorts of ugly tying of goods that
> you don't want in a free market. Given the completely ineffectual way the
> US enforces its anti-monopoly law, and the slowness of the EU at it the
> results might well be bad - but for other reasons.

I agree as well, the model could be heavily abused.  I'm not sure what the 
solution is as far as fighting monopolies and corporate greed.  But I'd rather 
that it was fought through education of consumers and not with a license 
agreement that should be giving people freedom.  A balance of freedom to the 
licensee and the licenser.  It's my opinion that GPLv3 potentially shifts the 
balance too far to the licensee.

Andrew McKay

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 21, 2007, Andrew McKay <amckay@iders.ca> wrote:

> A balance of freedom to the licensee and the licenser.  It's my
> opinion that GPLv3 potentially shifts the balance too far to the
> licensee.

It's more of a balance of freedom between licensee and licensee,
actually.  It's a lot about making sure no one can acquire a
privileged position, such that every licensee plays under the same
rules.  (The copyright holder is not *acquiring* a privileged
position, copyright law had already granted him/her that position.)

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Andrew McKay]

Alexandre Oliva wrote:
> On Jun 21, 2007, Andrew McKay <amckay@iders.ca> wrote:
> 
>> A balance of freedom to the licensee and the licenser.  It's my
>> opinion that GPLv3 potentially shifts the balance too far to the
>> licensee.
> 
> It's more of a balance of freedom between licensee and licensee,
> actually.  It's a lot about making sure no one can acquire a
> privileged position, such that every licensee plays under the same
> rules.  (The copyright holder is not *acquiring* a privileged
> position, copyright law had already granted him/her that position.)
> 

I do see what you're saying here.  But it does take the away the ability of a 
licensee to protect themselves from another malicious licensee.  If the ultimate 
goal of the Free Software community is to get source code out to the public, I 
think that was captured in GPLv2.  GPLv3 oversteps its bounds.

Anyways I think this topic has been quite covered.

Andrew

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 21, 2007, Andrew McKay <amckay@iders.ca> wrote:

> Alexandre Oliva wrote:
>> On Jun 21, 2007, Andrew McKay <amckay@iders.ca> wrote:
>> 
>>> A balance of freedom to the licensee and the licenser.  It's my
>>> opinion that GPLv3 potentially shifts the balance too far to the
>>> licensee.
>> 
>> It's more of a balance of freedom between licensee and licensee,
>> actually.  It's a lot about making sure no one can acquire a
>> privileged position, such that every licensee plays under the same
>> rules.  (The copyright holder is not *acquiring* a privileged
>> position, copyright law had already granted him/her that position.)

> I do see what you're saying here.  But it does take the away the
> ability of a licensee to protect themselves from another malicious
> licensee.

Sorry, I don't follow what the "it" refers to in your sentence.

> If the ultimate goal of the Free Software community is to get source
> code out to the public, I think that was captured in GPLv2.

That's a correct logical inference, but since the premise is false,
the conclusion is garbage.

GPLv2 goes far beyond getting source code out to the public.  It
contains the "no further restrictions" language, which is very
powerful.  It is pretty obvious that when Linus adopted GPLv2 he
didn't realize it reached that point.  That when Tivo invented
Tivoization, he decided he wanted to permit this, and thus grants an
implicit additional permission for anyone to do it with his code,
doesn't mean other participants in the Linux community feel the same
way, or read the GPLv2 the same way, and could be somehow stopped from
enforcing the license the way they meant it.

Ultimately, the current situation is that we have two
mutually-incompatible license intents being used in Linux, and no
matter how much those who want to grant the permission say so, this
doesn't trample other contributor's rights to enforce the license they
chose for their code.  Especially those who started contributing long
before the decision that "what TiVo does is good" was announced.

Now, since these two license intents are expressed by the same
license, and what the license demands is that derived works must be
under the same license, they are compatible, but since the intents are
distinct, what prevails is, as in any case of combination of different
licensing provisions, is the most restrictive provision.

So Linux does not permit tivoization today.  Linus does, Linux
doesn't.

All this fuss about the anti-tivoization provisions in GPLv3 is just a
consequence of reading the GPLv2 without fully understanding its
intended consequences, not having foresight to clarify the intent to
constrain the "no further restrictions" provisions to match the
alternate interpretation, and opposing the removal of the ambiguity
because it doesn't match the choice that *some* of the developers
would like it to go.

Who's the ambiguity good for?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Tomas Neme]

> powerful.  It is pretty obvious that when Linus adopted GPLv2 he
> didn't realize it reached that point.  That when Tivo invented
> Tivoization, he decided he wanted to permit this, and thus grants an
> implicit additional permission for anyone to do it with his code,
> doesn't mean other participants in the Linux community feel the same
> way, or read the GPLv2 the same way, and could be somehow stopped from
> enforcing the license the way they meant it.

The thing is, what matters in copyright and licencing matters is what
the author of the code understands, no the licence's author, if
ambiguous. And the kernel's rights holder is Linus. The authors of the
particular bits of code can complain about what tivo's doing, but
since TiVo's using the linux kernel, and GPLv2 says

"These requirements apply to the modified work as a whole.  If
identifiable sections of that work are not derived from the Program,
and can be reasonably considered independent and separate works in
themselves, then this License, and its terms, do not apply to those
sections when you distribute them as separate works.  But when you
distribute the same sections as part of a whole which is a work based
on the Program, the distribution of the whole must be on the terms of
this License, whose permissions for other licensees extend to the
entire whole, and thus to each and every part regardless of who wrote it."

Linus has the last word on it. IANAL, but as far as I can tell this
reads as "this licence applies to the whole, not to the parts by
themselves", and so does who the "holder" is, I'd believe.

So if you own a part of the kernel, then you can pursuit TiVo on your
own, if they did direct use of that part especifically and break (in
your opinion) what you feel GPLv2 means. You can form the CATV2
(CodersAgainstTiVo v2) and try to get TiVo to stop using the Linux
Kernel for their product (because, believe me, they WON'T release the
keys). Yeay, we lost Tivo's improvements on the kernel, and the
posibility of having a working kernel if anyone feels like
back-ingeneering TiVo for their own amusement.

T

-- 
|_|0|_|
|_|_|0|
|0|0|0|

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 22, 2007, "Tomas Neme" <lacrymology@gmail.com> wrote:

> The thing is, what matters in copyright and licencing matters is what
> the author of the code understands, no the licence's author, if
> ambiguous. And the kernel's rights holder is Linus.

Since he didn't get copyright assignments, each contributor is the
copyright holder of her/his own contribution.  And this means each
holder gets a say on how s/he understood GPLv2.

IANAL, but I think if Linus' intended interpretation had been
clarified all the way from the beginning, he could have grounds to
claim that everyone else had implicitly accepted that reading by
contributing to the project.

But since it was a decision made many years later, his clarification
on his reading of the license is in a way an additional permission
that affects only his own contributions; other authors are still
entitled to try to enforce their understanding of the legal terms of
the license, and they would have the spirit of the GPL and its
preamble on their side to guide the interpretation.  Even if contract
law states something like, in adhesion contracts, the party who writes
the contract gives the other party the benefits of any ambiguity in
the writing, the GPL is not a contract, it's a license, and per
copyright law, licenses are to be interpreted restrictively.

> Linus has the last word on it.

In the sense that he can decide to remove all contributions from
dissenting authors, yes, he does.  But he can't impose his more lax
interpretation upon other authors.  Under copyright, it's the more
restrictive reading that prevails, in that any holder who understands
his rights are being trampled can enforce them.  And since at least
one such author is vocal in his dissent, not even estoppel defenses
would apply.  But IANAL.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Tomas Neme]

> In the sense that he can decide to remove all contributions from
> dissenting authors, yes, he does.  But he can't impose his more lax
> interpretation upon other authors.  Under copyright, it's the more

yes, I saw my argument going weak as I wrote it, but what I said later:

> So if you own a part of the kernel, then you can pursuit TiVo on your
> own, if they did direct use of that part especifically and break (in
> your opinion) what you feel GPLv2 means. You can form the CATV2
> (CodersAgainstTiVo v2) and try to get TiVo to stop using the Linux
> Kernel for their product (because, believe me, they WON'T release the
> keys). Yeay, we lost Tivo's improvements on the kernel, and the
> posibility of having a working kernel if anyone feels like
> back-ingeneering TiVo for their own amusement.

is still right. What I meant, at least by the end of that email, was
that he has the last word on trying to stop TiVo from using The Linux
Kernel. Each author can still go and stop them from using his part,
and the derivative work that is The Linux Kernel.

But that brings another question: what if TiVo decided to remove all
code from the complaining parts and rewrite them? that wouldn't be The
Linux Kernel anymore, but it would be a derivative work of all the
parts that don't disagree with Tivoization, but is that legal?

-- 
|_|0|_|
|_|_|0|
|0|0|0|

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Tim Post]

On Wed, 2007-06-20 at 11:27 -0500, Andrew McKay wrote:

> I'm not going to address whether GPLv3 changed the spirit of GPLv2, but saying 
> that licensing the Linux Kernel under GPLv3 will result in more contributions is 
> absolute BS.

Is there a system in place that sort of keeps track? Argument aside, I'm
really interested to see how adoption in devices encourages
contributions (and what kinds of contributions it attracts). Anyone
would be I guess.

How do you guys keep up with it? My head almost blows up just thinking
about it given the traffic of just _this_ list.

TIA :)
--Tim

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Andrea Arcangeli]

On Sun, Jun 17, 2007 at 04:46:44PM -0300, Alexandre Oliva wrote:
> My perception is that the first easily dominates the second, and so
> you are better off without tivoization.

Your perception is quite flawed.

I see where you come from, I know your intentions are absolutely
genuine, but there's not a chance that by changing the GPL in any way
you want, you can move the needle on companies whose only real long
term threat is so broad and fundamental as the very existence of the
Internet! Stay away from any DRM issue since that is not a licensing
problem, and is a problem that is up to the economy to solve, all we
have to do is to write good code and grow the userbase, those
discussions are a total waste of time, and infact they're in the
interest of the pro-DRM people. See apple removal of drm mandated by
market forces (sure not mandated by the threat of GPLv3 ;).

Then (besides tivo that as said at the top is not a problem at all
regardless if it's legal or not with gplv2) there's the myth of
trusted computing apocalypse that promises an unbreakable DRM and no
computer capable of running a modified linux anymore, which cannot
materialize. At every respectable linux user group there's somebody
giving that TC speech. The thing, if I'm wrong and it really happens,
it means something went so fundamentally wrong that the symptom will
be a totally minor problem compared to the real cause that triggered
it. If nothing else wait the scenario to remotely materialize before
declaring preventive war to something purely theoretical, and _then_
release a v4 addressing it.

If all conspiracy theories should be added to v3 as potential threat,
then I would recommend to as well add a clause that if you're an alien
that wants to use some GPLv3 code in your alien-technology-driven ship
built to destroy planet earth, you can't unless you provide us with a
copy of your ship open specifications showing how to upload our
improved GPLv3 software to it, so we have a chance to build one too to
defend ourself LOL. Doesn't that sound fair enough too? I mean just in
case ;)

Back to the tivoization issue, the crypto key is the least of the
problems, the major linux cellphone vendor ships binary only modules
and I wouldn't even know how to upgrade the kernel regardless of any
crypto signature and even ignoring the binary only modules. If selling
the locked embedded package allows new market segments to grow around
linux (even if that's not _yet_ an ideal hacker-hackable cellphone)
that's still a net-positive because it sends a message to the venture
capital that may exploit the fact they're closed to grow market share
(see openmoko, not a "perception" of mine).

Open source licenses shouldn't forbid usages, not even the blatantly
unethical ones, "evil" is not tangible (I guess everything would be
easier in life if it was).

Let's tackle on the only real _tangible_ problem of gplv2 known
todate, that they can address with a few liner fix to gplv2. They
should _only_ do that, and release quickly a strightforward v3 that
nobody could ever pretend to argue about (they should have done that
already, more than half an year passed already and most of the time at
least here has been spent on purely theoretical things). They still
can do the right thing. All this arguing, busybox forking, are all
signals that something is wrong, it should start to ring a
bell. Frankly I think we all love the FSF and the GPL and we want to
help to avoid mistakes (I know I do).

I'm totally in favor to __experiment__ with the DRM clause but do that
on a new license. Leave GPL as it has always been, it works just
great, fix the only single tangible issue known todate, the rest is
all about conspiracy theories and at definining evil. Perfect fairness
cannot be obtained in this world, no matter what license or system you
apply.

Those developers that have been so totally trustful and used "any
later version" deserve better treatment (they deserve a quick fix too)
and you should obey to the promise of only modification in detail
according to point 9 or the "any later version" will not be
enforceable because outlawing a single usage means going the opposite
route of "promoting the reuse of the software" (written in gplv2,
search for it).

This whole email is irrelevant with the fact tivo may or may not be
legal with gplv2 depending on different countries. The single attempt
of trying to reduce the gpl userbase with new restriction, doesn't
qualify as a modification in detail here. For the record, I said the
first time quite anonymously in my blog back in Oct 06.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 17, 2007, Andrea Arcangeli <andrea@suse.de> wrote:

> Open source licenses shouldn't forbid usages, not even the blatantly
> unethical ones, "evil" is not tangible (I guess everything would be
> easier in life if it was).

Since you mention Open source...

What if I showed you that tivoization discriminates against field of
endeavor?

  http://opensource.org/docs/osd

  6. No Discrimination Against Fields of Endeavor

  The license must not restrict anyone from making use of the program
  in a specific field of endeavor. For example, it may not restrict
  the program from being used in a business, or from being used for
  genetic research.

Wait, TiVo, you're saying I cannot use this modified version of the
program to do $whatever in this computer you sold me, and that's
solely because you don't want to let me?

Is this forceful restriction on how or where the program can be used
not a discrimination against fields of endeavor?

It's not a restriction imposed by the license of the program, but it
is a restriction imposed by the manufacturer of the equipment that
distributes the program, therefore, the conditions under which the
program is distributed are not only non-Free Software, they also fail
to meet the Open Source definition.

Neat, huh?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Johannes Stezenbach]

On Sun, Jun 17, 2007, Alexandre Oliva wrote:
> 
> Serious, what's so hard to understand about:
> 
>   no tivoization => more users able to tinker their formerly-tivoized
>   computers => more users make useful modifications => more
>   contributions in kind
> 
> ?
> 
> Sure, there's a downside too:
> 
>   no tivoization => fewer contributions from manufacturers that demand
>   on tivoization
> 
> 
> My perception is that the first easily dominates the second, and so
> you are better off without tivoization.


You seem to focus on _preserving the four freedoms_, however
the FSF also has the goal of _promoting free software_.
This goal is even written down in the GPLv2 itself, in section 10.

I think those two goals are somewhat conflicting. If you want to
win people for free software, you need to make it easy for them
to accept your ideas. However, in order to make it easy you have to
make compromises wrt the four freedoms.


The FSF apparently has decided that they won't compromise on
the freedoms as stated in The Free Software Definition.
The message they send out is "if you don't agree to our ethical
principles, get lost."

OTOH Linus has decided he wants "world domination".
So he makes it easy for other people to join by saying
"if you know what tit-for-tat means, welcome."


On the surface this might look like two totally different goals.
But IMHO it's not so much the goal which is different, but
the _strategy_ to reach the goal. In the end the Linux
people also want truly free software, they just believe
they can get there without forcing people.

Honestly, I think Linus' strategy has been much more
successful so far than the FSF's. Of course there is no Linux
system without GNU software, that's not the point. But I
think without Linus' lenient interpretation of the GPLv2 there
would be much less people using free software than
there are today.


If you look at the lenght of this thread, don't you realize
that even when you talk to software developers on a mailing list
dealing with free software/open source, you have trouble
to get acceptance for your fundamentalistic view of the
ethical principles of free software?

And you haven't even started to talk to the business people,
executives and lawyers which you need to convince if you
want to make free software ubiquitous. "We have high ethical
standards" is something these people don't understand or
care about. "It's a bargain, you can use it for free and all you
have to do is give back your changes" is what might work
to win them.


The bottom line is that I think your perception is completely wrong.

For sure I would like to be able to change the software in my
TiVo, too, but when I look at the big picture then I think
it's more important to have free software in every device,
on every computer, and accept that a few "unfree" devices
along the way are a price worth to pay.


Johannes

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 18, 2007, Johannes Stezenbach <js@linuxtv.org> wrote:

> I think those two goals are somewhat conflicting. If you want to
> win people for free software, you need to make it easy for them
> to accept your ideas. However, in order to make it easy you have to
> make compromises wrt the four freedoms.

If you make compromises, it ceases to be free software.  *And*, in the
pragmatic plane, you lose the benefits from users whose freedoms will
be restricted by others.

> If you look at the lenght of this thread, don't you realize
> that even when you talk to software developers on a mailing list
> dealing with free software/open source, you have trouble
> to get acceptance for your fundamentalistic view of the
> ethical principles of free software?

I do.  I get rejection from a number of people from whom I expected to
get rejection, those with fundamentalist but opposing positions.
Meanwhile, in private, I get lots of voices of encouragement and
gratitude for what I'm doing.  I suppose they might even do that in
public if they didn't care about getting verbal abuse for exposing
dissenting opinions.

> And you haven't even started to talk to the business people,
> executives and lawyers which you need to convince if you
> want to make free software ubiquitous.

Oh, really?  How do you know?

People talk a lot about TiVo here, but do they the faintest idea of
how the conversations with TiVo are proceeding?  I thought so...

> "It's a bargain, you can use it for free and all you have to do is
> give back your changes" is what might work to win them.

And that's precisely what I've been working on.  But fundamentalism
can indeed blind people.  It works both ways, I guess.

> The bottom line is that I think your perception is completely wrong.

Do you think any part of this reasoning is wrong?

>   no tivoization => more users able to tinker their formerly-tivoized
>   computers => more users make useful modifications => more
>   contributions in kind

I know you see the other possibility:

>   no tivoization => fewer contributions from manufacturers that demand
>   on tivoization

and there's another, that's break-even for the community so I didn't
even mention it:

    no tivoization => ROM software, no difference for the community


So, you see, when people who oppose anti-tivoization measure the
outcome for the community, they only look at the second possibility,
assuming the vendor would immediately switch to some other software.
As if that was easy for the vendor, and as if the software sucked so
much that the vendor was just looking for a reason to switch.

But since the software is good, and moving to another software would
be costly in various dimentions, the vendor has an incentive to stick
with the software they have.

So the vendor will look into respecting users' freedoms, and they
might just do that, rather than switching to a tivoizable software or
facing the potential costs of ROM replacements at every software
update.

And if just a few vendors take the stance of respecting users'
freedoms, the community will gain not only more users, but also more
developers more motivated to improve the software to serve their own
interests, because they *will* be able to use the results of their
modifications on their devices.


So you see, the picture of anti-tivozation is not as bleak as people
try to frame it.  In fact, it's not bleak at all.  If one out of 10,
maybe even 1 out of 100 vendors start respecting users' freedoms, when
faced with anti-tivoization provisions, the community will already win
big time, because each vendor is likely to have thousands of
customers, some of which will use the freedoms to serve the goals of
the community, in the very terms the community claims to care about.


So, what flaw do you see in this reasoning?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Johannes Stezenbach]

On Mon, Jun 18, 2007, Alexandre Oliva wrote:
> 
> People talk a lot about TiVo here, but do they the faintest idea of
> how the conversations with TiVo are proceeding?  I thought so...

Oh, if you know something we don't, could you please fill us in?
And who was it who coined the "Tivoization" term, thus putting
TiVo into focus?

> So, you see, when people who oppose anti-tivoization measure the
> outcome for the community, they only look at the second possibility,
> assuming the vendor would immediately switch to some other software.
> As if that was easy for the vendor, and as if the software sucked so
> much that the vendor was just looking for a reason to switch.
> 
> But since the software is good, and moving to another software would
> be costly in various dimentions, the vendor has an incentive to stick
> with the software they have.

Hm, you only talk about people who already use free software,
but I tried to make you aware of the importance of
_promoting_ free software, i.e. winning new people and
companies for the free software idea.

I think the majority of embedded devices still run proprietary
RTOSes, and the majority of desktops still run Windows or Mac OS.
Don't you want to change that?

There are dozens of proprietary RTOSes, and along with
them dozens of proprietary toolchains, development environments
and trace/debug tools. Companies which worked in this field
for decades have invested money to create proprietary software
on top of them, and to train their staff to use them. Those companies
won't switch to Linux lightly. And it won't be a singular event,
but a process. They might start low, and maybe (hopefully) might
become well-playing free software contributors. But if you raise
the entry barrier too high, they won't get started at all.

OK, I don't have experience talking to big companies, but
I have talked to people working for smaller ones. They are
aware of the trend towards Linux, but are afraid that the
obligations of the GPL might be impractical for them.
Then they only have the choice to not use Linux, or to use
"creative workarounds".

It's true that what these companies do might have little direct benefit
for users buying their products, however the long term benefits of
getting the people in these companies exposed to free software ideas,
and in contact with the free software community, can only be
positive -- I think it's more important to spread the general
idea of free software into as many minds as possible than to
ensure that few follow the pure spirit of the FSF free software
definition in every detail.

> So you see, the picture of anti-tivozation is not as bleak as people
> try to frame it.  In fact, it's not bleak at all.  If one out of 10,
> maybe even 1 out of 100 vendors start respecting users' freedoms, when
> faced with anti-tivoization provisions, the community will already win
> big time, because each vendor is likely to have thousands of
> customers, some of which will use the freedoms to serve the goals of
> the community, in the very terms the community claims to care about.

Does this multiplicator also apply to new companies
which start using free software for their products?

I think the FSF strategy is suboptimal. The Linux
strategy works better for promoting free software.

In the end I want my devices to be open and hackable, too,
and I'm sure it will take an effort to convince companies
to open up. But I'm not convinced that the GPLv3 is a
step in the right direction towards that goal.


Johannes

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[david]

On Tue, 19 Jun 2007, Johannes Stezenbach wrote:

> On Mon, Jun 18, 2007, Alexandre Oliva wrote:
>>
>> People talk a lot about TiVo here, but do they the faintest idea of
>> how the conversations with TiVo are proceeding?  I thought so...
>
> Oh, if you know something we don't, could you please fill us in?
> And who was it who coined the "Tivoization" term, thus putting
> TiVo into focus?

what conversations are going on?

Tivo checked years ago and were told that what they were going to do was 
Ok. I don't know thatanyone is talking too Tivo about anything. they are 
just screaming about how evil Tivo is at every public opportunity.

David Lang

>> But since the software is good, and moving to another software would be 
>> costly in various dimentions, the vendor has an incentive to stick with 
>> the software they have.

but if regulations or other contracts require tamper-resistant hardware 
they have no choices other then to fork the existing GPLv2 versions or 
switch to alternate options for anything that switches to GPLv3

David Lang

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 18, 2007, david@lang.hm wrote:

> On Tue, 19 Jun 2007, Johannes Stezenbach wrote:
>>> But since the software is good, and moving to another software
>>> would be costly in various dimentions, the vendor has an incentive
>>> to stick with the software they have.

> but if regulations or other contracts require tamper-resistant
> hardware they have no choices other then to fork the existing GPLv2
> versions or switch to alternate options for anything that switches to
> GPLv3

Where by "alternate options" I hope you mean non-copyleft or
tivoizable (copyleft by definition) software, or (newer versions of)
the same software they already use, but in ROM.  (I point this out
because people keep forgetting all the available options when they
claim to enumerate all available options ;-)

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Is kdb package available for 2.6 Linux on PowerPC?

[gshan]

Hey Guys,

Recently, I need work on my TTY (console) driver so that it could 
support kdb. I didn't have much experiences on linux kdb. So my 
questions are:

1) 2.6 Linux for PowerPC could support kdb? where can I find the source 
code?

2) If 2.6 Linux for PowerPC doesn't support kdb, where could I get the 
package and integrated it to kernel?

Thanks in advance,
Gavin

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 18, 2007, Johannes Stezenbach <js@linuxtv.org> wrote:

> On Mon, Jun 18, 2007, Alexandre Oliva wrote:

>> People talk a lot about TiVo here, but do they the faintest idea of
>> how the conversations with TiVo are proceeding?  I thought so...

> Oh, if you know something we don't, could you please fill us in?

Honestly, I don't know either.  But I get an impression that there are
conversations underway.

> And who was it who coined the "Tivoization" term, thus putting
> TiVo into focus?

AFAIK TiVo invented the practice, did they not deserve the credit?

> Hm, you only talk about people who already use free software,
> but I tried to make you aware of the importance of
> _promoting_ free software, i.e. winning new people and
> companies for the free software idea.

Aah, I see.  Indeed, I'd missed that aspect.  Sorry about that.

My take on it is that bringing free loaders in doesn't help us much,
and bringing them in in a way that they don't learn the essential
aspects of the community will hurt the community in the long run.

So they must become aware that respecting others' freedoms is not only
the right thing to do, from a moral and ethical standpoint, but also
that this is precisely what enables our community to thrive, and to
enable everyone to get the best out of the software we cooperate to
develop.

> I think the majority of embedded devices still run proprietary
> RTOSes, and the majority of desktops still run Windows or Mac OS.
> Don't you want to change that?

Sure.  But getting those companies to adopt Free Software in a way
that turns it into non-Free Software doesn't change that in any way.

Of course we might get some additional contributions here and there,
but then more and more users would still be stuck, unable or limited
in the ways and incentives they have to participate in our community.
Permitting this is very short-sighted.  It might bring us apparent
advantages in the short run, but the more such disrespects there are,
the more there will be, and the fewer users will be able to become
developers.  In the end, this may kill the whole process, in a tragedy
of the commons.  In the article linked below, I argue this very point,
comparing how the demand for respecting users' freedoms is what keeps
the free-loaders away and makes the GPL the most cost-effective
license for software development, compared with permissive licenses
and non-Free licenses.  The very same arguments apply to a comparison
between a license that permits tivoization and one that doesn't,
because the latter is more likely to have more contributors to share
the load, and both equally reduce the likelihood of unmergeable forks.
http://www.lsd.ic.unicamp.br/~oliva/papers/free-software/BMind.pdf

> if you raise the entry barrier too high, they won't get started at
> all.

I acknowledge this argument, but I hear the same arguments against the
GPLv2, claiming the barrier is too high, and it's not from people who
believe that tivoization is already prohibited.

> They are aware of the trend towards Linux, but are afraid that the
> obligations of the GPL might be impractical for them.  Then they
> only have the choice to not use Linux, or to use "creative
> workarounds".

Or to respect users' freedoms, enabling/motivating those users to
become developers in our community.

> It's true that what these companies do might have little direct
> benefit for users buying their products, however the long term
> benefits of getting the people in these companies exposed to free
> software ideas, and in contact with the free software community, can
> only be positive

As long as they understand how the community works, be it from the
moral and ethical standpoint, be it from the pragmatic standpoint.  In
both cases, the end result is that they learn that, when they share
and cooperating, respecting users freedoms (enabling and providing
incentive for them to improve the software), everybody wins,
themselves included.

>> So you see, the picture of anti-tivozation is not as bleak as people
>> try to frame it.  In fact, it's not bleak at all.  If one out of 10,
>> maybe even 1 out of 100 vendors start respecting users' freedoms, when
>> faced with anti-tivoization provisions, the community will already win
>> big time, because each vendor is likely to have thousands of
>> customers, some of which will use the freedoms to serve the goals of
>> the community, in the very terms the community claims to care about.

> Does this multiplicator also apply to new companies
> which start using free software for their products?

Of course, even more so!  Then you win not only the contributions from
the user, but also from the company itself, which you didn't have
before.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Johannes Stezenbach]

On Mon, Jun 18, 2007, Alexandre Oliva wrote:
> On Jun 18, 2007, Johannes Stezenbach <js@linuxtv.org> wrote:
> 
> > Hm, you only talk about people who already use free software,
> > but I tried to make you aware of the importance of
> > _promoting_ free software, i.e. winning new people and
> > companies for the free software idea.
> 
> Aah, I see.  Indeed, I'd missed that aspect.  Sorry about that.
> 
> My take on it is that bringing free loaders in doesn't help us much,
> and bringing them in in a way that they don't learn the essential
> aspects of the community will hurt the community in the long run.
> 
> So they must become aware that respecting others' freedoms is not only
> the right thing to do, from a moral and ethical standpoint, but also
> that this is precisely what enables our community to thrive, and to
> enable everyone to get the best out of the software we cooperate to
> develop.

The keywords here are "learn" and "become aware": It's a process
which takes time, and which requires ongoing communication.

I argue that if you keep the free loaders out, you miss
the chance to communicate with and educate them.
Communication across borders doesn't work well, and you create
a border between the morally "good" and the "bad".

Of course you can't expect that every free loader will
learn and accept the free software philosopy, some just
won't. But to me that's acceptable, and the GPLv2, or indeed
Linus' tit-for-tat interpretation of the GPLv2, is IMHO
sufficient to protect my interests.

> Of course we might get some additional contributions here and there,
> but then more and more users would still be stuck, unable or limited
> in the ways and incentives they have to participate in our community.
> Permitting this is very short-sighted.  It might bring us apparent
> advantages in the short run, but the more such disrespects there are,
> the more there will be, and the fewer users will be able to become
> developers.  In the end, this may kill the whole process, in a tragedy
> of the commons.  In the article linked below, I argue this very point,
> comparing how the demand for respecting users' freedoms is what keeps
> the free-loaders away and makes the GPL the most cost-effective
> license for software development, compared with permissive licenses
> and non-Free licenses.  The very same arguments apply to a comparison
> between a license that permits tivoization and one that doesn't,
> because the latter is more likely to have more contributors to share
> the load, and both equally reduce the likelihood of unmergeable forks.
> http://www.lsd.ic.unicamp.br/~oliva/papers/free-software/BMind.pdf

I'm not arguing about the GPL, especially not against the GPLv2.

What I'm concerned about is that the language you use trying to
promote the GPLv3 is IMHO anti-promotion of free software.

I believe executives don't read licenses. What they'll read is
the random article about "GPLv3 to outlaw tivoization",
"FSF wants to keep free loaders away" etc.

What is more likely, that they'll go to www.gnu.org to read
and absorb the GNU philosopy in order to become an accepted
member of the free software community, or that they'll decide to
stick with their proprietary RTOS then?


Johannes

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Manu Abraham]

Johannes Stezenbach wrote:

> I argue that if you keep the free loaders out, you miss
> the chance to communicate with and educate them.
> Communication across borders doesn't work well, and you create
> a border between the morally "good" and the "bad".
> 
> Of course you can't expect that every free loader will
> learn and accept the free software philosopy, some just
> won't. But to me that's acceptable, and the GPLv2, or indeed
> Linus' tit-for-tat interpretation of the GPLv2, is IMHO
> sufficient to protect my interests.

Err .. when you say protection on one hand and on the other you state
it's hard to keep free loaders away, then don't you think that those 2
are two completely different things ?

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Johannes Stezenbach]

On Tue, Jun 19, 2007, Manu Abraham wrote:
> Johannes Stezenbach wrote:
> 
> > I argue that if you keep the free loaders out, you miss
> > the chance to communicate with and educate them.
> > Communication across borders doesn't work well, and you create
> > a border between the morally "good" and the "bad".
> > 
> > Of course you can't expect that every free loader will
> > learn and accept the free software philosopy, some just
> > won't. But to me that's acceptable, and the GPLv2, or indeed
> > Linus' tit-for-tat interpretation of the GPLv2, is IMHO
> > sufficient to protect my interests.
> 
> Err .. when you say protection on one hand and on the other you state
> it's hard to keep free loaders away, 

I didn't say that.

IMHO it isn't even useful to try to keep free loaders away,
it's better to try and integrate them gradually. That's part
of the game.
(Where "free loaders" is a term introduced by Alexandre, not by me.)

The GPLv2 is a sufficient tool to defend free software
against those that don't even grasp tit-for-tat. But if
they do, you can talk to them *as peers* and try to convince
them that there's more to free software than just tit-for-tat.
But it has to be their decision, IMHO it's wrong to force them.

The GPLv3 tries to be a tool to defend against those that
don't subscribe to the full Free Software Definition.


Johannes

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Manu Abraham]

Johannes Stezenbach wrote:
> On Tue, Jun 19, 2007, Manu Abraham wrote:
>> Johannes Stezenbach wrote:
>>
>>> I argue that if you keep the free loaders out, you miss
>>> the chance to communicate with and educate them.
>>> Communication across borders doesn't work well, and you create
>>> a border between the morally "good" and the "bad".
>>>
>>> Of course you can't expect that every free loader will
>>> learn and accept the free software philosopy, some just
>>> won't. But to me that's acceptable, and the GPLv2, or indeed
>>> Linus' tit-for-tat interpretation of the GPLv2, is IMHO
>>> sufficient to protect my interests.
>> Err .. when you say protection on one hand and on the other you state
>> it's hard to keep free loaders away, 
> 
> I didn't say that.
> 
> IMHO it isn't even useful to try to keep free loaders away,
> it's better to try and integrate them gradually. That's part
> of the game.
> (Where "free loaders" is a term introduced by Alexandre, not by me.)
> 
> The GPLv2 is a sufficient tool to defend free software
> against those that don't even grasp tit-for-tat. But if
> they do, you can talk to them *as peers* and try to convince
> them that there's more to free software than just tit-for-tat.

What i _feel_ is that "some" (vendors) think that even if they utilize
existing resources what is open and keep what they have done, completely
closed (ie, they use existing infrastructure as a building block, but
they keep the stuff completely closed, in many cases the argument with
regards to IP is not even valid, as looking at symbol tables etc we find
some badly copied code from other parts etc, sewn together. The logical
thought what i have is that they do this to avoid a competition in the
short run) -- and they feel that they are doing things in a quite legal way.

> But it has to be their decision, IMHO it's wrong to force them.

Trying to force _anything_ on _anyone_ doesn't achieve anything, other
than just anger.

We at the worst could of course argue that users should avoid going for
that specific product, but i don't know how much that would work. I say
thus, because legally it would not be possible to challenge such vendors
globally as rules and regulations are different with different
governments and or countries. In such a case a tit-for-tat doesn't work
at all.

If all the people were to agree on common aspects, there wouldn't be any
wars at all ?

> The GPLv3 tries to be a tool to defend against those that
> don't subscribe to the full Free Software Definition.

If v3 defends the definition better, that would be a better use case,
don't you think so ? (But i don't see how v3 also will defend the
definition across international territories, in such a case what i
outlined, since have been bitten by this many times, even talking to the
black sheep which never helped)

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 19, 2007, Johannes Stezenbach <js@linuxtv.org> wrote:

> (Where "free loaders" is a term introduced by Alexandre, not by me.)

It's actually from game theory.  Or something sufficiently mangled by
translation back and forth between English and Portuguese.  I think
the original is actually free riders.  My bad.

> The GPLv2 is a sufficient tool to defend free software
> against those that don't even grasp tit-for-tat.

It's not.  It doesn't even demand tit-for-tat.  This is a
*consequence* of the player understanding the spirit.  You can't
convince a person who doesn't believe in these ideas by pointing at
the license and saying "see, look, you're going to get contributions
back", because there are no such guarantees in the license.  And
that's how it should be.

> But it has to be their decision, IMHO it's wrong to force them.

Agreed.  Nobody is forcing anyone to use GPLed software.  Nobody is
forcing anyone to accept existing software under the GPLv3.

What every GPL stands to do is to defend the freedoms of users,
respecting the wishes of the authors expressed through the GPL that
the free software mains free.  Of course, authors who use the GPL for
other purposes may differ, but whether or not their claims that the
GPL advances their stances bear any resemblance with reality is
irrelevant.  This doesn't change what the GPL stands to do in any way,
it only changes what they expect the GPL to accomplish.  And whether
they're right or wrong in their expectations is besides the point.

> The GPLv3 tries to be a tool to defend against those that
> don't subscribe to the full Free Software Definition.

Not quite.  There appears to be an occurrence of a very common mistake
in your message.  Please forgive the long digression to try to dispell
it.

Many people think that the GPL is what the Free Software Definition is
all about, that Free Software somehow implies GPL, or some other
misunderstandings (to be read without a condescending tone ;-)

Free Software, and in particular the Free Software Definition, talks
about *respecting* users' freedoms.  There are many Free Software
licenses that accomplish this.

Some are very liberal in this sense.  They let you do whatever you
want with the code.  Even use it to disrespect users' freedoms.

Others are not so liberal.  They go *beyond* the Free Software
Definition, i.e., beyond merely respecting users' freedoms.  They
require agreement from recipients to not disrespect some users'
freedoms with the software, in some specific ways.  In addition to
*respecting* some freedoms, they *defend* some freedoms.

Others take the stance of defending all the four users' freedoms,
requiring agreement to not disrespect any users' freedoms with the
software, and to pass on this requirement, such that that software is
never used to disrespect users' freedoms.  These are called copyleft
licenses.

GPL is just one among all copyleft licenses, that are just some among
all licenses that defend some users' freedoms, that are just some
among all licenses that abide by the Free Software definition.  It
just so happens that it's the most widely-used Free Software license.


Think of it this way (if you understand open source but don't know
much of its history): the Open Source Definition is a rewritten
version of the Free Software Definition, intended to retain the same
meaning, but with a different focus.  That's why, as a general rule,
Free Software licenses are Open Source licenses, and Open Source
licenses are Free Software licenses.  AFAIK there is only one known
exception, and that's the open-source license Reciprocal Public
License, which goes to show that the Open Source definition is not
equivalent to the Free Software definition.

The differences are held to be in the motivations behind each of the
movements.  While Free Software takes the respect for the freedoms as
a moral issue (it's the right thing to do), Open Source takes it as a
pragmatic issue (it's better for everyone).  As luck would have it,
the pragmatic benefits are a consequence of the respect for users'
freedoms.

And more, pragmatists who see value in ensuring that the software
remains open source (which any OSI board member will insist that is
far more than merely keeping the source code open) are perceiving a
consequence of ensuring that the software remains Free, of defending
the freedoms of all users of the software.


Sure, there are other pragmatists that don't even care about the
software remaining open source, but only about the source code
remaining available.  We still have a lot in common with them, and we
can happily work together in projects under a number of licenses in
which our goals overlap.

It's not like everyone needs to move to GPLv3.  It's that there would
be moral and practical benefits for everyone if everyone did.  But if
those who don't want to don't, nothing is really lost.  It's just that
both Free Software and Open Source Software advocates who care not
only about abiding by their definitions, but also about defending and
advancing their goals (i.e., not willing to see their software being
used against their goals, and wanting to see more software like that),
don't get an advance in these defenses without the relicensing.

That's unfortunate, but it's not the end of the world.  Nobody can,
should or will force any copyright holder to adopt the GPLv3.  Any
claims to the contrary are emotional reactions to peer pressure, which
very clearly exists, since there are indeed numerous people who want
to advance their goals and would like to use the GPLv3 as a tool.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[bert hubert]

On Sat, Jun 16, 2007 at 08:10:53PM -0700, Linus Torvalds wrote:
> This is where we started. The same way you seem to think that "freedom" 
> has only the meaning *you* and the FSF give it, and that somehow the 
> spirit of the GPL includes the "four freedoms" that aren't even 
> _mentioned_ in it.
> 
> THAT IS NOT TRUE.

Compare for example the landmark case Griswold v. Connecticut where the US
supreme court, faced with a constitution that did not contain items they
desperately wanted it to contain, made up 'penumbras' and 'emanations'
stemming from the bill of rights, which then created the much sought after
guarantees of privacy for Americans.

 '[..] Specific guarantees in the Bill of Rights have penumbras, formed by
 emanations from those guarantees that help give them life and substance. 
 Various guarantees create zones of privacy.'

 and

  'Although the Bill of Rights does not explicitly mention "privacy",
   Justice William O. Douglas (writing for the majority) ruled that the
   right was to be found in the "penumbras" of other constitutional
   protections.'

This was frowned upon back then as 'overly creative', and still is - but
because the constitution could not be changed, and nobody was willing to
amend it, they had no choice but to interpret so extensively.

This is not the case now however.

http://www.law.umkc.edu/faculty/projects/ftrials/conlaw/griswold.html
http://en.wikipedia.org/wiki/Griswold_v._Connecticut

	Bert

-- 
http://www.PowerDNS.com      Open source, database driven DNS Software 
http://netherlabs.nl              Open and Closed source services

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Saturday 16 June 2007 12:57:59 Alexandre Oliva wrote:
> On Jun 16, 2007, Bernd Schmidt <bernds_cb1@t-online.de> wrote:
> > Alexandre Oliva wrote:
> >> On Jun 15, 2007, Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:
> >>> What this means for the FSF goals if Tivo get up one morning and switch
> >>> their system firmware to ROM however is interesting 8)
> >>
> >> I'm not the FSF, and I don't speak for it, but it seems to me that
> >> this would be "mission accomplished".
> >
> > This is insane.  You start with a lofty ideal involving "freedom", and
> > when you end up with a meaningless technicality (and in technical terms
> > a change for the worse) you consider it a victory?
>
> It accomplishes the mission in that everyone is on the same grounds.
> Same freedom for everyone.  If the vendor tries to keep a privilege
> over the software to itself, denying it to its customers, it's failing
> to comply with the spirit of the license.  It's really this simple.
> Is this so hard to understand?

-ELOGIC

They are not keeping a priviledge over the *SOFTWARE* at all. They are keeping 
a priviledge over the *HARDWARE*. But, of course, you've already proven to 
everyone here that you are unwilling and/or unable to understand that.

> The goal is not to push vendors away from GPLed software.  If they
> can't permit modification of the software, that's fine, they can still
> accomplish this.

replacement != modification

If you can't understand that simple fact, then its pointless to continue this 
discussion.

DRH

> What they can't do is deny it to customers while they retain it to
> themselves.  This is unfair, this is wrong, and this disrespects
> users' freedoms.  Therefore, the GPL should not permit it.



-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 16, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> They are not keeping a priviledge over the *SOFTWARE* at all. They
> are keeping a priviledge over the *HARDWARE*.

No, they're using the hardware (along with other pieces of software)
to deny users (but not themselves) the freedoms that the license of
software *meant* to defend, for that software, even if some believe it
doesn't actually defend them.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bernd Schmidt]

Alexandre Oliva wrote:
> On Jun 16, 2007, Bernd Schmidt <bernds_cb1@t-online.de> wrote:
> 
>> Alexandre Oliva wrote:
>>> On Jun 15, 2007, Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:
>>>
>>>> What this means for the FSF goals if Tivo get up one morning and switch
>>>> their system firmware to ROM however is interesting 8)
>>> I'm not the FSF, and I don't speak for it, but it seems to me that
>>> this would be "mission accomplished".
> 
>> This is insane.  You start with a lofty ideal involving "freedom", and
>> when you end up with a meaningless technicality (and in technical terms
>> a change for the worse) you consider it a victory?
> 
> It accomplishes the mission in that everyone is on the same grounds.
> Same freedom for everyone.

See, that's the problem I have with your arguments.  "Same freedom for
everyone" is a political slogan.  It is not a reasoned thought.  "We
must stop terrorists" is also a political slogan, and the consequence
"Tivo should install ROMs so they don't have more rights than users" is
about equivalent as a victory for freedom as disallowing liquids in hand
luggage is a victory against terrorism.  Both are nonsensical
consequences of a political agenda that is applied without thought.


Bernd

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 16, 2007, Bernd Schmidt <bernds_cb1@t-online.de> wrote:

> See, that's the problem I have with your arguments.  "Same freedom for
> everyone" is a political slogan.  It is not a reasoned thought.

Well, this is what got us GPLv2.  And the same reasoning is getting us
GPLv3, and it does get hardware manufacturers to think twice instead
of tivoizing hardware.  They can decide between respecting users'
freedoms and encouraging a community of developers around its product,
or they can decide that not letting users change the software is more
important or necessary, and give up the ability to install
modifications without user approval.  If half of the vendors go each
way, we'll get far more contributions in the end, so we're better off.

This is why I think the argument that anti-tivoization won't get us
more "giving back in kind" is irrational and contradictory.

> "Tivo should install ROMs so they don't have more rights than users"

TiVo doesn't have to install ROMs.  It can use the same technical
measures it uses today, then throw away the keys.

Or give the user half of the signing key, or some such.

How bad would this be for them?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Ingo Molnar <mingo@elte.hu> wrote:

> the GPLv2 license says no such thing, and you seem to be mighty confused 
> about how software licenses work.

> the GPL applies to software. It is a software license.

> the Tivo box is a piece of hardware.

> a disk is put into it with software copied to it already: a bootloader, 
> a Linux kernel plus a handful of applications. The free software bits 
> are available for download.

> the Tivo box is another (copyrighted) work, a piece of hardware.

> so how can, in your opinion, the hardware that Tivo produces, "take 
> away" some right that the user has to the GPL-ed software?

Consider egg yolk and egg shells.

I produce egg yolk.  I give it to you under terms that say "if you
pass this on, you must do so in such a way that doesn't stop anyone
from eating it"


You produce egg shells.  You carefully construct your shell around the
egg yolk and some white you got from a liberal third party.


Then you sell the egg shells, with white and yolk inside, under
contracts that specify "the shell must be kept intact, it can't be
broken or otherwise perforated".


Are you or are you not disrespecting the terms that apply to the yolk?


> by your argument, the user has some "right to modify the software", on 
> that piece of hardware it bought which had free software on it, correct? 

Yes.  This means the hardware distributor who put the software in
there must not place roadblocks that impede the user to get where she
wants with the software, not that the vendor must offer the user a
sport car to take her there.

The goal is not to burden the vendor.  The goal is to stop the vendor
from artificially burdening the user.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Thursday 14 June 2007 22:21:59 Alexandre Oliva wrote:
> On Jun 14, 2007, Ingo Molnar <mingo@elte.hu> wrote:
> > the GPLv2 license says no such thing, and you seem to be mighty confused
> > about how software licenses work.
> >
> > the GPL applies to software. It is a software license.
> >
> > the Tivo box is a piece of hardware.
> >
> > a disk is put into it with software copied to it already: a bootloader,
> > a Linux kernel plus a handful of applications. The free software bits
> > are available for download.
> >
> > the Tivo box is another (copyrighted) work, a piece of hardware.
> >
> > so how can, in your opinion, the hardware that Tivo produces, "take
> > away" some right that the user has to the GPL-ed software?
>
> Consider egg yolk and egg shells.
>
> I produce egg yolk.  I give it to you under terms that say "if you
> pass this on, you must do so in such a way that doesn't stop anyone
> from eating it"
>
>
> You produce egg shells.  You carefully construct your shell around the
> egg yolk and some white you got from a liberal third party.
>
>
> Then you sell the egg shells, with white and yolk inside, under
> contracts that specify "the shell must be kept intact, it can't be
> broken or otherwise perforated".
>
>
> Are you or are you not disrespecting the terms that apply to the yolk?

Bad analogy. I've already provided all the proof needed to prove that, 
while "tivoization" may be against the "intent" or "spirit" of the GPLv2 it 
is not in violation of it.

> > by your argument, the user has some "right to modify the software", on
> > that piece of hardware it bought which had free software on it, correct?
>
> Yes.  This means the hardware distributor who put the software in
> there must not place roadblocks that impede the user to get where she
> wants with the software, not that the vendor must offer the user a
> sport car to take her there.

Okay. That means that if I ship Linux on a ROM chip I have to somehow make it 
so that the person purchasing the chip can modify the copy of Linux installed 
on the chip *if* I want to follow both the spirit and the letter of the 
GPLv2. And no claiming that I'm missing the point - I'm drawing a logical 
conclusion from your statement above.

> The goal is not to burden the vendor.  The goal is to stop the vendor
> from artificially burdening the user.

I have no objection to this. What I object to is the manner in which it is 
being done. However, I must admit that, at this point, I do not know of a 
better method to achieve this goal.

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> On Thursday 14 June 2007 22:21:59 Alexandre Oliva wrote:

>> Consider egg yolk and egg shells.

>> I produce egg yolk.  I give it to you under terms that say "if you
>> pass this on, you must do so in such a way that doesn't stop anyone
>> from eating it"

>> You produce egg shells.  You carefully construct your shell around the
>> egg yolk and some white you got from a liberal third party.

>> Then you sell the egg shells, with white and yolk inside, under
>> contracts that specify "the shell must be kept intact, it can't be
>> broken or otherwise perforated".

>> Are you or are you not disrespecting the terms that apply to the yolk?

> Bad analogy.

It's just a very simple case in which an enclosure is being used to
disrespect the terms of something enclosed in it.

It's meant to show that the argument that "it's a software license, it
can't affect the hardware" is nonsense.

It's not meant to show whether TiVO is right or wrong.  This would
depend on agreement that the GPL requirements are similar to the
requirements of the egg yolk manufacturer.


>> > by your argument, the user has some "right to modify the
>> > software", on that piece of hardware it bought which had free
>> > software on it, correct?

>> Yes.  This means the hardware distributor who put the software in
>> there must not place roadblocks that impede the user to get where she
>> wants with the software, not that the vendor must offer the user a
>> sport car to take her there.

> Okay. That means that if I ship Linux on a ROM chip I have to
> somehow make it so that the person purchasing the chip can modify
> the copy of Linux installed on the chip *if* I want to follow both
> the spirit and the letter of the GPLv2.

I thought we'd already cleared up the issue about ROMs, and why
they're different.  Do I have to quote it again?  Must I allude to
"passing on the rights" every time I mention "imposing further
restrictions"? :-(

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Thursday 14 June 2007 23:54:31 Alexandre Oliva wrote:
> On Jun 14, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> > On Thursday 14 June 2007 22:21:59 Alexandre Oliva wrote:
> >> Consider egg yolk and egg shells.
> >>
> >> I produce egg yolk.  I give it to you under terms that say "if you
> >> pass this on, you must do so in such a way that doesn't stop anyone
> >> from eating it"
> >>
> >> You produce egg shells.  You carefully construct your shell around the
> >> egg yolk and some white you got from a liberal third party.
> >>
> >> Then you sell the egg shells, with white and yolk inside, under
> >> contracts that specify "the shell must be kept intact, it can't be
> >> broken or otherwise perforated".
> >>
> >> Are you or are you not disrespecting the terms that apply to the yolk?
> >
> > Bad analogy.
>
> It's just a very simple case in which an enclosure is being used to
> disrespect the terms of something enclosed in it.
>
> It's meant to show that the argument that "it's a software license, it
> can't affect the hardware" is nonsense.
>
> It's not meant to show whether TiVO is right or wrong.  This would
> depend on agreement that the GPL requirements are similar to the
> requirements of the egg yolk manufacturer.
>
> >> > by your argument, the user has some "right to modify the
> >> > software", on that piece of hardware it bought which had free
> >> > software on it, correct?
> >>
> >> Yes.  This means the hardware distributor who put the software in
> >> there must not place roadblocks that impede the user to get where she
> >> wants with the software, not that the vendor must offer the user a
> >> sport car to take her there.
> >
> > Okay. That means that if I ship Linux on a ROM chip I have to
> > somehow make it so that the person purchasing the chip can modify
> > the copy of Linux installed on the chip *if* I want to follow both
> > the spirit and the letter of the GPLv2.
>
> I thought we'd already cleared up the issue about ROMs, and why
> they're different.  Do I have to quote it again?  Must I allude to
> "passing on the rights" every time I mention "imposing further
> restrictions"? :-(

I wasn't referring to anything that had already been "cleared up". I was 
applying the logic of the statement of yours I quoted. The "cleared up" 
things all were in reference to the GPLv3 - my example was in reference to 
the "spirit" of the GPLv2 that you were stating. By simple extension of the 
logic you provided I came to the conclusion stated above.

The fact that you've claimed I'm wrong shows how flawed your logic is.

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Ingo Molnar]

* Alexandre Oliva <aoliva@redhat.com> wrote:

> > by your argument, the user has some "right to modify the software", 
> > on that piece of hardware it bought which had free software on it, 
> > correct?
> 
> Yes.  This means the hardware distributor who put the software in 
> there must not place roadblocks that impede the user to get where she 
> wants with the software, not that the vendor must offer the user a 
> sport car to take her there.

see the slippery slope in action? Lets just use this limited concession 
on your part and show that _even this_ leads to absurd results:

- a "roadblock" such as a too small button?

- a "roadblock" such as a soldered-on ROM instead of flash-ROM?

- a "roadblock" such as not opening up specifications to the hardware?

- a "roadblock" such as not releasing the source of the BIOS? (and here
  dont come with a "but the BIOS is not under the GPL" strawman
  argument. The Tivo hardware is not under the GPL license either! Your 
  whole argument was that even though the Tivo hardware's design is not 
  GPL-ed, if it runs free software it must not restrict the user's 
  rights and that it must offer the same rights as the hardware 
  manufacturer has. )

- a "roadblock" such as a virtual ROM implemented via an SHA1 key 
  embedded in the hardware?

each of these items limit your supposed "right to modify the software on 
that exact hardware". Each of these items puts a "roadblock" in the way 
and impedes the user to get where she wants with her software. So by 
your argument each of these items would be forbidden. That is 
nonsensical.

	Ingo

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 15, 2007, Ingo Molnar <mingo@elte.hu> wrote:

> * Alexandre Oliva <aoliva@redhat.com> wrote:

>> > by your argument, the user has some "right to modify the software", 
>> > on that piece of hardware it bought which had free software on it, 
>> > correct?

>> Yes.  This means the hardware distributor who put the software in 
>> there must not place roadblocks that impede the user to get where she 
>> wants with the software, not that the vendor must offer the user a 
>> sport car to take her there.

> see the slippery slope in action? Lets just use this limited concession 
> on your part and show that _even this_ leads to absurd results:

> - a "roadblock" such as a too small button?

Why is it too small?

> - a "roadblock" such as a soldered-on ROM instead of flash-ROM?

Why is it soldered-ROM on rather than flash-ROM?

> - a "roadblock" such as not opening up specifications to the hardware?

Why is it not open, and why does that get in the way of replacing the
software?

> - a "roadblock" such as not releasing the source of the BIOS?

Why is it not released, and why does that get in the way of replacing
the software?

> - a "roadblock" such as a virtual ROM implemented via an SHA1 key 
>   embedded in the hardware?

Why is the virtual ROM and the SHA1 key in the hardware?


Remember, the issue is intent.  If you do that for legitimate reasons,
such as technical limitations, industrial economic motives, etc,
you're probably fine.  But if you do that for the purpose of
restraining users' freedoms, then you're going against the intent (and
quite likely the letter) of the license.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Ingo Molnar]

* Alexandre Oliva <aoliva@redhat.com> wrote:

> > see the slippery slope in action? Lets just use this limited 
> > concession on your part and show that _even this_ leads to absurd 
> > results:
> 
> > - a "roadblock" such as a too small button?
> 
> Why is it too small?
> 
> > - a "roadblock" such as a soldered-on ROM instead of flash-ROM?
> 
> Why is it soldered-ROM on rather than flash-ROM?
> 
> > - a "roadblock" such as not opening up specifications to the hardware?
> 
> Why is it not open, and why does that get in the way of replacing the
> software?
> 
> > - a "roadblock" such as not releasing the source of the BIOS?
> 
> Why is it not released, and why does that get in the way of replacing
> the software?
> 
> > - a "roadblock" such as a virtual ROM implemented via an SHA1 key 
> >   embedded in the hardware?
> 
> Why is the virtual ROM and the SHA1 key in the hardware?
> 
> 
> Remember, the issue is intent.  If you do that for legitimate reasons, 
> such as technical limitations, industrial economic motives, etc, 
> you're probably fine.  But if you do that for the purpose of 
> restraining users' freedoms, then you're going against the intent (and 
> quite likely the letter) of the license.

Tivo does it for fully legitimate reasons as well: the only way it can 
be in the PVR business (and the only way it can employ and pay free 
software developers) is if it given a license to certain content. Those 
same users you are trying to "protect" are demanding this content! One 
condition of that content license is that the Tivo protects the 
downloaded content (such as pay-per-view movies). That same content, i'm 
sad to say, the same users who you are trying to "protect", would very 
much like to watch in a pay-per-view fashion, just without the 'pay' 
bit. I dont agree with content policies like that, but your demonization 
of Tivo is royally misplaced. Tivo has two choices: either it gives 
users the content they want to watch, or it goes out of business. Is 
that legitimate enough of a reason to restrict the hardware?

If you want to make a difference you shouldnt attempt to screw with 
Tivo, they are clearly the _victims_ of the content industry. For 
example you are apparently very capable of sending 'content' to lkml in 
the form of dozens of long emails. How about using that energy for a 
Creative Commons project? How about helping Mugshot become more popular? 
Putting Tivo out of business (or forcing Tivo over to Windows CE) does 
not make this world more free one iota - to the contrary!

> Remember, the issue is intent. [...]

Furthermore, there's no need for your patronizing tone here, and there's 
certainly no need to "remember" me of any issues. I very much know what 
i replied to. Here's the original quote of what you wrote:

> > by your argument, the user has some "right to modify the software", 
> > on that piece of hardware it bought which had free software on it, 
> > correct?
>
> Yes. [...]

your "Yes" was not qualified at all via " Yes, except for restriction 
that are 'well-intentioned' ".

your "Yes" led to clearly absurd results, and now that i've pointed out 
a few specific examples of that absurdity, you, instead of conceding 
that i might have a point or two, are now trying to change your "Yes" 
answer to "Yes, but ...". Shame on you!

furthermore, even going along with this newly found argument of yours, 
your new, refined position leads to absurd results just as much. 
Firstly, who are you to dictate the design of the hardware (which was 
created independently of any GNU code) to behavior that you consider 
"legitimate"? What gives you this false sense of entitlement? Secondly, 
who is going to decide what "legitimate" is. Is the FSF the new Police 
of Morality, which enforces that GNU software is only used on hardware 
that has limitations that the FSF considers "legitimate"?

	Ingo

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Tim Post]

On Fri, 2007-06-15 at 23:29 +0200, Ingo Molnar wrote:

> Tivo has two choices: either it gives 
> users the content they want to watch, or it goes out of business. Is 
> that legitimate enough of a reason to restrict the hardware?

Can I submit that they could just rent the use of their machines? It
seems to me like that would work. People can copy the copy of Linux to
their hearts content, but do not own the machines. Its then up to people
if they feel comfortable with that.

If nothing else, a little label that says "This appliance contains free
software which you can copy, but not modify in place". In other words,
you can't treat our gizmos like you would a computer. Therefore we rent
them. Its up to you if you get one after knowing that.

> If you want to make a difference you shouldnt attempt to screw with 
> Tivo, they are clearly the _victims_ of the content industry.

They are now more than ever a house hold word in every country. They
have the public at large championing them. I'd say they were a huge
success.

The situation you describe is very real however, I don't mean to negate
it. They did face undue hardship due to the stupid and unpredictable
industries that they built their business around.

But, they elected to do it. They deserve some sympathy and some kudos
for breaking ground, but I just can't quite see them as a total victim
in all of this. A victim never asked for it in the first place.

Best,
--Tim

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 16, 2007, Tim Post <tim.post@netkinetics.net> wrote:

> On Fri, 2007-06-15 at 23:29 +0200, Ingo Molnar wrote:
>> Tivo has two choices: either it gives 
>> users the content they want to watch, or it goes out of business. Is 
>> that legitimate enough of a reason to restrict the hardware?

> Can I submit that they could just rent the use of their machines?

I don't think this would escape the wording of section 6 in GPLv3dd4:

  [...] User Product is transferred to the recipient in perpetuity or
  for a fixed term (regardless of how the transaction is
  characterized), [...]

and IMHO that's as it should be to defend the freedoms of the user.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Friday 15 June 2007 23:44:00 Alexandre Oliva wrote:
> On Jun 16, 2007, Tim Post <tim.post@netkinetics.net> wrote:
> > On Fri, 2007-06-15 at 23:29 +0200, Ingo Molnar wrote:
> >> Tivo has two choices: either it gives
> >> users the content they want to watch, or it goes out of business. Is
> >> that legitimate enough of a reason to restrict the hardware?
> >
> > Can I submit that they could just rent the use of their machines?
>
> I don't think this would escape the wording of section 6 in GPLv3dd4:
>
>   [...] User Product is transferred to the recipient in perpetuity or
>   for a fixed term (regardless of how the transaction is
>   characterized), [...]
>
> and IMHO that's as it should be to defend the freedoms of the user.

In the case of renting a machine you can try to legislate new laws all you 
want. It doesn't make a difference. There are certain rights you don't get 
when renting something that you do when you own it.

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 16, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> On Friday 15 June 2007 23:44:00 Alexandre Oliva wrote:
>> On Jun 16, 2007, Tim Post <tim.post@netkinetics.net> wrote:
>> > On Fri, 2007-06-15 at 23:29 +0200, Ingo Molnar wrote:
>> >> Tivo has two choices: either it gives
>> >> users the content they want to watch, or it goes out of business. Is
>> >> that legitimate enough of a reason to restrict the hardware?
>> >
>> > Can I submit that they could just rent the use of their machines?
>> 
>> I don't think this would escape the wording of section 6 in GPLv3dd4:
>> 
>> [...] User Product is transferred to the recipient in perpetuity or
>> for a fixed term (regardless of how the transaction is
>> characterized), [...]
>> 
>> and IMHO that's as it should be to defend the freedoms of the user.

> In the case of renting a machine you can try to legislate new laws all you 
> want. It doesn't make a difference. There are certain rights you don't get 
> when renting something that you do when you own it.

You mean renting the computer with the software in it is not
distribution of the software?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Saturday 16 June 2007 04:21:04 Alexandre Oliva wrote:
> On Jun 16, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> > On Friday 15 June 2007 23:44:00 Alexandre Oliva wrote:
> >> On Jun 16, 2007, Tim Post <tim.post@netkinetics.net> wrote:
> >> > On Fri, 2007-06-15 at 23:29 +0200, Ingo Molnar wrote:
> >> >> Tivo has two choices: either it gives
> >> >> users the content they want to watch, or it goes out of business. Is
> >> >> that legitimate enough of a reason to restrict the hardware?
> >> >
> >> > Can I submit that they could just rent the use of their machines?
> >>
> >> I don't think this would escape the wording of section 6 in GPLv3dd4:
> >>
> >> [...] User Product is transferred to the recipient in perpetuity or
> >> for a fixed term (regardless of how the transaction is
> >> characterized), [...]
> >>
> >> and IMHO that's as it should be to defend the freedoms of the user.
> >
> > In the case of renting a machine you can try to legislate new laws all
> > you want. It doesn't make a difference. There are certain rights you
> > don't get when renting something that you do when you own it.
>
> You mean renting the computer with the software in it is not
> distribution of the software?

It is. But you don't have the same rights to a rented machine as you do to one 
you have purchased. In fact, in renting a machine you have to agree to 
a "renters contract" - and that can state *whatever* the person that is 
renting the machine to you feels like having it state. And yes, they can even 
have terms in it that violate the GPL. Not that a "renters contract" ("rental 
agreement" or whatever they call them in your jurisdiction) that has those 
terms can *legally* violate the GPL - but it doesn't stop them from existing.

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Tim Post]

On Sat, 2007-06-16 at 14:43 -0400, Daniel Hazelton wrote:

> >
> > You mean renting the computer with the software in it is not
> > distribution of the software?
> 
> It is. But you don't have the same rights to a rented machine as you do to one 
> you have purchased. In fact, in renting a machine you have to agree to 
> a "renters contract" - and that can state *whatever* the person that is 
> renting the machine to you feels like having it state.

I ended up owning a piece of &*@( I rented from rent-a-center on a trip
because I wiped Windows and installed Debian.

> And yes, they can even 
> have terms in it that violate the GPL. Not that a "renters contract" ("rental 
> agreement" or whatever they call them in your jurisdiction) that has those 
> terms can *legally* violate the GPL - but it doesn't stop them from existing.

I think you could be right. You (in 99% of all cases) agree to not
modify the machine anyway upon accepting it. Since you sign your rights
away at the minimum it would cost big time to enforce the license,
especially against a giant franchise.

That doesn't stop you from just letting people do what they want. If it
were me, I'd let people and that has nothing to do with the spirit of
the GPL, Windows is annoying and I wouldn't force someone to use it.

Probably, I'd just let them pick what they wanted and install it for
them.

I don't think they rent law degrees to go with computers however, you
have to get the blender or juicer to get that special.

Best,
--Tim

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 16, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> On Saturday 16 June 2007 04:21:04 Alexandre Oliva wrote:
>> On Jun 16, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
>> > In the case of renting a machine you can try to legislate new laws all
>> > you want. It doesn't make a difference. There are certain rights you
>> > don't get when renting something that you do when you own it.
>> 
>> You mean renting the computer with the software in it is not
>> distribution of the software?

> It is. But you don't have the same rights to a rented machine as you
> do to one you have purchased.

That's true.  But since it's distribution, the licensing terms of the
software in there must be followed, or the software must be removed.
It's really this simple.

It's not about the hardware.  It's about the software and what you
must not prevent others from doing with it.

> And yes, they can even have terms in it that violate the GPL. Not
> that a "renters contract" ("rental agreement" or whatever they call
> them in your jurisdiction) that has those terms can *legally*
> violate the GPL - but it doesn't stop them from existing.

By "legally violate the GPL", do you mean lawfully escape the terms of
the GPL, or that infringe the copyrights of the authors for violate
its legal terms?  I hope it's the latter.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Saturday 16 June 2007 18:01:59 Alexandre Oliva wrote:
> On Jun 16, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> > On Saturday 16 June 2007 04:21:04 Alexandre Oliva wrote:
> >> On Jun 16, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> >> > In the case of renting a machine you can try to legislate new laws all
> >> > you want. It doesn't make a difference. There are certain rights you
> >> > don't get when renting something that you do when you own it.
> >>
> >> You mean renting the computer with the software in it is not
> >> distribution of the software?
> >
> > It is. But you don't have the same rights to a rented machine as you
> > do to one you have purchased.
>
> That's true.  But since it's distribution, the licensing terms of the
> software in there must be followed, or the software must be removed.
> It's really this simple.
>
> It's not about the hardware.  It's about the software and what you
> must not prevent others from doing with it.
>
> > And yes, they can even have terms in it that violate the GPL. Not
> > that a "renters contract" ("rental agreement" or whatever they call
> > them in your jurisdiction) that has those terms can *legally*
> > violate the GPL - but it doesn't stop them from existing.
>
> By "legally violate the GPL", do you mean lawfully escape the terms of
> the GPL, or that infringe the copyrights of the authors for violate
> its legal terms?  I hope it's the latter.

Sorry, poor choice of words. I meant that they can violate the GPL, because 
they have the right to say "You can't modify the software on the device you 
are renting". I mis-stated it because I didn't make it clear that, even 
though that is their legal right, they would still be in violation of the 
GPL.

The reason that it is different from the TiVO case is that they have not 
stopped you from doing any modification - what they have prevented is the use 
of those modifications on the hardware they designed. The response from the 
FSF and people like you (Alexandre) is childish - at best. "They have one of 
our toys in their house but we can't play with it. WAAAAH!"

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Tim Post]

On Sat, 2007-06-16 at 00:44 -0300, Alexandre Oliva wrote:
> On Jun 16, 2007, Tim Post <tim.post@netkinetics.net> wrote:
> 
> > On Fri, 2007-06-15 at 23:29 +0200, Ingo Molnar wrote:
> >> Tivo has two choices: either it gives 
> >> users the content they want to watch, or it goes out of business. Is 
> >> that legitimate enough of a reason to restrict the hardware?
> 
> > Can I submit that they could just rent the use of their machines?
> 
> I don't think this would escape the wording of section 6 in GPLv3dd4:
> 
>   [...] User Product is transferred to the recipient in perpetuity or
>   for a fixed term (regardless of how the transaction is
>   characterized), [...]
> 
> and IMHO that's as it should be to defend the freedoms of the user.
> 

Yes, I think you're right. There may be no good solution for tivo.

I'm not yet ready to give up on middle ground! :) I'll just have to work
harder if I'm to think of it. I refuse to accept a situation where the
only good outcome results in people being hurt, one way or another.

You might see that as futility, it could very well be. But I feel
obligated to keep looking and thinking because I can.

My head hurts.

Best,
--Tim

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Lennart Sorensen]

On Thu, Jun 14, 2007 at 11:21:59PM -0300, Alexandre Oliva wrote:
> Consider egg yolk and egg shells.
> 
> I produce egg yolk.  I give it to you under terms that say "if you
> pass this on, you must do so in such a way that doesn't stop anyone
> from eating it"
> 
> 
> You produce egg shells.  You carefully construct your shell around the
> egg yolk and some white you got from a liberal third party.
> 
> 
> Then you sell the egg shells, with white and yolk inside, under
> contracts that specify "the shell must be kept intact, it can't be
> broken or otherwise perforated".

It would be more like not telling you how to change the egg yolk while
still having a working egg.  Only the egg shell guy knows how to put a
new egg yolk inside the shell and close the shell around it.  He isn't
going to say you can't break the shell, just that it you break the shell
the egg isn't going to work as a whole egg anymore, and he won't tell
you how to put it back together with a different yolk inside.  You can
still put the yolk inside another container that you do know how to
assemble around the egg stuff.

> Are you or are you not disrespecting the terms that apply to the yolk?

Very bad comparison.  

> Yes.  This means the hardware distributor who put the software in
> there must not place roadblocks that impede the user to get where she
> wants with the software, not that the vendor must offer the user a
> sport car to take her there.

What if I want to run a program that takes 512MB ram and the hardware
guys put in 128MB.  Now they are impeding me doing the change I wanted
to do to the software.

> The goal is not to burden the vendor.  The goal is to stop the vendor
> from artificially burdening the user.

Not putting in an infinite amount of resources is impeding the user too.
"artificially burdening" seems very hard to define.  When something is
hard to define, you are usually better of not trying because you will
get it wrong and screw up even worse as a result.

--
Len Sorensen

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 15, 2007, lsorense@csclub.uwaterloo.ca (Lennart Sorensen) wrote:

> On Thu, Jun 14, 2007 at 11:21:59PM -0300, Alexandre Oliva wrote:
>> Consider egg yolk and egg shells.
>> 
>> I produce egg yolk.  I give it to you under terms that say "if you
>> pass this on, you must do so in such a way that doesn't stop anyone
>> from eating it"
>> 
>> 
>> You produce egg shells.  You carefully construct your shell around the
>> egg yolk and some white you got from a liberal third party.
>> 
>> 
>> Then you sell the egg shells, with white and yolk inside, under
>> contracts that specify "the shell must be kept intact, it can't be
>> broken or otherwise perforated".

> It would be more like not telling you how to change the egg yolk while
> still having a working egg.

This might be more like GPLv3, or it might not.  But it still misses
the point.

The point is to show that the egg yolk license still is about the egg
yolk license, even though its effects do limit what the egg shell
manufacturer can do with that particular egg yolk.

I.e., the GPL is still about the software, and the hardware
manufacturer can't claim "but this is the hardware!" to escape
obligations determined by its choice of the GPL software.

>> Are you or are you not disrespecting the terms that apply to the yolk?

> Very bad comparison.  

Very bad understanding of the intent of the argument, and failing to
answer the relevant question ;-)

> What if I want to run a program that takes 512MB ram and the hardware
> guys put in 128MB.  Now they are impeding me doing the change I wanted
> to do to the software.

*Why* did they make this decision?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bron Gondwana]

On Fri, Jun 15, 2007 at 01:50:04AM +0200, Ingo Molnar wrote:
> the GPL applies to software. It is a software license.
> 
> the Tivo box is a piece of hardware.
> 
> a disk is put into it with software copied to it already: a bootloader, 
> a Linux kernel plus a handful of applications. The free software bits 
> are available for download.

#define Dell CFG_FAVOURITE_VENDOR

A Dell desktop machine is a piece of hardware.  The manufacturer has the
source code (hypothetically) to the BIOS.  The BIOS is required for the
machine to boot and run Linux.

Riddle me this (especially Alexandre, I'm just latching on to Ingo's
post because it has the right hook to grab) - are Dell required to give
out the source to the bios to enable people to have the same rights Dell
engineers do to modify the behaviour of the system?

Bron.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 15, 2007, Bron Gondwana <brong@fastmail.fm> wrote:

> #define Dell CFG_FAVOURITE_VENDOR

> A Dell desktop machine is a piece of hardware.  The manufacturer has the
> source code (hypothetically) to the BIOS.  The BIOS is required for the
> machine to boot and run Linux.

> Riddle me this (especially Alexandre, I'm just latching on to Ingo's
> post because it has the right hook to grab) - are Dell required to give
> out the source to the bios to enable people to have the same rights Dell
> engineers do to modify the behaviour of the system?

What is the license for the bios?  Does it say anything about 'no
further restrictions on the freedoms to modify and share the
software'?

Does it include any mechanisms to stop people from booting modified
versions of the Linux that ships with the machine?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Friday 15 June 2007 01:38:41 Alexandre Oliva wrote:
> On Jun 15, 2007, Bron Gondwana <brong@fastmail.fm> wrote:
> > #define Dell CFG_FAVOURITE_VENDOR
> >
> > A Dell desktop machine is a piece of hardware.  The manufacturer has the
> > source code (hypothetically) to the BIOS.  The BIOS is required for the
> > machine to boot and run Linux.
> >
> > Riddle me this (especially Alexandre, I'm just latching on to Ingo's
> > post because it has the right hook to grab) - are Dell required to give
> > out the source to the bios to enable people to have the same rights Dell
> > engineers do to modify the behaviour of the system?
>
> What is the license for the bios?  Does it say anything about 'no
> further restrictions on the freedoms to modify and share the
> software'?
>
> Does it include any mechanisms to stop people from booting modified
> versions of the Linux that ships with the machine?

Does it matter? The hardware is running a GPL'd project. You have repeatedly 
stated that if a system runs a GPL'd system then all rights to the system 
that the manufacturer has *must* be passed on to the end-user. The 
manufacturer can change the Bios. That's a right they have. Do they have to 
pass that on to the end-user.

Before you answer - this question is *NOT* based on any interpretation or 
reading of the GPLv3. What it is based on is statements you have repeatedly 
made. So no claims this being already covered, and no claims that this isn't 
a situation covered by the GPLv3.

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 15, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> You have repeatedly stated that if a system runs a GPL'd system then
> all rights to the system that the manufacturer has *must* be passed
> on to the end-user.

Not really, not to the entire system.  The spirit is not clear in this
regard, when it talks about "all rights", but I understand it means
"all rights related with the program", i.e., "you must let others do
with the program everything that you can".

> Before you answer - this question is *NOT* based on any interpretation or 
> reading of the GPLv3. What it is based on is statements you have repeatedly 
> made. So no claims this being already covered, and no claims that this isn't 
> a situation covered by the GPLv3.

Sorry that I have been unclear.  This just goes to show that what we
write isn't always the whole story, and quite often intent doesn't
shine through the words.  While legal terms have a stronger demand for
clarity and non-ambiguity, intent and other less-formal forms of
communication often depend on a lot of context for correct
interpretation.  And then, if multiple interpretations are possible,
the only resort is to ask the author and hope s/he still remembers
what s/he meant.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bron Gondwana]

On Fri, Jun 15, 2007 at 02:38:41AM -0300, Alexandre Oliva wrote:
> On Jun 15, 2007, Bron Gondwana <brong@fastmail.fm> wrote:
> 
> > #define Dell CFG_FAVOURITE_VENDOR
> 
> > A Dell desktop machine is a piece of hardware.  The manufacturer has the
> > source code (hypothetically) to the BIOS.  The BIOS is required for the
> > machine to boot and run Linux.
> 
> > Riddle me this (especially Alexandre, I'm just latching on to Ingo's
> > post because it has the right hook to grab) - are Dell required to give
> > out the source to the bios to enable people to have the same rights Dell
> > engineers do to modify the behaviour of the system?
> 
> What is the license for the bios?  Does it say anything about 'no
> further restrictions on the freedoms to modify and share the
> software'?

It's a necessary part of the boot process, without which Linux could
not be started.  Indeed, the Linux kernel interacts with it through a
(loosely, incompletely and frequently buggy) documented interface, much
like how binary modules interact with the linux kernel (even if they do
get loaded into the sacred ring0 execution space, ooh err)

What happens if you're debugging something you think is a bug in the
Linux kernel and then you run bang into some interactions that make you
think the bug might be in the BIOS instead.  Oh unhappy day, you don't
have access to the source code to the BIOS so you can't check.  Those
cretins at Dell (does a #define still work when it's 2 levels quoted?)
have denied your freedom to modify and debug the system they sold you
which is based _in_a_large_part_ on the GPL$mumble Linux kernel and
hence needs to be interoperable.

Regardless of your sophistry, it's a slipery slope by which Dell could
be forced to exert their corporate might back up the tree to the BIOS
vendor and get the right to release that BIOS source code to you or
stop distributing Linux on their machines. 

> Does it include any mechanisms to stop people from booting modified
> versions of the Linux that ships with the machine?

Maybe, and either way, a future update could, and you couldn't undo it
unless the BIOS flash system lets you "downgrade" again.

Bron.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 15, 2007, Bron Gondwana <brong@fastmail.fm> wrote:

> What happens if you're debugging something you think is a bug in the
> Linux kernel and then you run bang into some interactions that make you
> think the bug might be in the BIOS instead.

> have denied your freedom to modify and debug the system they sold you

If the bug is in the non-GPLed BIOS, not in the GPLed code, too bad.
One more reason to dislike non-Free Software.

The freedom the GPL defends is not the freedom to modify and debug the
system, but rather the covered software.

Now, if you find evidence that the "bug" is actually intentionally put
there to stop you from doing what you wanted with the software, then
there's clearly a violation of the spirit of the license, and you
might even have a case of copyright infringement, but IANAL.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bron Gondwana]

On Fri, Jun 15, 2007 at 04:26:34PM -0300, Alexandre Oliva wrote:
> On Jun 15, 2007, Bron Gondwana <brong@fastmail.fm> wrote:
> 
> > What happens if you're debugging something you think is a bug in the
> > Linux kernel and then you run bang into some interactions that make you
> > think the bug might be in the BIOS instead.
> 
> > have denied your freedom to modify and debug the system they sold you
> 
> If the bug is in the non-GPLed BIOS, not in the GPLed code, too bad.
> One more reason to dislike non-Free Software.

If the bug is in the non-GPLed binary module, not in the GPLed code, too
bad.  One more reason to dislike non-Free Software.

It's the same argument from the other direction.  The BIOS is linked
(inside the machine, sure) to the kernel for all intents and purposes
through a defined interface.  This doesn't affect the BIOS developers
who ship me a machine on to which I then install Linux, but it _does_
affect a hardware vendor who ships me a system with Linux pre-installed,
because it could easily be argued that they linked the BIOS with the
Linux kernel and hence produced a combined work (remember, they reserve
the right to modify the BIOS, but don't give that that right to me) and
the BIOS should now come under the GPL.

Talk about your chilling effects.  It's a strong reason for vendors not
to ship GPL3 or GPL2[your interpretation] code pre-installed while the
legal boundaries of work combination are in any way grey.

Regards,

Bron.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Friday 15 June 2007 22:16:30 Bron Gondwana wrote:
> On Fri, Jun 15, 2007 at 04:26:34PM -0300, Alexandre Oliva wrote:
> > On Jun 15, 2007, Bron Gondwana <brong@fastmail.fm> wrote:
> > > What happens if you're debugging something you think is a bug in the
> > > Linux kernel and then you run bang into some interactions that make you
> > > think the bug might be in the BIOS instead.
> > >
> > > have denied your freedom to modify and debug the system they sold you
> >
> > If the bug is in the non-GPLed BIOS, not in the GPLed code, too bad.
> > One more reason to dislike non-Free Software.
>
> If the bug is in the non-GPLed binary module, not in the GPLed code, too
> bad.  One more reason to dislike non-Free Software.
>
> It's the same argument from the other direction.  The BIOS is linked
> (inside the machine, sure) to the kernel for all intents and purposes
> through a defined interface.  This doesn't affect the BIOS developers
> who ship me a machine on to which I then install Linux, but it _does_
> affect a hardware vendor who ships me a system with Linux pre-installed,
> because it could easily be argued that they linked the BIOS with the
> Linux kernel and hence produced a combined work (remember, they reserve
> the right to modify the BIOS, but don't give that that right to me) and
> the BIOS should now come under the GPL.

This is the point that was being made, IMHO. But thank you for explaining it 
for the people that either could not understand, or will not willing to 
understand, the original point that was made.

DRH

>
> Talk about your chilling effects.  It's a strong reason for vendors not
> to ship GPL3 or GPL2[your interpretation] code pre-installed while the
> legal boundaries of work combination are in any way grey.
>
> Regards,
>
> Bron.



-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 15, 2007, Bron Gondwana <brong@fastmail.fm> wrote:

> because it could easily be argued that they linked the BIOS with the
> Linux kernel

How so?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bron Gondwana]

On Sat, Jun 16, 2007 at 05:22:21AM -0300, Alexandre Oliva wrote:
> On Jun 15, 2007, Bron Gondwana <brong@fastmail.fm> wrote:
> 
> > because it could easily be argued that they linked the BIOS with the
> > Linux kernel
> 
> How so?

(I'm going to refer to Linux as GPLix from here on since this argument
 is more general than a specific GPLed operating system)

Er, they installed it in the same piece of equipment, and the kernel
couldn't function without it in that work.  What's more 'linked' than
that.  It's a vital part of the boot process on that piece of hardware
in exactly the same way that the public-key check is a vital part of
the boot process.

If your printer^wPC isn't doing what you want and you know how to
change it to do what you want but it needs a BIOS patch.  Guess what,
you can't do it - your vendor can.  By using GPLix as part of their
boot process along with their non-GPL BIOS, they're subverting the
freedoms that the user should have in being able to control the entire
boot process.

Right?  Or are you unclear about the fact that there's a big grey area
cutting through this part of usage, and Linus sat down pretty clearly on
one side of it while you're arguing that the goalposts should be "where
I feel that my rights to make changes are being infringed".

While the vendor reserves the ability to change components of the
system (post sale, i.e. a BIOS flash update) and doesn't hand those
same rights on to you, they have partially Tivoised (hoover, kleenex,
you've got nothing on these guys for having your name associated with
a concept) the hardware.  By logical extention of your arguments over
the past few days, this denies them the ability to use any GPLed
software in 'the spirit of the licence' anyware on the machine because
they are denying you rights regarding the instance of the product
they shipped to you that they are retaining for themselves.  The very
freedoms you so vocally claim.

Now, the position I'm seeing here is that the above behaviour (every
single hardware manufacturer that has ever shipped a machine with
pre-installed Linux) violates the spirit of the GPL by the "retaining
exclusive freedoms to modify shipped product" rule, and hence their
BIOS is in the doghouse unless they either:

a) offer full source code access and rights per the holy spirit ghost
   of the GPL; or

b) deny themselves the ability to every offer a patch to said BIOS if
   bugs are found 


Point (b) is also exactly on topic for the discussion of enforcing
legal safety obligations in hardware on a peripheral rather than the
software drivers.  It's requiring that these limitations be placed in
a technically inferior location to hack around a legal "bug".  (A bug
is in the eye of the beholder, please wear glasses while cycling, it's
your own responsibility to protect your eyes)


Er, I think I'm done.  Yes.


Executive summary:

a) by not providing the BIOS source code but retaining the right to 
   change the BIOS the vendor is linking the GPLix kernel and the 
   BIOS (you can't run the kernel without it)

b) legislating intent is fraught.

c) by your arguments, (a) is violating the spirit and (b) is necessary
   to get around that.

Bron.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 16, 2007, Bron Gondwana <brong@fastmail.fm> wrote:

> On Sat, Jun 16, 2007 at 05:22:21AM -0300, Alexandre Oliva wrote:
>> On Jun 15, 2007, Bron Gondwana <brong@fastmail.fm> wrote:
>> 
>> > because it could easily be argued that they linked the BIOS with the
>> > Linux kernel
>> 
>> How so?

> Er, they installed it in the same piece of equipment, and the kernel
> couldn't function without it in that work.

I see what you're getting at.  You're thinking of a license that
doesn't respect the idea of "mere aggregation", right?

For starters, this wouldn't evidently not qualify as an Open Source
license, and I'm pretty sure it wouldn't qualify as a Free Software
license either.

> By using GPLix as part of their boot process along with their
> non-GPL BIOS, they're subverting the freedoms that the user should
> have in being able to control the entire boot process.

You're pushing the "freedom to change" too far.  Sure, I'd like to be
able to do that, and I prefer hardware that lets me do it, but it's
not like this BIOS in the scenario you described is being used as a
means to stop me from modifying the GPLed software.

I have never said that including a GPLed piece of software should
grant users the right to modify anything whatsoever in the system, or
grant them control over the entire system.  Others have, but it's not
true, it just shows how much mis-information is floating around.

All the GPL stands for is to defend the freedom of the users over the
particular program it applies to.  You can't impose further
restrictions on the user's ability to modify what *that* software
does.

If you wanted to change something else, but this something else is not
covered by the license, and is not being used to contradict the terms
of the license, well, too bad, you lose.

> b) deny themselves the ability to every offer a patch to said BIOS if
>    bugs are found 

> Point (b) is also exactly on topic for the discussion of enforcing
> legal safety obligations in hardware on a peripheral rather than the
> software drivers.

> It's requiring that these limitations be placed in a technically
> inferior location to hack around a legal "bug".

I don't think this last sentence is true.  If you implement hardware
locks that prevent modification of the software even by yourself, then
you're in compliance with the terms of the GPLv3dd4.  But IANAL.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Saturday 16 June 2007 13:14:29 Alexandre Oliva wrote:
> On Jun 16, 2007, Bron Gondwana <brong@fastmail.fm> wrote:
> > On Sat, Jun 16, 2007 at 05:22:21AM -0300, Alexandre Oliva wrote:
> >> On Jun 15, 2007, Bron Gondwana <brong@fastmail.fm> wrote:
> >> > because it could easily be argued that they linked the BIOS with the
> >> > Linux kernel
> >>
> >> How so?
> >
> > Er, they installed it in the same piece of equipment, and the kernel
> > couldn't function without it in that work.
>
> I see what you're getting at.  You're thinking of a license that
> doesn't respect the idea of "mere aggregation", right?
>
> For starters, this wouldn't evidently not qualify as an Open Source
> license, and I'm pretty sure it wouldn't qualify as a Free Software
> license either.

This situation is a general description that actually fits what TiVO has done. 
The difference in the TiVO case is that you (and everyone that thinks like 
you - ie: believes that the "tivoization" language in GPLv3 is good) 
equate "replace entirely" with "modification" when, in fact, the two are 
entirely separate acts.

> > By using GPLix as part of their boot process along with their
> > non-GPL BIOS, they're subverting the freedoms that the user should
> > have in being able to control the entire boot process.
>
> You're pushing the "freedom to change" too far.  Sure, I'd like to be
> able to do that, and I prefer hardware that lets me do it, but it's
> not like this BIOS in the scenario you described is being used as a
> means to stop me from modifying the GPLed software.
>
> I have never said that including a GPLed piece of software should
> grant users the right to modify anything whatsoever in the system, or
> grant them control over the entire system.  Others have, but it's not
> true, it just shows how much mis-information is floating around.
>
> All the GPL stands for is to defend the freedom of the users over the
> particular program it applies to.  You can't impose further
> restrictions on the user's ability to modify what *that* software
> does.

"You can't impose further restrictions on the user's ability to modify what 
*that* software does."

I don't see how TiVO has done this. They have placed no restrictions on 
*modification* at all. What they have done is placed a restriction on 
*REPLACEMENT* of the program. If you're going to argue that "replacement == 
modification" then it is an *easy* argument to make that every time someone 
*replaces* linux with a proprietary system the proprietary system magically 
becomes GPL'd.

And no, this isn't a logical fallacy on my part. It's on your part - all I've 
done is take the logic you have provided and extend it to cover a different 
situation.

DRH

> If you wanted to change something else, but this something else is not
> covered by the license, and is not being used to contradict the terms
> of the license, well, too bad, you lose.
>
> > b) deny themselves the ability to every offer a patch to said BIOS if
> >    bugs are found
> >
> > Point (b) is also exactly on topic for the discussion of enforcing
> > legal safety obligations in hardware on a peripheral rather than the
> > software drivers.
> >
> > It's requiring that these limitations be placed in a technically
> > inferior location to hack around a legal "bug".
>
> I don't think this last sentence is true.  If you implement hardware
> locks that prevent modification of the software even by yourself, then
> you're in compliance with the terms of the GPLv3dd4.  But IANAL.



-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 16, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> I don't see how TiVO has done this. They have placed no restrictions on 
> *modification* at all. What they have done is placed a restriction on 
> *REPLACEMENT* of the program.

Technicality.  In order for the software to remain free (which is what
the GPL is all about), the user must not be stopped from adapting the
software to suit his needs and running it for any purpose.  TiVo
places restrictions on it.  It's really this simple.

And then, TiVo doesn't really prohibit replacement.  You can replace
it as much as you like; just not as conveniently as TiVo can replace
it.  And then, if you do, it won't run, because it's not signed with a
key that they omit from the source code.  And they do this in order to
prevent the user from changing the behavior of the Free Software that
they use, while they keep this ability to themselves.

If these are not restrictions on the freedoms that the GPL is designed
to protect to ensure that Free Software remains Free for all its
users, I don't know what is.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

RE: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Schwartz]

> On Jun 16, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> > I don't see how TiVO has done this. They have placed no restrictions on
> > *modification* at all. What they have done is placed a restriction on
> > *REPLACEMENT* of the program.

> Technicality.  In order for the software to remain free (which is what
> the GPL is all about), the user must not be stopped from adapting the
> software to suit his needs and running it for any purpose.  TiVo
> places restrictions on it.  It's really this simple.

No, this is completely and utterly wrong. By this logic, Linux isn't free if
I can't run it on *YOUR* laptop. TiVo places restrictions on *hardware*. The
hardware is not free.

> And then, TiVo doesn't really prohibit replacement.  You can replace
> it as much as you like; just not as conveniently as TiVo can replace
> it.  And then, if you do, it won't run, because it's not signed with a
> key that they omit from the source code.  And they do this in order to
> prevent the user from changing the behavior of the Free Software that
> they use, while they keep this ability to themselves.

> If these are not restrictions on the freedoms that the GPL is designed
> to protect to ensure that Free Software remains Free for all its
> users, I don't know what is.

So why is it not a restriction on this freedom that I can't modify the copy
of Linux running on *your* desktop? If it helps you to understand the
situation better, think of TiVo as not really selling you the hardware.

To see why this isn't a GPL issue, imagine if TiVo explicitly didn't sell
the hardware. Imagine if they only rented it or sold it but retained the
right to control what software ran on it. Essentially, your TiVo would be
like my laptop -- you don't get to decide what software runs on it. But if
you get GPL'd software, you get source code.

Regardless of how the GPL came to be in the first place, the vast majority
of people who chose to use the GPL (including Linus himself) choose it so
that the code can't be modified and distributed and those modifications kept
secret. The idea is that any change widely distributed in binary form is
nearly assured to propogate back in source code form, and is assured to get
to those who paid for the binary.

Linus, and many other people, don't give a damn (from a GPL perspective)
about what TiVo does with their hardware. They may agree with it, disagree
with it, think it's legal, maybe even illegal, but they don't think it has
*anything* to do with the intent or spirit of the GPLv2 as *they* understand
it and for the reasons *they* chose it. They just want to get source code,
and they really don't care what other people do with it -- they care about
what *they* can do with it.

They just want the source code, and TiVo gives it to them. GPL was about
source code not being secret, to them and to many others.

> No, they're using the hardware (along with other pieces of software)
> to deny users (but not themselves) the freedoms that the license of
> software *meant* to defend, for that software, even if some believe it
> doesn't actually defend them.

At least to Linus, the GPL was never meant to defend the freedom to run
Linux on any hardware you want. It was just meant to ensure that you
couldn't keep the source code secret. I personally feel precisely the same
way and I think many other people do too.

I think that what TiVo is doing is wrong for completely different reasons
that have nothing to do with the fact that it happens to run Linux or that
Linux happens to be free software. But I think I've already made that clear
in other posts.

DS

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 16, 2007, "David Schwartz" <davids@webmaster.com> wrote:

>> On Jun 16, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

>> > I don't see how TiVO has done this. They have placed no restrictions on
>> > *modification* at all. What they have done is placed a restriction on
>> > *REPLACEMENT* of the program.

>> Technicality.  In order for the software to remain free (which is what
>> the GPL is all about), the user must not be stopped from adapting the
>> software to suit his needs and running it for any purpose.  TiVo
>> places restrictions on it.  It's really this simple.

> No, this is completely and utterly wrong. By this logic, Linux isn't free if
> I can't run it on *YOUR* laptop. TiVo places restrictions on *hardware*. The
> hardware is not free.

TiVo uses the hardware to stop the user from adapting the software to
suit his/her needs.  TiVo is imposing an artificial restriction on
what you can do with the software you use.

You don't use the software in my laptop.  The laptop is not yours.
You have no claims whatsoever about it.

The GPL is not about letting you do whatever you want.  It's about
ensuring every licensees respect others' freedoms, rather than
imposing artificial additional restrictions on the exercise of the
freedoms.

>> If these are not restrictions on the freedoms that the GPL is designed
>> to protect to ensure that Free Software remains Free for all its
>> users, I don't know what is.

> So why is it not a restriction on this freedom that I can't modify the copy
> of Linux running on *your* desktop?

If I gave, rented or sold the desktop to you, then I should respect
your freedom to do so.

I have no obligation to grant you access to my desktop.  If you're not
a user of this computer or of the software installed in it.

> If it helps you to understand the situation better, think of TiVo as
> not really selling you the hardware.

I see what you're getting at.  This might be relevant.  If I granted
you remote access to my desktop, I probably wouldn't want to grant you
permission to install and boot whatever kernel fancies you.

The difference is that, when I grant you remote access to my desktop,
I'm not distributing the software to you.  But when TiVo places its
DVR in your home, it is.

And then, again, there's the issue of motivation, the intent.  Why am
I not granting you permission to reboot my computer into a different
kernel?  Would you think my motivations are similar to TiVo's?  That
I'm doing this for the purpose of denying you the freedom to adapt the
software to your own needs?


> They just want the source code, and TiVo gives it to them. GPL was about
> source code not being secret, to them and to many others.

They chose the GPL because it worked this way for them.  But this is
not what the GPL is *all* about.  And GPLv3 shows the difference.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

RE: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Schwartz]

> On Jun 16, 2007, "David Schwartz" <davids@webmaster.com> wrote:

> > No, this is completely and utterly wrong. By this logic, Linux
> > isn't free if
> > I can't run it on *YOUR* laptop. TiVo places restrictions on
> > *hardware*. The
> > hardware is not free.

> TiVo uses the hardware to stop the user from adapting the software to
> suit his/her needs.  TiVo is imposing an artificial restriction on
> what you can do with the software you use.

Sure, and you use the hardware to stop me from modifying the Linux on your
laptop. You are imposing an artificial restriction on what I can do with
Linux.

If the restriction is in the source code of the program, I can remove it. If
it's not, it's outside the scope of the GPL.

> You don't use the software in my laptop.  The laptop is not yours.
> You have no claims whatsoever about it.

Exactly. And I have no *GPL* claims to my laptop either. The GPL doesn't
talk about who owns what hardware and it would be insane for it to do so.
Even though the TiVo hardware is yours, you have no more *GPL* claims to it
than you do to someone else's laptop. The GPL does not talk about who owns
what hardware.

The GPL (at least through version 2) is about free access to source code.

> The GPL is not about letting you do whatever you want.  It's about
> ensuring every licensees respect others' freedoms, rather than
> imposing artificial additional restrictions on the exercise of the
> freedoms.

Right, and those freedoms include getting the source code if you get the
object code. They include being able to import the source code into other
projects with compatible licenses. They include being able to modify the
source code however you like.

They just do not include being able to use the source code on whatever
hardware you want because that hardware could be restricted for any number
of reasons. One of them could be that it's not yours. Another of them could
be that the platform itself has restrictions.

> >> If these are not restrictions on the freedoms that the GPL is designed
> >> to protect to ensure that Free Software remains Free for all its
> >> users, I don't know what is.

> > So why is it not a restriction on this freedom that I can't
> > modify the copy
> > of Linux running on *your* desktop?

> If I gave, rented or sold the desktop to you, then I should respect
> your freedom to do so.

You are missing the point. Whether the laptop is mine or yours has no
bearing on the GPL terms. The GPL terms are about what you get when the
object code is distributed to you. To read into the GPL that you get certain
rights if you own hardware that runs GPL code and not if you rent such
hardware is just getting crazy. It's simply making arbitrary things up so
you get the results you want in the cases you care about and don't have to
deal with the crazy results you get in other cases you don't care about. It
makes no logical sense and is purely ad hoc.

> I have no obligation to grant you access to my desktop.  If you're not
> a user of this computer or of the software installed in it.

Right, and TiVo has no GPL obligation to grant you access to their hardware
platform, even if you own a physical implementation of it.

> > If it helps you to understand the situation better, think of TiVo as
> > not really selling you the hardware.

> I see what you're getting at.  This might be relevant.  If I granted
> you remote access to my desktop, I probably wouldn't want to grant you
> permission to install and boot whatever kernel fancies you.

> The difference is that, when I grant you remote access to my desktop,
> I'm not distributing the software to you.  But when TiVo places its
> DVR in your home, it is.

Assume the access includes the right to download copies of the software, in
that case, it is distribution. For GPL purposes, all that matters is whether
the software is distributed or not, and the rights must be the same
regardless of anything else.

> And then, again, there's the issue of motivation, the intent.  Why am
> I not granting you permission to reboot my computer into a different
> kernel?  Would you think my motivations are similar to TiVo's?  That
> I'm doing this for the purpose of denying you the freedom to adapt the
> software to your own needs?

That's all lovely stuff, but it has nothing to do with anything. The GPL
doesn't care what your motivations are. If you can't fulfill your GPL
obligations, no matter how nice your intentions, you can't distribute at
all.

> > They just want the source code, and TiVo gives it to them. GPL was about
> > source code not being secret, to them and to many others.

> They chose the GPL because it worked this way for them.  But this is
> not what the GPL is *all* about.  And GPLv3 shows the difference.

That's what it was about to many people, including Linus. It was about
getting source code.

DS

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 16, 2007, "David Schwartz" <davids@webmaster.com> wrote:

>> On Jun 16, 2007, "David Schwartz" <davids@webmaster.com> wrote:

>> > No, this is completely and utterly wrong. By this logic, Linux
>> > isn't free if
>> > I can't run it on *YOUR* laptop. TiVo places restrictions on
>> > *hardware*. The
>> > hardware is not free.

>> TiVo uses the hardware to stop the user from adapting the software to
>> suit his/her needs.  TiVo is imposing an artificial restriction on
>> what you can do with the software you use.

> Sure, and you use the hardware to stop me from modifying the Linux on your
> laptop.

Do I?  How so?

>> You don't use the software in my laptop.  The laptop is not yours.
>> You have no claims whatsoever about it.

> Exactly. And I have no *GPL* claims to my laptop either. The GPL
> doesn't talk about who owns what hardware and it would be insane for
> it to do so.  Even though the TiVo hardware is yours, you have no
> more *GPL* claims to it than you do to someone else's laptop. The
> GPL does not talk about who owns what hardware.

This is absolutely correct.

What it does is impose conditions for whoever wants to distribute the
software.  And GPLv3 makes it explicit that one such condition is to
permit the user to install and run modified versions of the program in
the hardware that ships with the program.  A condition that is
arguably already encoded in the "no further restrictions to the rights
granted" by the license" and to the requirement for complete
corresponding source code to accompany the binary.

That you disagree with it doesn't make you right.

But that it is within the spirit of the GPL defined by its authors
(which is all I'm trying to show here), it is.

> The GPL (at least through version 2) is about free access to source
> code.

Some think so, but this was GPLv1.

v2 added stuff such as:

  if a patent license would not permit royalty-free redistribution of
  the Program by all those who receive copies directly or indirectly
  through you, then the only way you could satisfy both it and this
  License would be to refrain entirely from distribution of the
  Program

Do you realize that the patent is unrelated with the program, but
nevertheless the copyright license establishes conditions about what
kind of patent licenses you may accept in order for you to have
permission to distribute the program.

Why should restrictions through patents be unacceptable, but
restrictions through hardware and software be acceptable.

Both are means to disrespect users' freedoms.

It is the duty of the FSF to defend these freedoms.  It's its public
mission.  That's a publicly stated goal of the GPL, for anyone who
cares to understand it, or miss it completely and then complain about
changes in spirit.

> They just do not include being able to use the source code on whatever
> hardware you want because that hardware could be restricted for any number
> of reasons.

That's true.  Per the license, it's only who distributes the hardware
to you that shouldn't impose such restrictions.

>> I see what you're getting at.  This might be relevant.  If I granted
>> you remote access to my desktop, I probably wouldn't want to grant you
>> permission to install and boot whatever kernel fancies you.

>> The difference is that, when I grant you remote access to my desktop,
>> I'm not distributing the software to you.  But when TiVo places its
>> DVR in your home, it is.

> Assume the access includes the right to download copies of the software, in
> that case, it is distribution. For GPL purposes, all that matters is whether
> the software is distributed or not, and the rights must be the same
> regardless of anything else.

I'm inclined to agree.

> The GPL doesn't care what your motivations are. If you can't fulfill
> your GPL obligations, no matter how nice your intentions, you can't
> distribute at all.

That's right.  But one of the obligations is to impose no further
restrictions on the exercise of the rights.  What is "imposing a
restriction"?  Installing the software in ROM isn't regarded as such,
it's just a technical decision.  Installing the software in modifiable
non-volatile storage, but denying the user the ability to change it,
is regarded as imposing a restriction.  (note the "denying")  It is a
matter of intent.

It's not because you only install say 32MB of RAM on the machine that
you're denying the user the ability to run OOo on the machine.  But if
you ship the computer with plenty of memory, but somehow configure the
hardware or the operating system so as to prevent the user from
upgrading an OOo that shipped with it, while you can still install
that upgrade, then you're actively placing limits on the user's
freedom WRT to that software, and an anti-tivoization clause would
then stop you from distributing the software under these conditions.

>> > They just want the source code, and TiVo gives it to them. GPL was about
>> > source code not being secret, to them and to many others.

>> They chose the GPL because it worked this way for them.  But this is
>> not what the GPL is *all* about.  And GPLv3 shows the difference.

> That's what it was about to many people, including Linus. It was about
> getting source code.

I've never disputed that this is how they perceive it.

I've never disputed that GPLv2 serves this goal.

I still think GPLv3 serves this same goal, and better than v2.

But this is not what my participation here is about.

My participation here is about showing that GPLv3, and anti-tivozation
in particular, don't violate the spirit of the defending users'
freedoms WRT the covered software, such that the Free Software remains
Free.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

RE: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Schwartz]

> > Sure, and you use the hardware to stop me from modifying the
> > Linux on your
> > laptop.

> Do I?  How so?

Any number of ways. For example, you probably don't connect the serial ports
to a device I have access to.

> >> You don't use the software in my laptop.  The laptop is not yours.
> >> You have no claims whatsoever about it.

> > Exactly. And I have no *GPL* claims to my laptop either. The GPL
> > doesn't talk about who owns what hardware and it would be insane for
> > it to do so.  Even though the TiVo hardware is yours, you have no
> > more *GPL* claims to it than you do to someone else's laptop. The
> > GPL does not talk about who owns what hardware.

> This is absolutely correct.

> What it does is impose conditions for whoever wants to distribute the
> software.  And GPLv3 makes it explicit that one such condition is to
> permit the user to install and run modified versions of the program in
> the hardware that ships with the program.

I'm sorry, who is "the user"? Who exactly is supposed to be able to install
and run modified versions? How does the GPLv3 specify who is supposed to be
authorized to do this?

The TiVo control over updating the software is a specific access control
measure. It says, "X is authorized to replace the software on this machine
but Y is not". Now, somebody has to make that decision. It's clearly chaos
if anyone can change the software on any machine.

How exactly does the GPLv3 specify who should and should not be able to
change the software on a particular physical machine?

> A condition that is
> arguably already encoded in the "no further restrictions to the rights
> granted" by the license" and to the requirement for complete
> corresponding source code to accompany the binary.

Except that the "right" to upload the software on some particular piece of
hardware was *never* a right granted by the GPL, nor could it be. That *HAS*
to be a right granted by whatever authority controls the use of that
hardware. It seems utterly nonsensical to argue otherwise.

> That you disagree with it doesn't make you right.

Anyone can disagree over anything. If I'm not right just because people
disagree with me, then nobody is ever right.

It's totally obvious that who gets to install what software on a given piece
of hardware is determined by the person who creates/owns that hardware and
they have to authorize anyone else to change it.

> But that it is within the spirit of the GPL defined by its authors
> (which is all I'm trying to show here), it is.

It is not. The GPL was never about who was allowed to modify the software on
particular pieces of hardware. It was about the lack of *legal* obstacles to
your doing so. It wasn't about *authorization* obstacles imposed on the
creators/owners of hardware. These are night and day different categories.

> > The GPL (at least through version 2) is about free access to source
> > code.
>
> Some think so, but this was GPLv1.
>
> v2 added stuff such as:
>
>   if a patent license would not permit royalty-free redistribution of
>   the Program by all those who receive copies directly or indirectly
>   through you, then the only way you could satisfy both it and this
>   License would be to refrain entirely from distribution of the
>   Program
>
> Do you realize that the patent is unrelated with the program, but
> nevertheless the copyright license establishes conditions about what
> kind of patent licenses you may accept in order for you to have
> permission to distribute the program.

Right, because this would put legal obstacles to your ability to use the
software, even on hardware you do control.

> Why should restrictions through patents be unacceptable, but
> restrictions through hardware and software be acceptable.

Because the former are legal obstacles and the latter are authorization
obstacles. Certainly I'm not free to run Linux on *YOUR* laptop. That's an
authorization obstacle -- it's *your* laptop. This is TiVo's hardware, and
assuming they don't include the right to control what software runs on the
hardware in the sale, you don't have that authorization right over that
hardware -- just like you don't have it for *my* computer.

> Both are means to disrespect users' freedoms.

The freedom to control what software runs on someone else's hardware?!

> > The GPL doesn't care what your motivations are. If you can't fulfill
> > your GPL obligations, no matter how nice your intentions, you can't
> > distribute at all.

> That's right.  But one of the obligations is to impose no further
> restrictions on the exercise of the rights.  What is "imposing a
> restriction"?  Installing the software in ROM isn't regarded as such,
> it's just a technical decision.  Installing the software in modifiable
> non-volatile storage, but denying the user the ability to change it,
> is regarded as imposing a restriction.  (note the "denying")  It is a
> matter of intent.

I totally disagree. Intent should be irrelevent. If the user cannot access,
modify and run the software (on hardware he controls and absent any
technical limits, of course) then you are violating the license, period. You
may not put *legal* obstacles in his way. However, he can't run the software
on hardware on which it is not compatible, though he is free to make it
compatible. He can't run it on hardware on which he does not have the right
to control the software.

> It's not because you only install say 32MB of RAM on the machine that
> you're denying the user the ability to run OOo on the machine.  But if
> you ship the computer with plenty of memory, but somehow configure the
> hardware or the operating system so as to prevent the user from
> upgrading an OOo that shipped with it, while you can still install
> that upgrade, then you're actively placing limits on the user's
> freedom WRT to that software, and an anti-tivoization clause would
> then stop you from distributing the software under these conditions.

What happens on some particular machine is totally irrelevent. Your GPL
rights do not in any way specially vest in some hardware and not others. The
whole point of GPL rights is they apply to *ANY* way you choose to use the
software. Again, it is a totaly change in the spirit of the GPL to treat the
particular hardware the software is distribute on specially.

> My participation here is about showing that GPLv3, and anti-tivozation
> in particular, don't violate the spirit of the defending users'
> freedoms WRT the covered software, such that the Free Software remains
> Free.

And I think they change it utterly by treating one piece of hardware
different from others for GPL purposes. GPL was always about equal freedom
to use the software on *ALL* hardware, not special rights to use it on one
piece of hardware. More importantly, the change in scope to claim rights
over things that are not derivative works and do not include any GPL'd code
is so massive that it's a change in spirit, IMO.

DS

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 18, 2007, "David Schwartz" <davids@webmaster.com> wrote:

>> > Sure, and you use the hardware to stop me from modifying the
>> > Linux on your
>> > laptop.

>> Do I?  How so?

> Any number of ways. For example, you probably don't connect the serial ports
> to a device I have access to.

But you're not the user of the software on my laptop.  I am.

> I'm sorry, who is "the user"? Who exactly is supposed to be able to install
> and run modified versions? How does the GPLv3 specify who is supposed to be
> authorized to do this?

Aah, good question.  Here's what the draft says about this:

  Mere interaction with a user through a computer network, with no
  transfer of a copy, is not conveying.

The requirements as to "installation information" apply to conveying
the program along with a user product.


> How exactly does the GPLv3 specify who should and should not be able to
> change the software on a particular physical machine?

IANAL, but my understanding is that (paraphrasing), when you convey
the software along with a user product, you must permit the recipient
of the software to install and run modified versions of the software
in the user product as well.

>> A condition that is
>> arguably already encoded in the "no further restrictions to the rights
>> granted" by the license" and to the requirement for complete
>> corresponding source code to accompany the binary.

> Except that the "right" to upload the software on some particular piece of
> hardware was *never* a right granted by the GPL, nor could it be.

It is a restriction on adapting the software installed in the machine,
and a restriction on running the software on that machine.  You can
argue these are not granted by GPLv2.  You may be right.  But per the
spirit of the GPL, they should be protected, and so GPLv3 fixes the
legal conditions such that they are.

> That *HAS* to be a right granted by whatever authority controls the
> use of that hardware.

What if the authority that controls the use of the hardware is
forbidding from restricting this possibility by law?  By contractual
provisions?  By a patent license?  By a copyright license?

> It's totally obvious that who gets to install what software on a
> given piece of hardware is determined by the person who creates/owns
> that hardware and they have to authorize anyone else to change it.

If who creates and who owns are different people, who gets to decide it?

> It is not. The GPL was never about who was allowed to modify the
> software on particular pieces of hardware. It was about the lack of
> *legal* obstacles to your doing so.

GPL has never been concerned *only* about *legal* obstacles.  In fact,
the only obstacle GPLv1 addressed by name was not a legal, but a
technical obstacle: denying access to source code.  Your distinction
is flawed.

>> Both are means to disrespect users' freedoms.

> The freedom to control what software runs on someone else's hardware?!

Freedom to control the software you use on the hardware you use it.

Someone else's hardware is just a distraction.  You're not a user of
software on someone else's hardware.  You have no rights over that.

> And I think they change it utterly by treating one piece of hardware
> different from others for GPL purposes.

No, it's tivoization that does this.

Tivoizers say "hey, you can still modify and run the software, just
not on *this* hardware".

GPLv3 says you must make this artificial distinction.  You must not
place barriers on the freedoms of the user WRT to the GPLv3 software
they use on the hardware you sold/rented/leased/lent/gave them along
with the GPLv3 software you meant them to use.

You can't waive your hands to escape your obligations saying "you can
run it elsewhere", in just the same way you can't escape your GPLv2
obligations to provide source code saying "you can download it
elsewhere"

> GPL was always about equal freedom to use the software on *ALL*
> hardware, not special rights to use it on one piece of hardware.

Exactly.  But tivoizers are making these distinctions, trying to frame
their hardware as somehow special, even though the users that receive
the hardware with the software become users of the software on that
very hardware, and that's why they must be able to enjoy the freedoms
on that very hardware.  Not being able to enjoy them elsewhere could
defensibly be not the vendor's fault.  Not being able to enjoy them on
that hardware is obviously the result of choices made by the vendor,
since the vendor *could* put the software there and get it to run.
Why couldn't the user?

> More importantly, the change in scope to claim rights over things
> that are not derivative works and do not include any GPL'd code is
> so massive that it's a change in spirit, IMO.

Show how patents whose licenses are implicitly granted under GPLv2 are
derivative works and your argument might begin to make sense.


Oh, and user products that GPLv3 talks about *do* include GPLv3 code,
otherwise the license is irrelevant for them, since GPLv3 code is not
being conveyed.  I guess you meant something else when you wrote "do
not include any GPL'ed code".

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[david]

On Mon, 18 Jun 2007, Alexandre Oliva wrote:

> On Jun 18, 2007, "David Schwartz" <davids@webmaster.com> wrote:
>
>>>> Sure, and you use the hardware to stop me from modifying the
>>>> Linux on your
>>>> laptop.
>
>>> Do I?  How so?
>
>> Any number of ways. For example, you probably don't connect the serial ports
>> to a device I have access to.
>
> But you're not the user of the software on my laptop.  I am.

ahh, but by your own argument you aren't

the software on your laptop is owned by people like Linus, Al Viro, David 
M, Alan Cox, etc.

they have the right to put a license on that software that would require 
you to give them access to your hardware (after all, that's the argument 
that you are useing to justify requireing Tivo to give you access to their 
hardware)

David Lang

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Michael Poole]

david@lang.hm writes:

> On Mon, 18 Jun 2007, Alexandre Oliva wrote:
>
>> On Jun 18, 2007, "David Schwartz" <davids@webmaster.com> wrote:
>>
>>>>> Sure, and you use the hardware to stop me from modifying the
>>>>> Linux on your
>>>>> laptop.
>>
>>>> Do I?  How so?
>>
>>> Any number of ways. For example, you probably don't connect the serial ports
>>> to a device I have access to.
>>
>> But you're not the user of the software on my laptop.  I am.
>
> ahh, but by your own argument you aren't
>
> the software on your laptop is owned by people like Linus, Al Viro,
> David M, Alan Cox, etc.

To be pedantic, the *copyrights* for certain software on his laptop
are owned by those people.  (Fortunately, they have been friendly
enough to engage in software quid-pro-quo with those rights.)

> they have the right to put a license on that software that would
> require you to give them access to your hardware (after all, that's
> the argument that you are useing to justify requireing Tivo to give
> you access to their hardware)

Even as straw men go, that is pretty incoherent.

First, end users buy and use the hardware in question.  It does not
belong to Tivo, so the analogy to his laptop fails there.

Second, the important access is not to the hardware, but to the bits
used to build the version of Linux that is distributed by Tivo.  This
is purely software.

Third, such a license would be neither a free software nor an open
source license.  No one argues it would be.

Michael Poole

RE: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Schwartz]

> First, end users buy and use the hardware in question.  It does not
> belong to Tivo, so the analogy to his laptop fails there.

No, this is incorrect. They buy *some* of the rights to the hardware but not
all of them. Specifically, they do not buy the right to choose what software
runs on that hardware. That right is still owned by TiVo.

You can argue that TiVo is being dishonest, breaking the law, being immoral,
or whatever in retaining this right or in failing to disclose that they
retain it. But you cannot coherently deny that TiVo retains this right when
they sell certain other rights to the hardware.

I do in fact argue that there are things that are wrong with TiVo doing
this. But they are not GPL-related things. I would make these same arguments
if the TiVo contained no GPL'd software and I in fact do make them about
products like the Xbox.

DS

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Michael Poole]

David Schwartz writes:

>> First, end users buy and use the hardware in question.  It does not
>> belong to Tivo, so the analogy to his laptop fails there.
>
> No, this is incorrect. They buy *some* of the rights to the hardware but not
> all of them. Specifically, they do not buy the right to choose what software
> runs on that hardware. That right is still owned by TiVo.

Do you have a reference to the contract establishing that cession of
rights from the buyer to Tivo?  To the extent that some contract
purports to restrict the user in ways contrary to the GPL, I suspect
Tivo might have a hard time defending it in court.

> You can argue that TiVo is being dishonest, breaking the law, being immoral,
> or whatever in retaining this right or in failing to disclose that they
> retain it. But you cannot coherently deny that TiVo retains this right when
> they sell certain other rights to the hardware.

By the first sale doctrine, someone who buys an item has practically
unlimited rights to deal with it or dispose of it as the buyer wishes.
The only things that would restrict that are statute or a contract
entered as part of the sale -- most likely a EULA or other shrink-wrap
agreement.  Given that most such recognized agreements deal with
software or services rather than hardware, I am not sure a court would
recognize a hardware EULA as being binding.  (I suspect this is the
direction you were heading with the paragraph below.)

Michael Poole

> I do in fact argue that there are things that are wrong with TiVo doing
> this. But they are not GPL-related things. I would make these same arguments
> if the TiVo contained no GPL'd software and I in fact do make them about
> products like the Xbox.
>
> DS

RE: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Schwartz]

> David Schwartz writes:

> >> First, end users buy and use the hardware in question.  It does not
> >> belong to Tivo, so the analogy to his laptop fails there.

> > No, this is incorrect. They buy *some* of the rights to the
> > hardware but not
> > all of them. Specifically, they do not buy the right to choose
> > what software
> > runs on that hardware. That right is still owned by TiVo.

> Do you have a reference to the contract establishing that cession of
> rights from the buyer to Tivo?

No, and I submit that this is at least arguably something wrong that TiVo is
doing. Note that Microsoft does this too when you buy an Xbox. It has
nothing to do with the GPL.

> To the extent that some contract
> purports to restrict the user in ways contrary to the GPL, I suspect
> Tivo might have a hard time defending it in court.

I agree, however, this doesn't restrict the user in ways contrary to the
GPL. The GPL does not say that you have to be allowed to modify the Linux
running on some particular piece of hardware because that is a legitimate
authorization decision. TiVo not letting you change the software is the same
as me not letting you change the software on my laptop.

> > You can argue that TiVo is being dishonest, breaking the law,
> > being immoral,
> > or whatever in retaining this right or in failing to disclose that they
> > retain it. But you cannot coherently deny that TiVo retains
> > this right when
> > they sell certain other rights to the hardware.

> By the first sale doctrine, someone who buys an item has practically
> unlimited rights to deal with it or dispose of it as the buyer wishes.

This is solely a right against copyright claims. You would be correct if
TiVo were going to sue you for violating some copyright they hold in the
hardware or software if you modified the software.

> The only things that would restrict that are statute or a contract
> entered as part of the sale -- most likely a EULA or other shrink-wrap
> agreement.  Given that most such recognized agreements deal with
> software or services rather than hardware, I am not sure a court would
> recognize a hardware EULA as being binding.  (I suspect this is the
> direction you were heading with the paragraph below.)

Yep, but that has nothing whatsoever to do with the GPL. The exact same
argument applies with the Xbox. It's about whether authorization to modify a
device should or must come with buying that device.

The GPL was never about allowing you to load modified software onto hardware
where the legitimate creators/owners of that hardware say, "no, you may not
modify the software running on this hardware".

DS

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Michael Poole]

David Schwartz writes:

>> David Schwartz writes:
>
>> >> First, end users buy and use the hardware in question.  It does not
>> >> belong to Tivo, so the analogy to his laptop fails there.
>
>> > No, this is incorrect. They buy *some* of the rights to the
>> > hardware but not
>> > all of them. Specifically, they do not buy the right to choose
>> > what software
>> > runs on that hardware. That right is still owned by TiVo.
>
>> Do you have a reference to the contract establishing that cession of
>> rights from the buyer to Tivo?
>
> No, and I submit that this is at least arguably something wrong that TiVo is
> doing. Note that Microsoft does this too when you buy an Xbox. It has
> nothing to do with the GPL.

There is a significant difference between what is a legally recognized
right and what no one has litigated over.  I tend to not recognize the
latter as the former until I see specific backing for the idea that
the purported right has been recognized by law, a court, or all
involved parties.

>> To the extent that some contract
>> purports to restrict the user in ways contrary to the GPL, I suspect
>> Tivo might have a hard time defending it in court.
>
> I agree, however, this doesn't restrict the user in ways contrary to the
> GPL. The GPL does not say that you have to be allowed to modify the Linux
> running on some particular piece of hardware because that is a legitimate
> authorization decision. TiVo not letting you change the software is the same
> as me not letting you change the software on my laptop.

I disagree that the two are the same -- for the fundamental reason
that you have not distributed the software on your laptop to me.  Tivo
has distributed the software on Tivo DVRs to their customers.  The act
of distribution is governed by (in the case of Linux) copyright law
and the GPL.

>> > You can argue that TiVo is being dishonest, breaking the law,
>> > being immoral,
>> > or whatever in retaining this right or in failing to disclose that they
>> > retain it. But you cannot coherently deny that TiVo retains
>> > this right when
>> > they sell certain other rights to the hardware.
>
>> By the first sale doctrine, someone who buys an item has practically
>> unlimited rights to deal with it or dispose of it as the buyer wishes.
>
> This is solely a right against copyright claims. You would be correct if
> TiVo were going to sue you for violating some copyright they hold in the
> hardware or software if you modified the software.

Do you propose that Tivo would (or could) sue a customer for some
non-copyright tort if the customer were to run a Linux kernel that has
not been authorized by Tivo on a Tivo-manufactured DVR?  As far as I
can tell, the legal concerns in question are all copyright issues.

>> The only things that would restrict that are statute or a contract
>> entered as part of the sale -- most likely a EULA or other shrink-wrap
>> agreement.  Given that most such recognized agreements deal with
>> software or services rather than hardware, I am not sure a court would
>> recognize a hardware EULA as being binding.  (I suspect this is the
>> direction you were heading with the paragraph below.)
>
> Yep, but that has nothing whatsoever to do with the GPL. The exact same
> argument applies with the Xbox. It's about whether authorization to modify a
> device should or must come with buying that device.
>
> The GPL was never about allowing you to load modified software onto hardware
> where the legitimate creators/owners of that hardware say, "no, you may not
> modify the software running on this hardware".

True.  The GPL always about allowing someone to modify software that
they received from someone else.  Tivo's Linux kernel images qualify
both as softare that they distribute to others and software that is
loaded onto hardware that they created.  The concern at hand is not
about hardware that Tivo owns or software that Tivo never distributes
-- except where it is also source code for software that they *do*
distribute.

Michael Poole

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 18, 2007, "David Schwartz" <davids@webmaster.com> wrote:

>> First, end users buy and use the hardware in question.  It does not
>> belong to Tivo, so the analogy to his laptop fails there.

> No, this is incorrect. They buy *some* of the rights to the hardware but not
> all of them.

Wow, really?  I thought TiVo actually sold the computer.

Not that it would make a difference as far as GPLv3 is concerned.
It's still a user product, and it still contains GPLed software, and
TiVo distributes that software to other users.

> But you cannot coherently deny that TiVo retains this right when
> they sell certain other rights to the hardware.

Heh.  I mis-parsed "sell rights to the hardware".  How can the
hardware buy something?


Whatever rights TiVo wants to retain or keep from the user is of
little concern here, as long as this doesn't get in the way of the
user's exercise of the freedoms that the GPL stands to defend.  If it
wants to retain more rights than that, then it may have to refrain
from using GPLed software, or face the risk of a court finding it
couldn't have done that in the first place.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

RE: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Schwartz]

> > But you're not the user of the software on my laptop.  I am.

> ahh, but by your own argument you aren't

Let's not confuse owner with user and let's not confuse ownership of
copyrights with ownership of particular copies.

> the software on your laptop is owned by people like Linus, Al Viro, David
> M, Alan Cox, etc.

No. The copyright to the software is owned by those people. But particular
copies of copyrighted items can be owned by other people.

> they have the right to put a license on that software that would require
> you to give them access to your hardware (after all, that's the argument
> that you are useing to justify requireing Tivo to give you access
> to their hardware)

That's right, they do have that right so long as they condition it on the
exercise of something I could not do without their permission. (Ignoring for
the moment the fact that the software is a derivative work of GPL'd
software.)

I'm not sure whether you think this disagrees with or refutes anything I've
said.

DS

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alan Cox]

> the software on your laptop is owned by people like Linus, Al Viro, David 
> M, Alan Cox, etc.

Not quite that simple. An easier way to think about this one is books.
You own the book but you don't own the right to reproduce the words
within. You can however boil the book, use it as bog roll or read it (not
in that order I suggest)

Alan

RE: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Schwartz]

> > Any number of ways. For example, you probably don't connect the
> > serial ports
> > to a device I have access to.

> But you're not the user of the software on my laptop.  I am.

Even when I get web pages from your web server?

> > I'm sorry, who is "the user"? Who exactly is supposed to be
> > able to install
> > and run modified versions? How does the GPLv3 specify who is
> > supposed to be
> > authorized to do this?

> Aah, good question.  Here's what the draft says about this:
>
>   Mere interaction with a user through a computer network, with no
>   transfer of a copy, is not conveying.
>
> The requirements as to "installation information" apply to conveying
> the program along with a user product.

In other words, the GPLv3 *compels* a critical authorization decision to
follow the physical possession of the device. Do you see that, as far as the
GPLv2 is concerned, this is from outer space?

> > How exactly does the GPLv3 specify who should and should not be able to
> > change the software on a particular physical machine?

> IANAL, but my understanding is that (paraphrasing), when you convey
> the software along with a user product, you must permit the recipient
> of the software to install and run modified versions of the software
> in the user product as well.

Which is totally alien to everything in the GPLv2, word and spirit. It never
required any authorization decisions be made any particular way, nor even
hinted that authorization decisions were within its scope.

In fact, there were many discussions where it was made clear that GPLv2
specifically allowed you to make the authorization decisions any way you
want, because it permitted anyone else to remove them. For GPLv3,
apparently, that anyone can remove them is not good enough.

> >> A condition that is
> >> arguably already encoded in the "no further restrictions to the rights
> >> granted" by the license" and to the requirement for complete
> >> corresponding source code to accompany the binary.
>
> > Except that the "right" to upload the software on some
> > particular piece of
> > hardware was *never* a right granted by the GPL, nor could it be.

> It is a restriction on adapting the software installed in the machine,
> and a restriction on running the software on that machine.  You can
> argue these are not granted by GPLv2.  You may be right.  But per the
> spirit of the GPL, they should be protected, and so GPLv3 fixes the
> legal conditions such that they are.

What does the spirit of the GPLv2 say about who is authorized to modify the
software on some particular piece of hardware? This is not per the spirit of
the GPL, it's totally alien to the spirit of the GPL. It has always been
explicitly clear (I can dig up the old discussions if needed) the the GPL
stayed totally away from authorization. Otherwise, you could argue that the
fact that a non-root user can't install a modified kernel "violates the
spirit of the GPL".

> > That *HAS* to be a right granted by whatever authority controls the
> > use of that hardware.

> What if the authority that controls the use of the hardware is
> forbidding from restricting this possibility by law?  By contractual
> provisions?  By a patent license?  By a copyright license?

Those kinds of things are totally alien to the GPL, which was about getting
the source code and being able to modify it and use it on any hardware for
which you were authorized to do so.

> > It's totally obvious that who gets to install what software on a
> > given piece of hardware is determined by the person who creates/owns
> > that hardware and they have to authorize anyone else to change it.

> If who creates and who owns are different people, who gets to decide it?

That's a question on which I would likely agree with you, but it has *ZERO*
to do with the GPL. The GPL was never, until GPLv3, about who gets to make
authorization decisions.

> > It is not. The GPL was never about who was allowed to modify the
> > software on particular pieces of hardware. It was about the lack of
> > *legal* obstacles to your doing so.

> GPL has never been concerned *only* about *legal* obstacles.  In fact,
> the only obstacle GPLv1 addressed by name was not a legal, but a
> technical obstacle: denying access to source code.  Your distinction
> is flawed.

You are taking my claim out of contect. I am distinguishing legal obstacles
from *authorization* obstacles, not technical obstacles. Tivoization is
about authorization even though that authorization is enforced by technical
means.

> >> Both are means to disrespect users' freedoms.
>
> > The freedom to control what software runs on someone else's hardware?!

> Freedom to control the software you use on the hardware you use it.

But that's not a freedom, that's an authorization right that belongs to
someone. Someone gets to choose what software runs on what hardware.

> Someone else's hardware is just a distraction.  You're not a user of
> software on someone else's hardware.  You have no rights over that.

You are. In the case of TiVo, the hardware (specifically the right to decide
what software runs on that hardware) is someone else's. That is part of the
bundle of rights that owning a piece of hardware includes. That is a right
you simply do not have with TiVo.

With respect to control over what software runs on it, your TiVo is someone
else's.

> > And I think they change it utterly by treating one piece of hardware
> > different from others for GPL purposes.

> No, it's tivoization that does this.

How so?

> Tivoizers say "hey, you can still modify and run the software, just
> not on *this* hardware".

Exactly. The GPL is about rights that apply to *all* hardware, not some one
specific piece. That's a massive change in the spirit of the GPL. (Special
rights to one piece of hardware.)

> GPLv3 says you must make this artificial distinction.  You must not
> place barriers on the freedoms of the user WRT to the GPLv3 software
> they use on the hardware you sold/rented/leased/lent/gave them along
> with the GPLv3 software you meant them to use.

Which is a massive departure from the previous GPL spirit which was about
being able to use the software on *ANY* hardware you controlled, not some
special pieces more than others.

> You can't waive your hands to escape your obligations saying "you can
> run it elsewhere", in just the same way you can't escape your GPLv2
> obligations to provide source code saying "you can download it
> elsewhere"

That's a nonsensical comparison. You can run it on any hardware for which
you have the right to say what software runs.

> > GPL was always about equal freedom to use the software on *ALL*
> > hardware, not special rights to use it on one piece of hardware.

> Exactly.  But tivoizers are making these distinctions, trying to frame
> their hardware as somehow special, even though the users that receive
> the hardware with the software become users of the software on that
> very hardware, and that's why they must be able to enjoy the freedoms
> on that very hardware.  Not being able to enjoy them elsewhere could
> defensibly be not the vendor's fault.  Not being able to enjoy them on
> that hardware is obviously the result of choices made by the vendor,
> since the vendor *could* put the software there and get it to run.
> Why couldn't the user?

Because that is not a right the vendor chooses to give to the user. You may
dislike this decision, but it's not irrational. There are any number of
reasons you might want a device to be "trusted".

> > More importantly, the change in scope to claim rights over things
> > that are not derivative works and do not include any GPL'd code is
> > so massive that it's a change in spirit, IMO.

> Show how patents whose licenses are implicitly granted under GPLv2 are
> derivative works and your argument might begin to make sense.

The GPL does not claim any control over those patents. If it included
mandatory licensing of them, then you would have a point.

> Oh, and user products that GPLv3 talks about *do* include GPLv3 code,
> otherwise the license is irrelevant for them, since GPLv3 code is not
> being conveyed.  I guess you meant something else when you wrote "do
> not include any GPL'ed code".

The TiVo loader does not include any GPL'ed code. The TiVo signing keys do
not contain any GPL'ed code. If you are not claiming the GPLv3 exerts any
control over the loader or the keys, then what is left to assure the user
can replace the software on his TiVo?

DS

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 18, 2007, "David Schwartz" <davids@webmaster.com> wrote:

>> > Any number of ways. For example, you probably don't connect the
>> > serial ports
>> > to a device I have access to.

>> But you're not the user of the software on my laptop.  I am.

> Even when I get web pages from your web server?

Yes.  I'm (hypothetically) running the web server such that it serves
web pages to you or anyone else.  You don't become user of a software
just because you establish a network session with it.  You clearly
need more than that.

That said, the precise threshold isn't clear.  For complex web
applications that run part on the client and part on the server, or
even almost entirely on the server, one can argue that a user is
indeed using the software even though it runs mostly on the server.
Some people call this kind of situation the ASP loophole in the GPL.
I'm not sure I want to disturb users of this list with the details
about this, but I'd be pleased to discuss this off the list.

>> The requirements as to "installation information" apply to conveying
>> the program along with a user product.

> In other words, the GPLv3 *compels* a critical authorization decision to
> follow the physical possession of the device. Do you see that, as far as the
> GPLv2 is concerned, this is from outer space?

Not really.  When a user receives a copy of the software, there's
distribution going on, and that's when the user can start having any
expectations of having her freedoms respected as to that software.

>> > How exactly does the GPLv3 specify who should and should not be able to
>> > change the software on a particular physical machine?

>> IANAL, but my understanding is that (paraphrasing), when you convey
>> the software along with a user product, you must permit the recipient
>> of the software to install and run modified versions of the software
>> in the user product as well.

> Which is totally alien to everything in the GPLv2, word and spirit. It never
> required any authorization decisions be made any particular way, nor even
> hinted that authorization decisions were within its scope.

It's the authorization decisions that are alien to GPLv2.  That's just
yet another form of denying users the freedoms that they ought to
receive along with the software.

> What does the spirit of the GPLv2 say about who is authorized to modify the
> software on some particular piece of hardware?

It doesn't.  Why should it have to?  Whether someone is authorized or
not is a direct consequence of the freedoms.  The moment the software
was distributed to you, you're entitled to the freedoms.  Imposing
restrictions on them is a violation of the spirit, if not the letter,
of the license.

>> What if the authority that controls the use of the hardware is
>> forbidding from restricting this possibility by law?  By contractual
>> provisions?  By a patent license?  By a copyright license?

> Those kinds of things are totally alien to the GPL, which was about getting
> the source code and being able to modify it and use it on any hardware for
> which you were authorized to do so.

This is a very limited reading of the GPL that leaves out one of its
most important provisions: the bit about "no further restrictions".

> The GPL was never, until GPLv3, about who gets to make
> authorization decisions.

I can agree with that.  As long as the authorization decisions are not
used as means to deprive users' of the freedoms that must not be
restricted, they can be whatever the distributor fancies.

> You are taking my claim out of contect. I am distinguishing legal obstacles
> from *authorization* obstacles, not technical obstacles.

It doesn't matter how elaborate the excuse to disrespect the freedoms
of the user is.  If there are further restrictions to them, then this
violates the spirit, if not the letter, of the GPL.

>> Someone else's hardware is just a distraction.  You're not a user of
>> software on someone else's hardware.  You have no rights over that.

> You are. In the case of TiVo, the hardware (specifically the right
> to decide what software runs on that hardware) is someone
> else's. That is part of the bundle of rights that owning a piece of
> hardware includes. That is a right you simply do not have with TiVo.

Ah, ok, so I was sloppy above and you caught that.

If someone else places hardware on your home for you to use, even if
they still own it, then you can be a user of someone else's hardware.
And at that point the GPL kicks in, because the software was
distributed to you (even if the hardware wasn't sold), and with the
distributed software come the freedoms, which, per the GPL, the
distributor must not disrespect.

>> > And I think they change it utterly by treating one piece of hardware
>> > different from others for GPL purposes.

>> No, it's tivoization that does this.

> How so?

Like this:

>> Tivoizers say "hey, you can still modify and run the software, just
>> not on *this* hardware".

Tivoization is treating the hardware that comes along with the
software as if it was different from others.  But it isn't.

> Exactly. The GPL is about rights that apply to *all* hardware, not some one
> specific piece.

Exactly!  Just like the GPL doesn't permit the distributor to state
"BTW, you can't install or run this software on your mother's
computer", it doesn't permit the distributor to state "BTW, you can't
install or run this software on this computer I'm selling you".  The
"no further restrictions" applies equally to all computers.  It's not
just because you have some control over some particular hardware that
you deliver along with the software that you're entitled to use that
to limit the user's freedoms.

>> GPLv3 says you must make this artificial distinction.  You must not
>> place barriers on the freedoms of the user WRT to the GPLv3 software
>> they use on the hardware you sold/rented/leased/lent/gave them along
>> with the GPLv3 software you meant them to use.

> Which is a massive departure from the previous GPL spirit which was about
> being able to use the software on *ANY* hardware you controlled, not some
> special pieces more than others.

It doesn't make the sold hardware special.  How come you think it
does?  It's exactly the opposite.  It just says the distributor can't
make the hardware special, so as to restrain the users' freedoms that
are inseparable from the software.

>> You can't waive your hands to escape your obligations saying "you can
>> run it elsewhere", in just the same way you can't escape your GPLv2
>> obligations to provide source code saying "you can download it
>> elsewhere"

> That's a nonsensical comparison. You can run it on any hardware for which
> you have the right to say what software runs.

And why don't I have the right to say what software runs on the
hardware I received along with the GPLed software?  Because the
tivoizer doesn't want me to.  The tivoizer is placing barriers such
that I cannot adapt the GPLed software included in that device to my
own needs.  How is that not a further restriction to the four
freedoms?  How is that not making that hardware special?

> Because that is not a right the vendor chooses to give to the user.

As in, the vendor can turn to the user and sue her for patent
infringement, after distributing GPLed software to her, just because
the use of the patent is not a right the vendor chooses to give to the
user?

> You may dislike this decision, but it's not irrational.

I never said it was irrational.  I just said it's a further
restriction on the exercise of the freedoms that must accompany the
software wherever it goes.

>> > More importantly, the change in scope to claim rights over things
>> > that are not derivative works and do not include any GPL'd code is
>> > so massive that it's a change in spirit, IMO.

>> Show how patents whose licenses are implicitly granted under GPLv2 are
>> derivative works and your argument might begin to make sense.

> The GPL does not claim any control over those patents. If it included
> mandatory licensing of them, then you would have a point.

It doesn't because the US law makes that implicit.  GPLv3 makes it
explicit because it was found that it wasn't like this everywhere.

>> Oh, and user products that GPLv3 talks about *do* include GPLv3 code,
>> otherwise the license is irrelevant for them, since GPLv3 code is not
>> being conveyed.  I guess you meant something else when you wrote "do
>> not include any GPL'ed code".

> The TiVo loader does not include any GPL'ed code. The TiVo signing keys do
> not contain any GPL'ed code. If you are not claiming the GPLv3 exerts any
> control over the loader or the keys, then what is left to assure the user
> can replace the software on his TiVo?

Absence of disrespect for users' freedoms.  Any measure taken by the
vendor to disrespect them is a failure to comply with the obligations
imposed by the spirit, if not the letter, of the license.

And, just in case, IANAL ;-)

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

RE: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Schwartz]

> This is a very limited reading of the GPL that leaves out one of its
> most important provisions: the bit about "no further restrictions".

Why is the fact that only the root user can load a kernel module not a
further restriction? Simple -- anyone who is bothered by that restriction
can remove it on any hardware for which they have the right to load modified
software. Anyone who does not have the right to load modified software on
some hardware simply does not have the right to change it.

Why is that not a "further restriction"? It means that I can't load kernel
modules on hardware that I don't have the right to load modified software
on.

> > The GPL was never, until GPLv3, about who gets to make
> > authorization decisions.

> I can agree with that.  As long as the authorization decisions are not
> used as means to deprive users' of the freedoms that must not be
> restricted, they can be whatever the distributor fancies.

Right, which is the freedom to modify the software. The freedom to get the
source code. The freedom to use the source code however you want, absent
legitimate authorization decisions to the contrary.

> > You are taking my claim out of contect. I am distinguishing
> > legal obstacles
> > from *authorization* obstacles, not technical obstacles.
>
> It doesn't matter how elaborate the excuse to disrespect the freedoms
> of the user is.  If there are further restrictions to them, then this
> violates the spirit, if not the letter, of the GPL.

I agree. However, "you can't load your modified sofware on *MY* hardware" is
not a further restriction. If it was, we get absurdities.

> >> Someone else's hardware is just a distraction.  You're not a user of
> >> software on someone else's hardware.  You have no rights over that.
>
> > You are. In the case of TiVo, the hardware (specifically the right
> > to decide what software runs on that hardware) is someone
> > else's. That is part of the bundle of rights that owning a piece of
> > hardware includes. That is a right you simply do not have with TiVo.

> Ah, ok, so I was sloppy above and you caught that.

> If someone else places hardware on your home for you to use, even if
> they still own it, then you can be a user of someone else's hardware.

Definitely.

> And at that point the GPL kicks in, because the software was
> distributed to you (even if the hardware wasn't sold), and with the
> distributed software come the freedoms, which, per the GPL, the
> distributor must not disrespect.

Absolutely.

> >> Tivoizers say "hey, you can still modify and run the software, just
> >> not on *this* hardware".

> Tivoization is treating the hardware that comes along with the
> software as if it was different from others.  But it isn't.

Of course it is. They have the authorization right on that hardware, and
they don't have that right on my laptop. For any piece of hardware, there
has to be someone who decides who can and can't choose what software runs on
that hardware.

> > Exactly. The GPL is about rights that apply to *all* hardware,
> > not some one
> > specific piece.

> Exactly!  Just like the GPL doesn't permit the distributor to state
> "BTW, you can't install or run this software on your mother's
> computer", it doesn't permit the distributor to state "BTW, you can't
> install or run this software on this computer I'm selling you".

That would mean it doesn't permit the distribute to state "BTW, you can't
install, modify or run this software on *OUR* computers that run our
corporate network". Don't you see how obviously absurd that is?

Someone has to be authorized to decide what software runs on some particular
piece of hardware. The GPL cannot mean that other people get to modify and
run software on that particular piece of hardware.

> The
> "no further restrictions" applies equally to all computers.  It's not
> just because you have some control over some particular hardware that
> you deliver along with the software that you're entitled to use that
> to limit the user's freedoms.

I agree. However, that doesn't mean that people who own or control
particular pieces of hardware can't put authorization barriers that prevent
you from running whatever software you want on thos pieces of hardware.

> > Which is a massive departure from the previous GPL spirit which
> > was about
> > being able to use the software on *ANY* hardware you
> > controlled, not some
> > special pieces more than others.

> It doesn't make the sold hardware special.  How come you think it
> does?

Because it becomes the only piece of hardware in the entire universe on
which the GPL gives you the right to run the software. On every other piece
of hardware, you must obtain that right from whoever owns the right to
decide what software runs on that hardware.

> It's exactly the opposite.  It just says the distributor can't
> make the hardware special, so as to restrain the users' freedoms that
> are inseparable from the software.

Don't you see that the rule that "this one thing cannot be special" makes
that one thing special since everything else *can* be special.

> > That's a nonsensical comparison. You can run it on any hardware
> > for which
> > you have the right to say what software runs.

> And why don't I have the right to say what software runs on the
> hardware I received along with the GPLed software?  Because the
> tivoizer doesn't want me to.  The tivoizer is placing barriers such
> that I cannot adapt the GPLed software included in that device to my
> own needs.  How is that not a further restriction to the four
> freedoms?  How is that not making that hardware special?

You can adapt it to your own needs, you just can't run it on hardware you
don't fully own. You do not fully own the TiVo because you do not own the
right to run modified software on it. It is just like *my* laptop -- you
don't own the right to choose what software runs on it. Someone has to have
that right, and in the case of TiVo, it's not you.

> > Because that is not a right the vendor chooses to give to the user.

> As in, the vendor can turn to the user and sue her for patent
> infringement, after distributing GPLed software to her, just because
> the use of the patent is not a right the vendor chooses to give to the
> user?

I don't know what patent you are talking about.

> > You may dislike this decision, but it's not irrational.

> I never said it was irrational.  I just said it's a further
> restriction on the exercise of the freedoms that must accompany the
> software wherever it goes.

No more than having to be 'root' to load a kernel module. You are free to
remove it from any hardware for which you have the right to choose what
software runs.

DS

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 18, 2007, "David Schwartz" <davids@webmaster.com> wrote:

> Why is the fact that only the root user can load a kernel module not a
> further restriction?

Because the user (under whose control the computer is, be it person or
company) set up the root password herself?

>> > The GPL was never, until GPLv3, about who gets to make
>> > authorization decisions.

>> I can agree with that.  As long as the authorization decisions are not
>> used as means to deprive users' of the freedoms that must not be
>> restricted, they can be whatever the distributor fancies.

> Right, which is the freedom to modify the software. The freedom to get the
> source code. The freedom to use the source code however you want, absent
> legitimate authorization decisions to the contrary.

What makes them lawful, given the "no further restrictions"?

> However, "you can't load your modified sofware on *MY* hardware" is
> not a further restriction.

As long as you didn't hand me the hardware along with the software,
for me to become a user of the software on that hardware, I agree.

>> >> Tivoizers say "hey, you can still modify and run the software, just
>> >> not on *this* hardware".

>> Tivoization is treating the hardware that comes along with the
>> software as if it was different from others.  But it isn't.

> Of course it is. They have the authorization right on that hardware, and
> they don't have that right on my laptop.

Ok, I stand corrected.  They have that right.

However, since they distribute GPLed software along with the hardware,
such that I'd be a user of the software on that hardware, they should
not impose further restrictions on my freedoms that the GPL stands to
defend WRT the GPLed software.  So, they must not use their
authorization right to deny me, the user of the software, in the
hardware that they meant me to use the software, the freedom to adapt
the software for my own needs, and run it for any purpose.

> That would mean it doesn't permit the distribute to state "BTW, you can't
> install, modify or run this software on *OUR* computers that run our
> corporate network".

No, because the user is not becoming a user of the software on their
own computers.  Only in the computer that was shipped along with the
software.

> Don't you see how obviously absurd that is?

Yes, it would be, if it were so.

>> The "no further restrictions" applies equally to all computers.
>> It's not just because you have some control over some particular
>> hardware that you deliver along with the software that you're
>> entitled to use that to limit the user's freedoms.

> I agree. However, that doesn't mean that people who own or control
> particular pieces of hardware can't put authorization barriers that
> prevent you from running whatever software you want on thos pieces
> of hardware.

That's correct, as long as they didn't give me that hardware with
GPLed software in it.  The moment they do, I become recipient and user
of GPLed software in that computer, and they should relinquish their
power to impose restrictions on my exercise of the freedoms WRT that
software.  And there's no reason whatsoever to exclude restrictions
such as those implemented by means of authorization.

>> > Which is a massive departure from the previous GPL spirit which
>> > was about being able to use the software on *ANY* hardware you
>> > controlled, not some special pieces more than others.

>> It doesn't make the sold hardware special.  How come you think it
>> does?

> Because it becomes the only piece of hardware in the entire universe on
> which the GPL gives you the right to run the software. On every other piece
> of hardware, you must obtain that right from whoever owns the right to
> decide what software runs on that hardware.

I see.  Good point.  Agreed.  That hardware is indeed special.  Per
the GPL, it's the only one in which the distributor must NOT exercise
any restraints whatsoever on my exercise of the freedoms.

Which in turns makes it non-special, in that, from the point of both
the distributor and the user, it becomes just like any other random
piece of hardware: the distributor doesn't limit the freedoms the user
had, and the user isn't limited in enjoying the freedoms she had.

>> It's exactly the opposite.  It just says the distributor can't
>> make the hardware special, so as to restrain the users' freedoms that
>> are inseparable from the software.

> Don't you see that the rule that "this one thing cannot be special" makes
> that one thing special since everything else *can* be special.

It took me several attempts to understand what you meant.  Yes.  It's
special in that it can't be made special.  Poof, there goes the
universe ;-)

>> > Because that is not a right the vendor chooses to give to the user.

>> As in, the vendor can turn to the user and sue her for patent
>> infringement, after distributing GPLed software to her, just because
>> the use of the patent is not a right the vendor chooses to give to the
>> user?

> I don't know what patent you are talking about.

I'm talking about the implicit patent licenses that arise from
distributing software under GPLv2.

>> > You may dislike this decision, but it's not irrational.

>> I never said it was irrational.  I just said it's a further
>> restriction on the exercise of the freedoms that must accompany the
>> software wherever it goes.

> No more than having to be 'root' to load a kernel module. You are free to
> remove it from any hardware for which you have the right to choose what
> software runs.

Yup.  And I get that right (because the distributor must not stop me)
when I receive software under the GPL along with the computer in which
I'm expected to use it.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

RE: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Schwartz]

> On Jun 18, 2007, "David Schwartz" <davids@webmaster.com> wrote:

> > Why is the fact that only the root user can load a kernel module not a
> > further restriction?

> Because the user (under whose control the computer is, be it person or
> company) set up the root password herself?

Well, duh. TiVo, under whose control the software running on my Tivo is, set
up the signing key themself. *Someone* has to decide what software runs, the
GPL cannot rationally decide who that is because it's an
application-specific authorization decision.

> >> > The GPL was never, until GPLv3, about who gets to make
> >> > authorization decisions.
>
> >> I can agree with that.  As long as the authorization decisions are not
> >> used as means to deprive users' of the freedoms that must not be
> >> restricted, they can be whatever the distributor fancies.

> > Right, which is the freedom to modify the software. The freedom
> > to get the
> > source code. The freedom to use the source code however you want, absent
> > legitimate authorization decisions to the contrary.

> What makes them lawful, given the "no further restrictions"?

That the person who decides what software runs on that hardware can remove
them if they please. *Someone* has to decide what software runs on a
particular piece of hardware, right?

> > However, "you can't load your modified sofware on *MY* hardware" is
> > not a further restriction.

> As long as you didn't hand me the hardware along with the software,
> for me to become a user of the software on that hardware, I agree.

This is, again, an argument that is totally alien to the GPL. The idea that
you have 'special' rights to the software on some hardware but not others is
simply insane. It is totally out of left field with respect to the GPL. The
GPL is about being able to use the software on *ANY* hardware for which you
have the right to decide what software runs.

> However, since they distribute GPLed software along with the hardware,
> such that I'd be a user of the software on that hardware, they should
> not impose further restrictions on my freedoms that the GPL stands to
> defend WRT the GPLed software.  So, they must not use their
> authorization right to deny me, the user of the software, in the
> hardware that they meant me to use the software, the freedom to adapt
> the software for my own needs, and run it for any purpose.

You can argue this, but it's not a GPL argument. It's a reasonable argument,
but it has nothing whatsoever to do with GPL rights. GPL rights are about
being able to use the software on *any* hardware you want, not special
rights to use the software on some one particular piece of hardware. GPL
rights are rights against obstacles to getting the source code and legally
modifying it and distributing it, not rights against authorization obstacles
placed by people who own hardware.

> > That would mean it doesn't permit the distribute to state "BTW,
> > you can't
> > install, modify or run this software on *OUR* computers that run our
> > corporate network".

> No, because the user is not becoming a user of the software on their
> own computers.  Only in the computer that was shipped along with the
> software.

Yes, they are becoming a user. They might very well be using those
computers. It's absolutely absurd to argue that the right to choose what
software runs on a piece of hardware must go to the user of that hardware.
In any event, it's totally alien to the GPL which is not at all about who
gets to decide what software runs on what hardware.

> >> The "no further restrictions" applies equally to all computers.
> >> It's not just because you have some control over some particular
> >> hardware that you deliver along with the software that you're
> >> entitled to use that to limit the user's freedoms.

> > I agree. However, that doesn't mean that people who own or control
> > particular pieces of hardware can't put authorization barriers that
> > prevent you from running whatever software you want on thos pieces
> > of hardware.

> That's correct, as long as they didn't give me that hardware with
> GPLed software in it.  The moment they do, I become recipient and user
> of GPLed software in that computer, and they should relinquish their
> power to impose restrictions on my exercise of the freedoms WRT that
> software.  And there's no reason whatsoever to exclude restrictions
> such as those implemented by means of authorization.

You can state what the GPLv3 does as many times as you want, but special
rights to particular pieces of hardware is *TOTALLY* alien to the spirit of
the GPL. The GPL was always about equal rights to use the software in any
hardware.

DS

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 19, 2007, "David Schwartz" <davids@webmaster.com> wrote:

>> On Jun 18, 2007, "David Schwartz" <davids@webmaster.com> wrote:

>> > Why is the fact that only the root user can load a kernel module not a
>> > further restriction?

>> Because the user (under whose control the computer is, be it person or
>> company) set up the root password herself?

> Well, duh. TiVo, under whose control the software running on my Tivo is, set
> up the signing key themself. *Someone* has to decide what software runs, the
> GPL cannot rationally decide who that is because it's an
> application-specific authorization decision.

Right.  All GPL can say is that you cannot impose further restrictions
on how the user adapts the software, and since the user runs the
software on that computer, that means you must not restrict the user's
ability to upgrade or otherwise replace that software there, when you
gave the user the software along with the computer.


>> > However, "you can't load your modified sofware on *MY* hardware" is
>> > not a further restriction.

>> As long as you didn't hand me the hardware along with the software,
>> for me to become a user of the software on that hardware, I agree.

> This is, again, an argument that is totally alien to the GPL.

No, it's not.  It is intended to ensure that free software remains
free for all its users.  When you receive the software, you become a
user.  That's when you receive the rights, and that's what creates the
obligation on the distributor to not impose restrictions on the
freedoms, no matter by how means such restrictions could be legally or
technically accomplished.

> The idea that you have 'special' rights to the software on some
> hardware but not others is simply insane.

I agree, to some extent.  It's not so much about the rights, but about
the restrictions the vendor can impose on hardware.

It's just that, for this particular hardware, as you say, the
manufacturer has (or had) special rights.  This means it can decide
what software runs, whom it gives the hardware to, etc.

However, by distributing software under the GPL, the vendor accepts
the condition to not use any means whatsoever to impose restrictions
on the recipient's exercise of the rights granted by the license by
means of the distribution of the software.

There's no reason to make the hardware special, or the right of
authorization special, as a possible excuse to impose restrictions on
the user.  It amounts to just that: an attempt to excuse oneself from
the condition of not imposing restrictions on the enjoyment of the
freedoms.

> The GPL is about being able to use the software on *ANY* hardware
> for which you have the right to decide what software runs.

Yes.  And, per the "pass on all rights you have" spirit in the
preamble, that translates into "no further restrictions" in the legal
terms, the user *must* receive this right from the distributor of the
software.


Oh, but what if the distributor doens't have this right in the first
place?  Well, let's see...

Either the distributor received the hardware with the software inside
it, which means it should have received this right along with the
software from whoever gave it the software, so it has this right, or
it installed the software itself, which means it does have this
right.  In both cases.

>> No, because the user is not becoming a user of the software on their
>> own computers.  Only in the computer that was shipped along with the
>> software.

> Yes, they are becoming a user. They might very well be using those
> computers.

This means they already were users there of those computers, just not
necessarily of that software on those computers.

And then, if they choose to copy and run the software on other
computers where they are entitled to install software, they're free to
do so, the vendor of that other piece of hardware must not impose
restrictions on that either.

> You can state what the GPLv3 does as many times as you want, but special
> rights to particular pieces of hardware is *TOTALLY* alien to the spirit of
> the GPL.

I agree.  That's the bug in GPLv2 that the anti-tivoization provision
is trying to fix.

> The GPL was always about equal rights to use the software in any
> hardware.

Exactly.  Thank you.  It finally sank in, it seems.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

RE: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Schwartz]

> Right.  All GPL can say is that you cannot impose further restrictions
> on how the user adapts the software, and since the user runs the
> software on that computer, that means you must not restrict the user's
> ability to upgrade or otherwise replace that software there, when you
> gave the user the software along with the computer.

You keep smuggling in the same assumption without ever defending it. There
is a user. There is a person who gets to decide what software runs on a
particular piece of hardware. You keep assuming they must be the same
person. There are *MANY* legitimate reasons why the user of a piece of
hardware should not be the same person who controls what software runs on
that hardware.

> >> As long as you didn't hand me the hardware along with the software,
> >> for me to become a user of the software on that hardware, I agree.

> > This is, again, an argument that is totally alien to the GPL.

> No, it's not.  It is intended to ensure that free software remains
> free for all its users.

Exactly, on all hardware. Not "especially free" on some one particular
piece.

> When you receive the software, you become a
> user.  That's when you receive the rights, and that's what creates the
> obligation on the distributor to not impose restrictions on the
> freedoms, no matter by how means such restrictions could be legally or
> technically accomplished.

Exactly. And they place no restrictions on your ability to modify or use
that software on any hardware you like, provided of course you are the
person who gets to decide what software runs on that hardware.

> > The idea that you have 'special' rights to the software on some
> > hardware but not others is simply insane.

> I agree, to some extent.  It's not so much about the rights, but about
> the restrictions the vendor can impose on hardware.

The GPL is about what restrictions a particule piece of hardware, that
contains no GPL'd software, can impose?

> It's just that, for this particular hardware, as you say, the
> manufacturer has (or had) special rights.  This means it can decide
> what software runs, whom it gives the hardware to, etc.

Exactly. For any given piece of hardware, there must be some person or
entity that decides what software runs on it.

> However, by distributing software under the GPL, the vendor accepts
> the condition to not use any means whatsoever to impose restrictions
> on the recipient's exercise of the rights granted by the license by
> means of the distribution of the software.

Agreed.

> There's no reason to make the hardware special, or the right of
> authorization special, as a possible excuse to impose restrictions on
> the user.  It amounts to just that: an attempt to excuse oneself from
> the condition of not imposing restrictions on the enjoyment of the
> freedoms.

There is always the restriction that if you aren't the person who gets to
choose what software runs on a particular piece of hardware, then you can't
run modified software on that hardware.

There has to be someone who makes that decision for any given piece of
hardware. The idea that this person *MUST* be the user is totally alien to
the GPL. It's got nothing whatsoever to do with *ANY* of the freedoms the
GPL was protecting. All of those freedoms very critically apply to *ALL*
hardware in the entire universe.

> > The GPL is about being able to use the software on *ANY* hardware
> > for which you have the right to decide what software runs.

> Yes.  And, per the "pass on all rights you have" spirit in the
> preamble, that translates into "no further restrictions" in the legal
> terms, the user *must* receive this right from the distributor of the
> software.

That right is a right to that particular piece of hardware, it is not a
right to the GPL'd software. Your argument suggests that if I let you use my
laptop, I must let you modify the Linux kernel on it. That's just craziness.
The GPL was never about who was authorized to install modified software on
particular pieces of hardware.

> > You can state what the GPLv3 does as many times as you want, but special
> > rights to particular pieces of hardware is *TOTALLY* alien to
> > the spirit of the GPL.

> I agree.  That's the bug in GPLv2 that the anti-tivoization provision
> is trying to fix.

You can see it as a bug, and if you think Tivoization of free software is
bad, then that view makes sense. However, if you see the GPL as being about
getting the software, being free to modify the software, being able to
install that software on *ANY* hardware (whether or not that hardware
shipped with open-souruce software) and if you see the GPL as avoiding any
restrictions on authorization decisions, then the GPLv3 is not "fixing"
something but radically doing something else entirely.

>From my point of view, the biggest problem with the GPLv3 is not the change
in spirit but the change in scope. The GPLv3 attempts to control the Tivo
hardware, firmware, and keys, none of which contain any GPL'd software at
all. The idea that because you use GPL'd software, restrictions are imposed
on non-GPL'd works bothers me tremendously.

DS

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 19, 2007, "David Schwartz" <davids@webmaster.com> wrote:

>> Right.  All GPL can say is that you cannot impose further restrictions
>> on how the user adapts the software, and since the user runs the
>> software on that computer, that means you must not restrict the user's
>> ability to upgrade or otherwise replace that software there, when you
>> gave the user the software along with the computer.

> You keep smuggling in the same assumption without ever defending it. There
> is a user. There is a person who gets to decide what software runs on a
> particular piece of hardware. You keep assuming they must be the same
> person.

No, I'm just saying that whoever gets to decide cannot restrict the
user's freedoms as to the software the user received.


Consider this:

I get GPLed software.

I make improvements to it.

I give it to you, but I leave out the sources of my changes.

You ask me for sources, because without them you can't enjoy the
freedom to adapt the software.

I say "No, they're mine.  I have the right to keep them and release
them however I like.  Copyright law says so.!

You talk to the copyright holder, and he revokes my license and gets a
court order such that I can't distribute the software any more.


You see?  It's not because I had a right that I can use it to impose
restrictions on your freedoms, after I distribute the software to you.

Right to control what software runs on the hardware is no different.
For any hardware on which I can run the software, I'm a user there,
and I'm entitled to the rights granted by the license.


It's really this simple.  Don't complicate the issue by trying to make
hardware special.  It's just an illusion to try to convince yourself
that you can deprive users of freedoms provided by the GPL.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

RE: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Schwartz]

> > You keep smuggling in the same assumption without ever
> > defending it. There
> > is a user. There is a person who gets to decide what software runs on a
> > particular piece of hardware. You keep assuming they must be the same
> > person.

> No, I'm just saying that whoever gets to decide cannot restrict the
> user's freedoms as to the software the user received.

I agree. However, the freedom to run modified software on hardware for which
one is not the person who gets to decide what software runs on that hardware
is not one of those freedoms.

> Consider this:
>
> I get GPLed software.
>
> I make improvements to it.
>
> I give it to you, but I leave out the sources of my changes.
>
> You ask me for sources, because without them you can't enjoy the
> freedom to adapt the software.
>
> I say "No, they're mine.  I have the right to keep them and release
> them however I like.  Copyright law says so.!
>
> You talk to the copyright holder, and he revokes my license and gets a
> court order such that I can't distribute the software any more.

Right.

> You see?  It's not because I had a right that I can use it to impose
> restrictions on your freedoms, after I distribute the software to you.

Right.

> Right to control what software runs on the hardware is no different.
> For any hardware on which I can run the software, I'm a user there,
> and I'm entitled to the rights granted by the license.

Exactly. However, that right does not include the right to run the software
on any particular piece of hardware. It includes the right to get the source
code, redistribute it, modify it, and so on. It includes the right to run
the software on *ANY* hardware, so long as one is authorized to choose what
software runs on that hardware.

> It's really this simple.  Don't complicate the issue by trying to make
> hardware special.  It's just an illusion to try to convince yourself
> that you can deprive users of freedoms provided by the GPL.

The right to run GPL'd software on some particular piece of hardware was
never a GPL right. That is an *authorization* decision, similar to who has
'root' access on a Linux box and can modify and install the kernel.

You keep conflating things like access to the source code and the legal
right to modify that source code with the authorization right to install
that modified kernel on some particular piece of hardware. The GPL was never
about how such authorization decisions are made.

If I make you a user of my laptop, do I have to let you install a modified
kernel? What if I even let you access the kernel source, so I have
distributed it to you.

The GPL was never, ever about such authorization decisions. They are
completely alien to both the wording and the spirit of the GPL.

DS

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 19, 2007, "David Schwartz" <davids@webmaster.com> wrote:

>> Right to control what software runs on the hardware is no different.
>> For any hardware on which I can run the software, I'm a user there,
>> and I'm entitled to the rights granted by the license.

> Exactly. However, that right does not include the right to run the software
> on any particular piece of hardware.

I think that's correct, in general.  But when I receive the software
along with a particular piece of hardware, on which I'm to expected to
run it, I'm also receiving all other freedoms that ought to be
respected by whoever distributed the software to me, which means the
distributor can't impose restrictions on my running modified versions
of the program there (or anywhere else I can run the software, for
that matter).

> It includes the right to run the software on *ANY* hardware, so long
> as one is authorized to choose what software runs on that hardware.

That's correct, too, with the provision that the distributor cannot
impose restrictions on running modified versions of the program.  Not
on any other hardware I can control, not on the hardware with which I
received the software along with the freedoms to control it in as far
as running the GPLed software goes.

> You keep conflating things like access to the source code and the legal
> right to modify that source code with the authorization right to install
> that modified kernel on some particular piece of hardware. The GPL was never
> about how such authorization decisions are made.

That's correct.  All it says is that you can't impose further
restrictions.  This means you can't use anything (source code
deprivation, patents, copyright, authorization, nothing) to disrespect
users' freedoms.  Is this so hard to accept?  There's nothing special
about these authorization rights you're talking about that distinguish
them from copyrights or patent rights or any other rights.  The GPL
says: don't use them against users' freedoms regarding the GPLed
software.  No excuses.

> If I make you a user of my laptop, do I have to let you install a modified
> kernel?

Does interacting with your laptop over the network make me a user of
your kernel?

> What if I even let you access the kernel source, so I have
> distributed it to you.

You have distributed the source to me, I can modify it.

But did you distribute the kernel binary installed on your computer to
me?  That's the one I'm (possibly) using per the above.  Whether you
also gave me sources is irrelevant, unless you gave them to me as part
of your obligation of distributing the corresponding sources, in case
we conclude you distributed the kernel binary installed on your
computer to me when you granted me remote access to it.

And then, GPLv3 makes it clear that, in the case of remote access,
you're not conveying the binary to me, therefore the conditions for
conveying do not apply.

> The GPL was never, ever about such authorization decisions. They are
> completely alien to both the wording and the spirit of the GPL.

You can repeat that as much as you want, this won't change the fact
that the GPL has never permitted you to use whatever rights you have
to impose restrictions on users' freedoms as to GPLed software once
you've (implicitly) accepted the conditions of the license.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Lennart Sorensen]

On Mon, Jun 18, 2007 at 06:12:57PM -0300, Alexandre Oliva wrote:
> Aah, good question.  Here's what the draft says about this:
> 
>   Mere interaction with a user through a computer network, with no
>   transfer of a copy, is not conveying.
> 
> The requirements as to "installation information" apply to conveying
> the program along with a user product.

So if I go use a computer running some GPL software, and I copy the
contents of /bin to a CD and bring it home, does the owner of the
machine now owe me a copy of the GPL sources?

--
Len Sorensen

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 20, 2007, lsorense@csclub.uwaterloo.ca (Lennart Sorensen) wrote:

> On Mon, Jun 18, 2007 at 06:12:57PM -0300, Alexandre Oliva wrote:
>> Aah, good question.  Here's what the draft says about this:
>> 
>> Mere interaction with a user through a computer network, with no
>> transfer of a copy, is not conveying.
>> 
>> The requirements as to "installation information" apply to conveying
>> the program along with a user product.

> So if I go use a computer running some GPL software, and I copy the
> contents of /bin to a CD and bring it home, does the owner of the
> machine now owe me a copy of the GPL sources?

According to one of the rationales of GPLv3, it is understood that
lending someone a computer for a short period of time does not amount
to conveying the software in it.  I assume this is backed by strong
legal reasoning I won't pretend to know or understand.  IANAL.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Lennart Sorensen]

On Sun, Jun 17, 2007 at 12:52:38AM -0300, Alexandre Oliva wrote:
> What it does is impose conditions for whoever wants to distribute the
> software.  And GPLv3 makes it explicit that one such condition is to
> permit the user to install and run modified versions of the program in
> the hardware that ships with the program.  A condition that is
> arguably already encoded in the "no further restrictions to the rights
> granted" by the license" and to the requirement for complete
> corresponding source code to accompany the binary.
> 
> That you disagree with it doesn't make you right.
> 
> But that it is within the spirit of the GPL defined by its authors
> (which is all I'm trying to show here), it is.
> 
> > The GPL (at least through version 2) is about free access to source
> > code.
> 
> Some think so, but this was GPLv1.
> 
> v2 added stuff such as:
> 
>   if a patent license would not permit royalty-free redistribution of
>   the Program by all those who receive copies directly or indirectly
>   through you, then the only way you could satisfy both it and this
>   License would be to refrain entirely from distribution of the
>   Program
> 
> Do you realize that the patent is unrelated with the program, but
> nevertheless the copyright license establishes conditions about what
> kind of patent licenses you may accept in order for you to have
> permission to distribute the program.
> 
> Why should restrictions through patents be unacceptable, but
> restrictions through hardware and software be acceptable.
> 
> Both are means to disrespect users' freedoms.

A patent prevents you from using the software in any way at all, while a
hardware restriction prevents you from using the software on that
particular hardware, but not on lots of other hardware.  Very big
difference.

> It is the duty of the FSF to defend these freedoms.  It's its public
> mission.  That's a publicly stated goal of the GPL, for anyone who
> cares to understand it, or miss it completely and then complain about
> changes in spirit.

I wouldn't call it a duty.  It is the chosen mission perhaps, but nobody
is making them do it.

> That's true.  Per the license, it's only who distributes the hardware
> to you that shouldn't impose such restrictions.

So what would happen if some company was to make software for a tivo and
released their binaries signed with some specific key, and they released
information on how to check this was signed with their key, and then
some other companies went and made tivo hardware and decided that they
would only allow code signed by the first companies key to run on it,
because that company had software which was acceptable to the
DMCA/RIAA/MPAA/etc and allowed them to get access to the hardware they
wanted to use in their box.  The second company now sells hardware to
make money, and the first company sells tv guide updates service to
people who want to use their software releases fully.

What does the GPL do now?  The software company still releases the
sources to the GPL software, but their binary releases are signed with a
key they don't give you.  They didn't provide you with any hardware, you
have to buy that from the hardware company that makes a product that
happens to run that software because it has the right bits of hardware
to record tv programs and such.  The hardware company put restrictions
on what software the box will run, although techicly the software
company that has the signing key could make lots of compatible software
for that particular locked down hardware, including vxworks or windows
based code if they chose to do so, while the hardware company just makes
hardware and decided to only allow software with the signature to run.
They didn't distribute any software, the buyer has to go get that from
the software company if they want the box to be useful (probably not a
good business plan for the majority of customers, but still possible in
theory).

> That's right.  But one of the obligations is to impose no further
> restrictions on the exercise of the rights.  What is "imposing a
> restriction"?  Installing the software in ROM isn't regarded as such,
> it's just a technical decision.  Installing the software in modifiable
> non-volatile storage, but denying the user the ability to change it,
> is regarded as imposing a restriction.  (note the "denying")  It is a
> matter of intent.
> 
> It's not because you only install say 32MB of RAM on the machine that
> you're denying the user the ability to run OOo on the machine.  But if
> you ship the computer with plenty of memory, but somehow configure the
> hardware or the operating system so as to prevent the user from
> upgrading an OOo that shipped with it, while you can still install
> that upgrade, then you're actively placing limits on the user's
> freedom WRT to that software, and an anti-tivoization clause would
> then stop you from distributing the software under these conditions.
> 
> I've never disputed that this is how they perceive it.
> 
> I've never disputed that GPLv2 serves this goal.
> 
> I still think GPLv3 serves this same goal, and better than v2.
> 
> But this is not what my participation here is about.
> 
> My participation here is about showing that GPLv3, and anti-tivozation
> in particular, don't violate the spirit of the defending users'
> freedoms WRT the covered software, such that the Free Software remains
> Free.

Well many people in the community disagrees, and you can't change their
mind on that it would seem, just as they can't change yours (or the FSFs
for that matter).  I would not be surprised to see some code forks when
the GPLv3 is finally released.

--
Len Sorensen

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 20, 2007, lsorense@csclub.uwaterloo.ca (Lennart Sorensen) wrote:

> On Sun, Jun 17, 2007 at 12:52:38AM -0300, Alexandre Oliva wrote:
>> Why should restrictions through patents be unacceptable, but
>> restrictions through hardware and software be acceptable.

>> Both are means to disrespect users' freedoms.

> A patent prevents you from using the software in any way at all,
> while a hardware restriction prevents you from using the software on
> that particular hardware, but not on lots of other hardware.  Very
> big difference.

So, one disrespects a lot, the other disrespects a little.  Is that
relevant, when the requirement is "no further restrictions"?

>> It is the duty of the FSF to defend these freedoms.  It's its public
>> mission.  That's a publicly stated goal of the GPL, for anyone who
>> cares to understand it, or miss it completely and then complain about
>> changes in spirit.

> I wouldn't call it a duty.  It is the chosen mission perhaps, but nobody
> is making them do it.

Everyone who donates to it does so understanding what the mission is.
Detracting from that mission would be failing the public commitment.

Sure, it may have been self-imposed in the beginning, but maybe not
even then.  At least in Brazil, foundations are started by an initial
donor, who determines its mission, and it has a legal obligation to
pursue that mission, and IIRC it cannot be changed except by a court
order, and even than within certain limits.

> So what would happen if some company was to make software for a tivo and
> released their binaries signed with some specific key, and they released
> information on how to check this was signed with their key, and then
> some other companies went and made tivo hardware and decided that they
> would only allow code signed by the first companies key to run on it,

I was pretty sure this had been covered in the section about technical
barriers to modification in the third draft's rationale, but I can't
find it right now.  http://gplv3.fsf.org/gpl3-dd3-rationale.pdf

Anyhow, the argument I read went like: if there's an agreement between
the parties to do this, then the copyright holder can probably enforce
the license regardless of the software and hardware distributor being
different parties, since the software is being distributed with
information whose purpose is to enable the hardware to deny the user
the freedom to run modified versions of the software.

However, if there's no such agreement, if the copyright holder has no
copyright claims over the hardware or works shipped in it, there's
nothing the copyright holder can do about it, and that's probably how
it should be, since a copyright license (!= contract) can't possibly
prohibit people from creating hardware limited in function, it can
only tell people that, in order for them to have permission to modify
or distribute the covered work, they must abide by certain conditions.
And if they don't want to abide by the conditions, and they don't
manage to obtain a license from the copyright holders that doesn't
impose conditions they can't accept, they just can't modify or
distribute the work.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[david]

On Wed, 20 Jun 2007, Alexandre Oliva wrote:

> On Jun 20, 2007, lsorense@csclub.uwaterloo.ca (Lennart Sorensen) wrote:
>>> It is the duty of the FSF to defend these freedoms.  It's its public
>>> mission.  That's a publicly stated goal of the GPL, for anyone who
>>> cares to understand it, or miss it completely and then complain about
>>> changes in spirit.
>
>> I wouldn't call it a duty.  It is the chosen mission perhaps, but nobody
>> is making them do it.
>
> Everyone who donates to it does so understanding what the mission is.
> Detracting from that mission would be failing the public commitment.

true, but selecting the GPL as the license for your project is not 
donating to the FSF.

David Lang

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 20, 2007, david@lang.hm wrote:

> On Wed, 20 Jun 2007, Alexandre Oliva wrote:
>> On Jun 20, 2007, lsorense@csclub.uwaterloo.ca (Lennart Sorensen) wrote:
>>>> It is the duty of the FSF to defend these freedoms.  It's its public
>>>> mission.  That's a publicly stated goal of the GPL, for anyone who
>>>> cares to understand it, or miss it completely and then complain about
>>>> changes in spirit.
>> 
>>> I wouldn't call it a duty.  It is the chosen mission perhaps, but nobody
>>> is making them do it.
>> 
>> Everyone who donates to it does so understanding what the mission is.
>> Detracting from that mission would be failing the public commitment.

> true, but selecting the GPL as the license for your project is not
> donating to the FSF.

Oh, that's what you meant.  Indeed, absolutely not.

The GPL is "just" a set of permissions you, as an author, grant to
anyone who comes across your program.

Whether you share FSF's goals or not, you can do that.

If you share FSF's goals of not only respecting users' freedoms, but
also defending them as much as deemed legally possible under copyright
law, you can also offer your code under any later version of the GPL,
such that it remains usable by the community who cares about this.

In theory, this shouldn't be a problem for anyone who chose the GPLv2,
since all of the permissions granted by GPLv3 are granted by GPLv2,
and this is how it should be.  The difference is that GPLv3 plugs some
holes that were found in GPLv2, in a similar way that GPLv2 plugged
holes found in GPLv1, and GPL "plugs holes" in LGPL, which "plugs
holes" in other even more permissive licenses.

Each GPL revision is expected to plug holes ("address new problems",
as in the legal terms of GPLv2) that might enable licensees to deny
other licensees the rights you meant to grant them.

This will necessarily make each revision incompatible with the
previous, for being stricter, thus imposing further restrictions, even
if only by removing exploitable ambiguities.  This should have been
clear since GPLv1, anyone who understands the goals of the GPL and
with enough foresight to understand the recommendation of permitting
relicensing under newer versions should be able to see this.

So, since new restrictions are always on licensees' ways to deny other
licensees the enjoyment of the permissions you meant to grant them, if
you mean to permit people to use your work in the ways permitted by
GPLv2, not permitting them to be used in GPLv3 software amounts to
pure selfishness: "if I won't get to use your code, you don't get to
use mine."  Tit-for-tat, for sure, but certainly not in the spirit of
sharing clearly established early in the preamble of every version of
the GPL.

Permitting such relicensing wouldn't deny anyone any freedom, and it
wouldn't create any obligations whatsoever for the licensor.  Whoever
wanted to use the work under the more liberal terms of the earlier
version of the GPL under which the work was licensed still could: this
license can't be unilaterally revoked, not by you, not by anyone else.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Lennart Sorensen]

On Wed, Jun 20, 2007 at 05:52:40PM -0300, Alexandre Oliva wrote:
> On Jun 20, 2007, lsorense@csclub.uwaterloo.ca (Lennart Sorensen) wrote:
> > A patent prevents you from using the software in any way at all,
> > while a hardware restriction prevents you from using the software on
> > that particular hardware, but not on lots of other hardware.  Very
> > big difference.
> 
> So, one disrespects a lot, the other disrespects a little.  Is that
> relevant, when the requirement is "no further restrictions"?

What about the freedom to buy devices with certified code on it, while
still being able to look through the source code and verify for yourself
that it is correct and not full of bugs?  Would it be better if the
devices that have to be certified and locked down used secret code so
that the purchaser can't verify the code?

Apparently the only restrictions ever permitted are the ones the FSF
thinks of.

> > So what would happen if some company was to make software for a tivo and
> > released their binaries signed with some specific key, and they released
> > information on how to check this was signed with their key, and then
> > some other companies went and made tivo hardware and decided that they
> > would only allow code signed by the first companies key to run on it,
> 
> I was pretty sure this had been covered in the section about technical
> barriers to modification in the third draft's rationale, but I can't
> find it right now.  http://gplv3.fsf.org/gpl3-dd3-rationale.pdf

So really what the GPL v3 wants to have is to make sure that the user
can reproduce from the sources a bit for bit identical copy of the
binaries?  Too bad compilers that put time stamps and such into the
binary would make that imposible.  I don't think there is any way that
can be written into the GPL that can prevent all loop holes for how to
make signed binaries.

> Anyhow, the argument I read went like: if there's an agreement between
> the parties to do this, then the copyright holder can probably enforce
> the license regardless of the software and hardware distributor being
> different parties, since the software is being distributed with
> information whose purpose is to enable the hardware to deny the user
> the freedom to run modified versions of the software.

There doesn't have to be an agreement.  The software company could just
release specs for a hardware design and let others freely go and build
them from that design.

> However, if there's no such agreement, if the copyright holder has no
> copyright claims over the hardware or works shipped in it, there's
> nothing the copyright holder can do about it, and that's probably how
> it should be, since a copyright license (!= contract) can't possibly
> prohibit people from creating hardware limited in function, it can
> only tell people that, in order for them to have permission to modify
> or distribute the covered work, they must abide by certain conditions.
> And if they don't want to abide by the conditions, and they don't
> manage to obtain a license from the copyright holders that doesn't
> impose conditions they can't accept, they just can't modify or
> distribute the work.

But if the hardware ships with only code that simply waits for the user
to provide some code for it to isntall (which has to be signed in a way
the hardware likes), then the hardware has nothing to do with the
license of the software.

The signed binaries from the service provider/software developer on the
other hand is GPL and the sources are released with changes.  They just
happen to sign their binaries in a way that allows them to install on
the hardware in question.  It could also install on hardware that
doesn't check the signature as long as it is functionally identical
otherwise.

I hope no one does this, but I still don't see how the GPLv3 draft deals
with this case, or even how it could deal with it.

--
Len Sorensen

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 21, 2007, lsorense@csclub.uwaterloo.ca (Lennart Sorensen) wrote:

> Apparently the only restrictions ever permitted are the ones the FSF
> thinks of.

Where does this nonsensical idea come from?  How does it follow that,
from FSF offering a licensing option to authors, you conclude that
nobody could ever establish whatever other restrictions they liked?

> So really what the GPL v3 wants to have is to make sure that the user
> can reproduce from the sources a bit for bit identical copy of the
> binaries?

No, this is not enough to enable someone to adapt the software to
one's own needs.

> Too bad compilers that put time stamps and such into the
> binary would make that imposible.

This would be the copyright author imposing such a restriction, not
the software distributor.

> I don't think there is any way that can be written into the GPL that
> can prevent all loop holes for how to make signed binaries.

Which is one possible reason to explain why the FSF switched to the
'Installation Information' approach.

> There doesn't have to be an agreement.  The software company could just
> release specs for a hardware design and let others freely go and build
> them from that design.

Aah, so the software company has designed a mechanism to restrict
users' freedoms, and is just leaving it up to third parties to
complete the implementation?  I think these design documents could be
used in a court to prove intent to impose restrictions on the users,
but IANAL.

>> However, if there's no such agreement, if the copyright holder has no
>> copyright claims over the hardware or works shipped in it, there's
>> nothing the copyright holder can do about it, and that's probably how
>> it should be, since a copyright license (!= contract) can't possibly
>> prohibit people from creating hardware limited in function, it can
>> only tell people that, in order for them to have permission to modify
>> or distribute the covered work, they must abide by certain conditions.
>> And if they don't want to abide by the conditions, and they don't
>> manage to obtain a license from the copyright holders that doesn't
>> impose conditions they can't accept, they just can't modify or
>> distribute the work.

> But if the hardware ships with only code that simply waits for the user
> to provide some code for it to isntall (which has to be signed in a way
> the hardware likes), then the hardware has nothing to do with the
> license of the software.

Correct.  That's pretty much what I said, isn't it?

> I hope no one does this, but I still don't see how the GPLv3 draft deals
> with this case, or even how it could deal with it.

It doesn't, and it probably shouldn't.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Saturday 16 June 2007 15:27:37 Alexandre Oliva wrote:
> On Jun 16, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> > I don't see how TiVO has done this. They have placed no restrictions on
> > *modification* at all. What they have done is placed a restriction on
> > *REPLACEMENT* of the program.
>
> Technicality.  In order for the software to remain free (which is what
> the GPL is all about), the user must not be stopped from adapting the
> software to suit his needs and running it for any purpose.  TiVo
> places restrictions on it.  It's really this simple.

Your arguments are all based on technicalities, so why are you complaining 
when I do the same?

As it stands there is *NOTHING* that singular distinction makes all the 
difference in the world. What you are arguing - based on your *BELIEF* that 
such *REPLACEMENT* is a modification. 

By the way, Alexandre, I'm not so much of an *IDIOT* as to believe that you 
don't understand the difference. You are arguing about it because you *WANT* 
the difference to not exist. You are arguing about it because it makes your 
argument that what TiVO did broke the "spirit" of the license. If you want 
I'll go dig out the exact place where RMS said that he didn't care about 
hardware. 

You want another "technicality"? How about one that you *AGREED* is valid? 
That your right to configure a device ends at the point where it connects to 
a network? Well, unless you want to sacrifice *ALL* the stuff that makes a 
TiVO actually worth using, you *HAVE* to connect it to their network. At that 
point *ALL* of its configuration details - yes, even the Operating System - 
fall under their control. In the US there are laws that restrict this right 
when applied to "telecommunications" companies - but TiVO *isn't* 
a "telecommunications" company.

> And then, TiVo doesn't really prohibit replacement.  You can replace
> it as much as you like; just not as conveniently as TiVo can replace
> it.  And then, if you do, it won't run, because it's not signed with a
> key that they omit from the source code.  And they do this in order to
> prevent the user from changing the behavior of the Free Software that
> they use, while they keep this ability to themselves.

If your argument is that the final output binary is created by combining the 
signing key and an interrim binary then you *MIGHT* have a point. The simple 
fact is that that argument depends on whether the kernel itself is modified 
by the signing process or if the signing process generates a separate 
signature which is then verified as part of the boot process.

> If these are not restrictions on the freedoms that the GPL is designed
> to protect to ensure that Free Software remains Free for all its
> users, I don't know what is.

"Free as in beer" is the phrasing used, I believe. I see nothing in that TiVO 
has done that negates this. I do disagree with it - if I buy a TiVO box, I 
own it and should be able to do what I want with it. However, this does not 
negate the fact that it does connect to their network, and as a device that 
does such they are allowed to configure it in *ANY* manner they choose. What 
you and the FSF are trying to do is strip that right from them.

If you have such a respect for peoples freedoms - and I don't doubt that you 
actually believe you do - then why are you stripping freedoms from people?

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 16, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> On Saturday 16 June 2007 15:27:37 Alexandre Oliva wrote:
>> On Jun 16, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
>> > I don't see how TiVO has done this. They have placed no restrictions on
>> > *modification* at all. What they have done is placed a restriction on
>> > *REPLACEMENT* of the program.
>> 
>> Technicality.  In order for the software to remain free (which is what
>> the GPL is all about), the user must not be stopped from adapting the
>> software to suit his needs and running it for any purpose.  TiVo
>> places restrictions on it.  It's really this simple.

> Your arguments are all based on technicalities, so why are you complaining 
> when I do the same?

My arguments are based on the intent behind the license, its spirit.

You keep falling back to legal technicalities, that have zero to do
with the interpretation of the intent.

That's why.


http://en.wikipedia.org/wiki/Letter_and_spirit_of_the_law

> As it stands there is *NOTHING* that singular distinction makes all the 
> difference in the world. What you are arguing - based on your *BELIEF* that 
> such *REPLACEMENT* is a modification. 

Maybe modification is not the best word, because it carries a lot of
legal background from copyright law.

How about adaptation.  From freedom #1, freedom to study the software
and adapt it to your needs.  Do you see how tivoization imposes an
artificial restriction to this freedom?

> If you want I'll go dig out the exact place where RMS said that he
> didn't care about hardware.

This is still true.  This is not about the hardware.  This is about
the software, and how the user is stopped from adapting it to her own
needs, while the vendor saves this prerogative to itself.

> That your right to configure a device ends at the point where it
> connects to a network? Well, unless you want to sacrifice *ALL* the
> stuff that makes a TiVO actually worth using, you *HAVE* to connect
> it to their network.

So, if you visit www.fsfla.org, I 0w|\| your computer?

If you join a bit torrent, I can replace the operating system on your
computer?

Sorry, I don't buy that.  You're leaving something out of this
picture, and that's probably quite important.

>> If these are not restrictions on the freedoms that the GPL is designed
>> to protect to ensure that Free Software remains Free for all its
>> users, I don't know what is.

> "Free as in beer" is the phrasing used, I believe.

Huh?  Are you implying that the Free Software foundation wrote this
meaning "zero cost"?

> If you have such a respect for peoples freedoms - and I don't doubt
> that you actually believe you do - then why are you stripping
> freedoms from people?

Because they're disrespecting others' freedoms.  Freedoms aren't
absolute.  One's freedom ends where another's freedom starts.
Tivoization exceeds the hardware manufacturer's freedoms and
disrespects users' freedoms and disrespect some author's ethical
intent.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Saturday 16 June 2007 21:49:56 Alexandre Oliva wrote:
> On Jun 16, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> > On Saturday 16 June 2007 15:27:37 Alexandre Oliva wrote:
> >> On Jun 16, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> >> > I don't see how TiVO has done this. They have placed no restrictions
> >> > on *modification* at all. What they have done is placed a restriction
> >> > on *REPLACEMENT* of the program.
> >>
> >> Technicality.  In order for the software to remain free (which is what
> >> the GPL is all about), the user must not be stopped from adapting the
> >> software to suit his needs and running it for any purpose.  TiVo
> >> places restrictions on it.  It's really this simple.
> >
> > Your arguments are all based on technicalities, so why are you
> > complaining when I do the same?
>
> My arguments are based on the intent behind the license, its spirit.

But each of those arguments is based on a technicality. By your reasoning I 
could kill everybody living in the middle east to stop the wars there and not 
be wrong - after all, you say "But I'm using those technicalities to show the 
letter and spirit of the license" - ie: "The ends justify the means".

> You keep falling back to legal technicalities, that have zero to do
> with the interpretation of the intent.
>
> That's why.
>
>
> http://en.wikipedia.org/wiki/Letter_and_spirit_of_the_law

Do you know how many lawyers make a living because the "spirit" of a law has 
no legal weight?

> > As it stands there is *NOTHING* that singular distinction makes all the
> > difference in the world. What you are arguing - based on your *BELIEF*
> > that such *REPLACEMENT* is a modification.
>
> Maybe modification is not the best word, because it carries a lot of
> legal background from copyright law.
>
> How about adaptation.  From freedom #1, freedom to study the software
> and adapt it to your needs.  Do you see how tivoization imposes an
> artificial restriction to this freedom?

Nothing stopping people from doing that with the GPL'd software running on a 
TiVO. 

> > If you want I'll go dig out the exact place where RMS said that he
> > didn't care about hardware.
>
> This is still true.  This is not about the hardware.  This is about
> the software, and how the user is stopped from adapting it to her own
> needs, while the vendor saves this prerogative to itself.
>
> > That your right to configure a device ends at the point where it
> > connects to a network? Well, unless you want to sacrifice *ALL* the
> > stuff that makes a TiVO actually worth using, you *HAVE* to connect
> > it to their network.
>
> So, if you visit www.fsfla.org, I 0w|\| your computer?
>
> If you join a bit torrent, I can replace the operating system on your
> computer?
>
> Sorry, I don't buy that.  You're leaving something out of this
> picture, and that's probably quite important.

Nope. Because I'm connecting the the *INTERNET*. The internet is not owned by 
any one person or "legal entity" - therefore there is nobody that can demand 
a certain configuration. Note that I also made it a point to mention that it 
only applies to certain classes of networks - in the US there are laws that 
remove the "complete control over configuration" from telecommunications 
companies. But get a cable-modem in the US and your ISP has the right to 
configure it in *ANY* way they choose.

The TiVO service runs as a network - and a non-public one at that. They own 
the network, they control what hardware and with what configurations is 
allowed to connect. Whats more is that they have the right to actively 
control that configuration.

You do realize, Alexandre, that you can't make me look stupid by just cutting 
out a part of a statement I've made and making silly comments about it. If 
you are going to quote something I've said, make sure you quote the *ENTIRE* 
effective part and not just the bit you think will make you look smart. All 
it does is make you look like an ass.

> >> If these are not restrictions on the freedoms that the GPL is designed
> >> to protect to ensure that Free Software remains Free for all its
> >> users, I don't know what is.
> >
> > "Free as in beer" is the phrasing used, I believe.
>
> Huh?  Are you implying that the Free Software foundation wrote this
> meaning "zero cost"?

Nope. I was making sure that you understood your own propaganda. "Free as in 
beer" - if I get a free beer I'm getting the beer, not the glass. If you 
aren't intelligent enough to understand what I'm saying: I get the software 
and *ALL* rights to it that everyone *BUT* the licensor has under the GPL. 
What you are doing is saying "It is what is said, but not what is meant."

The funniest part of it is that you are claiming that the "spirit" of the GPL 
is to force each licensee to give up *MORE* rights than they are asked to. In 
other words... TiVO is a licensee of the kernel - they received certain 
rights through the GPL that they are required to pass along to anyone they 
give a copy to. Those rights are passed on. What they don't do is allow a 
copy of the "covered work" to run on the hardware - copies that might do 
things like allow people to break the copyright on content that the box can 
create copies of.

The thing is, I already know your answer. It's in the mail that this is a 
reply to. The "spirit of the law" is something everyone wishes people would 
follow. But each person has their own interpretation of what the "Spirit" is 
and there is no real way to know what the "spirit" of a law is. Because 
language is such a slippery beast even the writings of the people that wrote 
the law can't define what the "spirit" of a law is. I've repeated myself to 
many times about the unreliability of human testimony on the topic. So what 
is being done by you and everyone else that is part of the FSF is doing is 
saying "Our view of the 'Spirit' of the license is the only correct one. 
Everyone else is wrong."

You are, of course, entitled to your opinion. The FSF is also entitled to it's 
opinion. But as has been shown, that opinion may not be correct. And it is a 
simple to prove fact that a large number of people hold a different opinion. 

> > If you have such a respect for peoples freedoms - and I don't doubt
> > that you actually believe you do - then why are you stripping
> > freedoms from people?
>
> Because they're disrespecting others' freedoms.  Freedoms aren't
> absolute.  One's freedom ends where another's freedom starts.
> Tivoization exceeds the hardware manufacturer's freedoms and
> disrespects users' freedoms and disrespect some author's ethical
> intent.

By this logic I could release software under a license that says "if you want 
to use this in a commercial product you have to send any person who buys the 
product a copy of the complete technical specifications - including any 
cryptographic keys - on request." And claim that its protecting the rights of 
the end-user in regards to my software. - Oh, wait, the FSF beat me to it 
with the GPL3!

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 17, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> But each of those arguments is based on a technicality.

They're based on the Free Software definition, that establishes the
four freedoms that the GPL was designed to respect and defend.

Each version of the GPL may miss the mark.  But this doesn't mean
that's not their spirit.

> Do you know how many lawyers make a living because the "spirit" of a law has 
> no legal weight?

Yes.  What's your point?

All I'm trying to show is that the tivoization provision in GPLv3 is
not a departure from the spirit of the GPL.

Is this so hard to understand?


I'm not trying to say why Linus and others chose the GPLv2.

I'm not trying to determine what their motivations were.

I'm not trying to force them to change to GPLv3.

I'm not trying to convince them that tivozation is a bad thing.

I'm only trying to show that anti-tivozation is in line with the
spirit of the GPL.


tivoization, which means to restrict a user's ability to adapt the
software to their own needs and run it for any purpose, while the
hardware manufacturer keeps this to itself, is against the spirit of
the GPL.

Not whatever reasons the Linux developers had to release their code
under GPLv2.  But the spirit that the authors of the GPL tried to
encode in it.

Is this so difficult to accept?

>> > That your right to configure a device ends at the point where it
>> > connects to a network? Well, unless you want to sacrifice *ALL* the
>> > stuff that makes a TiVO actually worth using, you *HAVE* to connect
>> > it to their network.

>> So, if you visit www.fsfla.org, I 0w|\| your computer?

> Nope. Because I'm connecting the the *INTERNET*.

Is the connection with the TiVo network not through some other
carrier too?

> The TiVO service runs as a network - and a non-public one at that. They own 
> the network, they control what hardware and with what configurations is 
> allowed to connect. Whats more is that they have the right to actively 
> control that configuration.

As long as this doesn't violate any other laws or agreements they've
entered, that is.  And this includes license agreements.

> You do realize, Alexandre, that you can't make me look stupid by
> just cutting out a part of a statement I've made and making silly
> comments about it.

Didn't mean to, sorry if it seemed that way.  I still don't quite
understand the distinction you're trying to make.

> The funniest part of it is that you are claiming that the "spirit"
> of the GPL is to force each licensee to give up *MORE* rights than
> they are asked to.

No, the GPL doesn't force anything.  It can't.  All it does is to
demand respect for others' freedoms in case one decides to modify or
distribute the software.  It's only if you do modify or distribute the
software that you must respect others' freedoms.  And TiVo does
distribute the software.  But it doesn't respect the freedoms.

It might as well stop distributing the software.

> What they don't do is allow a 
> copy of the "covered work" to run on the hardware

It's not just that.  They actively stop you from being able to do so.
They do this so as to prevent you from changing the behavior of the
program that runs on that box.  They disrespect the freedoms to adapt
the program and to run it for any purpose.

> By this logic I could release software under a license that says "if
> you want to use this in a commercial product you have to send any
> person who buys the product a copy of the complete technical
> specifications - including any cryptographic keys - on request."

And, guess what, you *can* do that.  And it's up to the hardware
manufacturer to decide whether they want to use distribute your
software along with the hardware or not.

Whether this would qualify as a Free Software license, and whether it
would be in the spirit of the GPL, is a separate issue.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Saturday 16 June 2007 23:31:00 Alexandre Oliva wrote:
> On Jun 17, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> > But each of those arguments is based on a technicality.
>
> They're based on the Free Software definition, that establishes the
> four freedoms that the GPL was designed to respect and defend.

And what gives you or the FSF to define what "Free Software" is? What makes 
the definition you are using is the correct one?

But it doesn't matter. You're backpedalling, tossing up a smokescreen because 
you've been caught working under a double-standard.

> Each version of the GPL may miss the mark.  But this doesn't mean
> that's not their spirit.
>
> > Do you know how many lawyers make a living because the "spirit" of a law
> > has no legal weight?
>
> Yes.  What's your point?

That you were arguing a pitifully bad point.

> All I'm trying to show is that the tivoization provision in GPLv3 is
> not a departure from the spirit of the GPL.
>
> Is this so hard to understand?

Nope. As I've stated before it doesn't matter that you believe it. What 
matters is that there is no single, definable "spirit" of the license. 
The "spirit" is what each person who places their work under the license 
believes it to be.

>
> I'm not trying to say why Linus and others chose the GPLv2.
>
> I'm not trying to determine what their motivations were.
>
> I'm not trying to force them to change to GPLv3.
>
> I'm not trying to convince them that tivozation is a bad thing.

But you have done this multiple times. You may not have been trying to, but 
you were.

> I'm only trying to show that anti-tivozation is in line with the
> spirit of the GPL.

With *YOUR* view of what the spirit is. 

>
> tivoization, which means to restrict a user's ability to adapt the
> software to their own needs and run it for any purpose, while the
> hardware manufacturer keeps this to itself, is against the spirit of
> the GPL.

I'm a firm believer in letting people hang themselves, but this is a bit much. 
The TiVO company didn't do that. They kept the ability to *REPLACE* the 
version on the device that connects to their network to themselves. You have 
access to the source code TiVO uses, complete with their modifications... You 
can modify it in any way you choose *AND* you can distribute the code 
yourself. Hell, you can even *RUN* it for any purpose you want. What you 
can't do is replace the functional code on the device connected to their 
network.

> Not whatever reasons the Linux developers had to release their code
> under GPLv2.  But the spirit that the authors of the GPL tried to
> encode in it.
>
> Is this so difficult to accept?
>
> >> > That your right to configure a device ends at the point where it
> >> > connects to a network? Well, unless you want to sacrifice *ALL* the
> >> > stuff that makes a TiVO actually worth using, you *HAVE* to connect
> >> > it to their network.
> >>
> >> So, if you visit www.fsfla.org, I 0w|\| your computer?
> >
> > Nope. Because I'm connecting the the *INTERNET*.
>
> Is the connection with the TiVo network not through some other
> carrier too?

But your server doesn't run the internet. TiVO may use phone lines to connect 
a device to their server (and this is an example - I don't know how TiVO 
devices actually connect) but the network being connected to has a single 
owner who can set such terms. 

I'll repeat, in full, my earlier examples of this:
The first:
I buy a cable modem. Until the second I connect the cable-line to it so I can 
get a connection to the internet I can configure it in whatever manner I 
please. The second the line is connected, even though I *OWN* the hardware, I 
lose all control over its configuration.

The second:
I buy a DSL modem. Until I want to actually connect to the internet it can 
have whatever settings I want it to have. The second I want to connect to the 
internet it has to be configured the way that the ISP wants.
 
> > The TiVO service runs as a network - and a non-public one at that. They
> > own the network, they control what hardware and with what configurations
> > is allowed to connect. Whats more is that they have the right to actively
> > control that configuration.
>
> As long as this doesn't violate any other laws or agreements they've
> entered, that is.  And this includes license agreements.

And if the license doesn't explicitly state something as being a violation of 
its terms then it doesn't matter.

> > You do realize, Alexandre, that you can't make me look stupid by
> > just cutting out a part of a statement I've made and making silly
> > comments about it.
>
> Didn't mean to, sorry if it seemed that way.  I still don't quite
> understand the distinction you're trying to make.
>
> > The funniest part of it is that you are claiming that the "spirit"
> > of the GPL is to force each licensee to give up *MORE* rights than
> > they are asked to.
>
> No, the GPL doesn't force anything.  It can't.  All it does is to
> demand respect for others' freedoms in case one decides to modify or
> distribute the software.  It's only if you do modify or distribute the
> software that you must respect others' freedoms.  And TiVo does
> distribute the software.  But it doesn't respect the freedoms.

I agree with you. If I buy the hardware I should have control over it - 
period. However, it is complying with the license.

> It might as well stop distributing the software.

I actually disagree with you. My feeling is that any exposure to open-source 
is a good thing.

> > What they don't do is allow a
> > copy of the "covered work" to run on the hardware
>
> It's not just that.  They actively stop you from being able to do so.
> They do this so as to prevent you from changing the behavior of the
> program that runs on that box.  They disrespect the freedoms to adapt
> the program and to run it for any purpose.

But that freedom isn't guaranteed in the license. It also isn't very well 
stated. By the logic of that statement, I should be able to adapt the Linux 
Kernel to be a clone of "bc". (Well, I can, but that isn't the point)

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 17, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> On Saturday 16 June 2007 23:31:00 Alexandre Oliva wrote:
>> On Jun 17, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
>> > But each of those arguments is based on a technicality.
>> 
>> They're based on the Free Software definition, that establishes the
>> four freedoms that the GPL was designed to respect and defend.

> And what gives you or the FSF to define what "Free Software" is? What makes 
> the definition you are using is the correct one?

Huh?  Would you expect the *Free* *Software* Foundation to use the term
*Free* *Software* present in the license it wrote to mean what?

> But it doesn't matter. You're backpedalling, tossing up a smokescreen because 
> you've been caught working under a double-standard.

Heh.  That's hilarious.

> Nope. As I've stated before it doesn't matter that you believe it. What 
> matters is that there is no single, definable "spirit" of the license. 
> The "spirit" is what each person who places their work under the license 
> believes it to be.

You're mixing two separate issues.

One thing is the intent behind the writing of the license.  This is
what the spirit of the license is.

Another thing is how the copyright holder of a work perceived that
license, and the motivations for his choice of that license.  I don't
think this has a name.  Let's call this spirit of the licensing, for
the sake of the argument.

What I've been talking about is not the spirit of the licensing.  I
respect that, even when I don't agree with it.

But when people claim the GPL is changing its spirit, they're accusing
the FSF of changing the spirit of the license.  And this hasn't
happened.  This is the point I'm standing for.

Does this clarify the issue?

>> I'm not trying to say why Linus and others chose the GPLv2.
>> 
>> I'm not trying to determine what their motivations were.
>> 
>> I'm not trying to force them to change to GPLv3.
>> 
>> I'm not trying to convince them that tivozation is a bad thing.

> But you have done this multiple times. You may not have been trying to, but 
> you were.

You mean I've done all of the above multiple times?  Show me?

Odds of success for the last one are pretty high, because the
discussion somehow sidetracked into that and I've probably been sloppy
about it, but as for the other points, I very much doubt it you'll
find me doing any of them.  100% sure you won't find anything about
forcing anyone to change to GPLv3.

I've just realized that "determine" above is ambiguous.  I meant it as
"decide", rather than "understand".  I was definitely trying to
understand their motivations, once the debate moved onto that front as
well.

>> I'm only trying to show that anti-tivozation is in line with the
>> spirit of the GPL.

> With *YOUR* view of what the spirit is. 

Why, sure.  And given how close I am to the FSFs and how closely I
understand the reasoning behind the GPL, do you really think my view
does not match the intent of the FSF for the GPL?

>> tivoization, which means to restrict a user's ability to adapt the
>> software to their own needs and run it for any purpose, while the
>> hardware manufacturer keeps this to itself, is against the spirit of
>> the GPL.

> The TiVO company didn't do that. They kept the ability to *REPLACE*
> the version on the device that connects to their network to
> themselves. You have access to the source code TiVO uses, complete
> with their modifications... You can modify it in any way you choose
> *AND* you can distribute the code yourself. Hell, you can even *RUN*
> it for any purpose you want.

How is this enough to adapt the software to my needs and run it for
any purpose?

How can you possibly claim they're not imposing restrictions on my
abilities to adapt the software to my needs (freedom #1) and run the
software for any purpose (freedom #0), if that's the whole point
behind their technical measures?

>> Is the connection with the TiVo network not through some other
>> carrier too?

> But your server doesn't run the internet.

But it runs my home network, and you've connected to it.  Now what?

I'm sure I'm still missing something in your characterization of the
situation about networks, but I'm not sure I care enough to pursue
this point.  Feel free to drop it, I don't think it's relevant for the
discussion on whether GPLv3 changes the spirit of the GPL.

>> > The TiVO service runs as a network - and a non-public one at that. They
>> > own the network, they control what hardware and with what configurations
>> > is allowed to connect. Whats more is that they have the right to actively
>> > control that configuration.

>> As long as this doesn't violate any other laws or agreements they've
>> entered, that is.  And this includes license agreements.

> And if the license doesn't explicitly state something as being a
> violation of its terms then it doesn't matter.

Not quite.  Copyright licenses are to be interpreted restrictively.
Unless it states you can do something, you can't.

>> And TiVo does distribute the software.  But it doesn't respect the
>> freedoms.

> I agree with you. If I buy the hardware I should have control over it - 
> period. However, it is complying with the license.

>> It might as well stop distributing the software.

> I actually disagree with you. My feeling is that any exposure to open-source 
> is a good thing.

Oh, but that's an option they have nevertheless.  GPLv3 or not.

>> > What they don't do is allow a
>> > copy of the "covered work" to run on the hardware

>> It's not just that.  They actively stop you from being able to do so.
>> They do this so as to prevent you from changing the behavior of the
>> program that runs on that box.  They disrespect the freedoms to adapt
>> the program and to run it for any purpose.

> But that freedom isn't guaranteed in the license.

I see you've moved away from the spirit and back to the legal terms.
My answer to this is: maybe.  There's no consensus about it.  And
then, I've never claimed it was.  I am not a lawyer to make such
claims.

What I've been trying to say is that disrespecting this freedom was
not in line with the spirit of the GPL, the spirit of respecting and
defending the 4 freedoms.  So, when GPLv3 adds provisions to make this
practice clearly not permitted, it's not changing the spirit.  And it
might not even be changing the legal effects in this regard.

> By the logic of that statement, I should be able to adapt the Linux
> Kernel to be a clone of "bc". (Well, I can, but that isn't the
> point)

You can.  I can appreciate that this isn't the point, but I don't see
what the point is that you were trying to make.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Graham Murray]

Daniel Hazelton <dhazelton@enter.net> writes:

> The second:
> I buy a DSL modem. Until I want to actually connect to the internet it can 
> have whatever settings I want it to have. The second I want to connect to the 
> internet it has to be configured the way that the ISP wants.

But only those configurations which enable it to work, such as the
modulation type, PPPOE/PPPOA, VPI/VCI, PPP username and password
etc. You still have control over the rest of the configuration such as
NAT/No NAT, LAN IP address(es), firewall, MTU, QoS, etc.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Michael Poole]

Daniel Hazelton writes:

> But your server doesn't run the internet. TiVO may use phone lines to connect 
> a device to their server (and this is an example - I don't know how TiVO 
> devices actually connect) but the network being connected to has a single 
> owner who can set such terms. 
> 
> I'll repeat, in full, my earlier examples of this:
> The first:
> I buy a cable modem. Until the second I connect the cable-line to it so I can 
> get a connection to the internet I can configure it in whatever manner I 
> please. The second the line is connected, even though I *OWN* the hardware, I 
> lose all control over its configuration.
>
> The second:
> I buy a DSL modem. Until I want to actually connect to the internet it can 
> have whatever settings I want it to have. The second I want to connect to the 
> internet it has to be configured the way that the ISP wants.

Jung va gur jbeyq znxrf lbh guvax gurer vf n hfrshy nanybtl
orgjrra pbzzhavpngvba fgnaqneqf naq pbclevtug yvprafrf?

One moment, let me retune.

What in the world makes you think there is a useful analogy
between communication standards and copyright licenses?

Neither law nor common sense give much common ground to the two,
except in the general sense of two parties interacting.  One is a set
of rules so the two can interact through some information channel.
The other is a set of rules so that one can exploit a creative work
developed by the other.

I suppose that you think it is acceptable for someone to offer access
to binary and source versions of GPLed software (with or without
modifications from commonly available versions) -- but only on the
condition that people never download the source versions?  That
certainly corresponds to the idea that Tivo can keep proprietary
extensions to the kernel if Tivo's customers want to connect to Tivo's
network services.

Your reliance on counterfactual arguments severely undermines your
position -- whether the fiction is how Tivo devices connect (a quick
search on Google indicates that Tivo recommends a broadband Ethernet
connection rather than a phone line) or that we should analyze based
on DRM signatures distributed separately from the kernel (when they
are not).  We are arguing about the universe we inhabit, not some
alternative where the GPL might actually be the Groundhog Petting
License.

Michael Poole

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Sunday 17 June 2007 09:54:39 Michael Poole wrote:
> Daniel Hazelton writes:
> > But your server doesn't run the internet. TiVO may use phone lines to
> > connect a device to their server (and this is an example - I don't know
> > how TiVO devices actually connect) but the network being connected to has
> > a single owner who can set such terms.
> >
> > I'll repeat, in full, my earlier examples of this:
> > The first:
> > I buy a cable modem. Until the second I connect the cable-line to it so I
> > can get a connection to the internet I can configure it in whatever
> > manner I please. The second the line is connected, even though I *OWN*
> > the hardware, I lose all control over its configuration.
> >
> > The second:
> > I buy a DSL modem. Until I want to actually connect to the internet it
> > can have whatever settings I want it to have. The second I want to
> > connect to the internet it has to be configured the way that the ISP
> > wants.
>
> Jung va gur jbeyq znxrf lbh guvax gurer vf n hfrshy nanybtl
> orgjrra pbzzhavpngvba fgnaqneqf naq pbclevtug yvprafrf?
>
> One moment, let me retune.
>
> What in the world makes you think there is a useful analogy
> between communication standards and copyright licenses?

I don't. I was *REPEATING* an example of how TiVO has a *RIGHT* to change the 
kernel or any other facet of the device connecting to their network. That 
right *ISN'T* tied to copyright - as you have stated. Since it isn't, why is 
the FSF trying to mandate that it is with the tivoization clauses in GPLv3?

> Neither law nor common sense give much common ground to the two,
> except in the general sense of two parties interacting.  One is a set
> of rules so the two can interact through some information channel.
> The other is a set of rules so that one can exploit a creative work
> developed by the other.
>
> I suppose that you think it is acceptable for someone to offer access
> to binary and source versions of GPLed software (with or without
> modifications from commonly available versions) -- but only on the
> condition that people never download the source versions?  That
> certainly corresponds to the idea that Tivo can keep proprietary
> extensions to the kernel if Tivo's customers want to connect to Tivo's
> network services.

Nope. Because that isn't a right they have that is disconnected from copyright 
law. Or did you not read the entire post and just decide to try and make me 
look stupid?

> Your reliance on counterfactual arguments severely undermines your
> position -- whether the fiction is how Tivo devices connect (a quick
> search on Google indicates that Tivo recommends a broadband Ethernet
> connection rather than a phone line) or that we should analyze based
> on DRM signatures distributed separately from the kernel (when they
> are not).  We are arguing about the universe we inhabit, not some
> alternative where the GPL might actually be the Groundhog Petting
> License.

At no point have I claimed to know how TiVO was accomplishing the "connection 
to the service" or how they were handling the signing of the kernel. If the 
final binary is altered by the signing process then, IMHO, the signed kernel 
and everything involved in creating it constitutes a "derived work".

My argument is that, as much as people want it to be different, in mandating 
that people give up rights that are (potentially) disconnected from the 
copyright - in order to "Defend the Four Freedoms" and keep "all users of the 
licensed work equal in their freedoms" - the GPLv3 is flawed and potentially 
unenforceable. 

DRH

> Michael Poole



-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Michael Poole]

Daniel Hazelton writes:

> On Sunday 17 June 2007 09:54:39 Michael Poole wrote:
>> Daniel Hazelton writes:
>> > But your server doesn't run the internet. TiVO may use phone lines to
>> > connect a device to their server (and this is an example - I don't know
>> > how TiVO devices actually connect) but the network being connected to has
>> > a single owner who can set such terms.
>> >
>> > I'll repeat, in full, my earlier examples of this:
>> > The first:
>> > I buy a cable modem. Until the second I connect the cable-line to it so I
>> > can get a connection to the internet I can configure it in whatever
>> > manner I please. The second the line is connected, even though I *OWN*
>> > the hardware, I lose all control over its configuration.
>> >
>> > The second:
>> > I buy a DSL modem. Until I want to actually connect to the internet it
>> > can have whatever settings I want it to have. The second I want to
>> > connect to the internet it has to be configured the way that the ISP
>> > wants.
>>
>> Jung va gur jbeyq znxrf lbh guvax gurer vf n hfrshy nanybtl
>> orgjrra pbzzhavpngvba fgnaqneqf naq pbclevtug yvprafrf?
>>
>> One moment, let me retune.
>>
>> What in the world makes you think there is a useful analogy
>> between communication standards and copyright licenses?
>
> I don't. I was *REPEATING* an example of how TiVO has a *RIGHT* to change the 
> kernel or any other facet of the device connecting to their network. That 
> right *ISN'T* tied to copyright - as you have stated. Since it isn't, why is 
> the FSF trying to mandate that it is with the tivoization clauses in GPLv3?

The FSF *ISN'T* trying to mandate what *CONNECTS* to *TiVO's*
*NETWORK*.  *BOY* *ISN'T* *SHOUTING* *FUN*?  If Tivo wants to restrict
what connects to their network, the GPL and FSF will not stop them.
The major new (relative to GPLv2) things the FSF is trying to restrict
are new technical and legal methods that software distributors have
tried to use to convert free software into non-free software.

Tivo's right to dictate or change aspects of devices connecting to
their network is very much tied to copyright: Tivo needs appropriate
license if they are modifying or distributing anyone else's work.  I
am not sure why you think Tivo's right to free association trumps its
freely entered obligations to copyright owners.

>> Neither law nor common sense give much common ground to the two,
>> except in the general sense of two parties interacting.  One is a set
>> of rules so the two can interact through some information channel.
>> The other is a set of rules so that one can exploit a creative work
>> developed by the other.
>>
>> I suppose that you think it is acceptable for someone to offer access
>> to binary and source versions of GPLed software (with or without
>> modifications from commonly available versions) -- but only on the
>> condition that people never download the source versions?  That
>> certainly corresponds to the idea that Tivo can keep proprietary
>> extensions to the kernel if Tivo's customers want to connect to Tivo's
>> network services.
>
> Nope. Because that isn't a right they have that is disconnected from copyright 
> law. Or did you not read the entire post and just decide to try and make me 
> look stupid?

I read your post, but it was full of nonsense.  Tivo has every right
to restrict what connects to their network.  Tivo does not have the
right to infringe copyrights in order to make that restriction
effective.

You have said -- using enough words that you probably deceived
yourself -- that if Tivo distributes a specially mangled version of
Linux in order to restrict what connects to their network, and they
keep the mangling method proprietary, the GPL cannot shed light on
whether that hoarding is allowed.  Wrong: it can and does.

Michael Poole

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Sunday 17 June 2007 14:46:05 Michael Poole wrote:
> Daniel Hazelton writes:
> > On Sunday 17 June 2007 09:54:39 Michael Poole wrote:
> >> Daniel Hazelton writes:
> >> > But your server doesn't run the internet. TiVO may use phone lines to
> >> > connect a device to their server (and this is an example - I don't
> >> > know how TiVO devices actually connect) but the network being
> >> > connected to has a single owner who can set such terms.
> >> >
> >> > I'll repeat, in full, my earlier examples of this:
> >> > The first:
> >> > I buy a cable modem. Until the second I connect the cable-line to it
> >> > so I can get a connection to the internet I can configure it in
> >> > whatever manner I please. The second the line is connected, even
> >> > though I *OWN* the hardware, I lose all control over its
> >> > configuration.
> >> >
> >> > The second:
> >> > I buy a DSL modem. Until I want to actually connect to the internet it
> >> > can have whatever settings I want it to have. The second I want to
> >> > connect to the internet it has to be configured the way that the ISP
> >> > wants.
> >>
> >> Jung va gur jbeyq znxrf lbh guvax gurer vf n hfrshy nanybtl
> >> orgjrra pbzzhavpngvba fgnaqneqf naq pbclevtug yvprafrf?
> >>
> >> One moment, let me retune.
> >>
> >> What in the world makes you think there is a useful analogy
> >> between communication standards and copyright licenses?
> >
> > I don't. I was *REPEATING* an example of how TiVO has a *RIGHT* to change
> > the kernel or any other facet of the device connecting to their network.
> > That right *ISN'T* tied to copyright - as you have stated. Since it
> > isn't, why is the FSF trying to mandate that it is with the tivoization
> > clauses in GPLv3?
>
> The FSF *ISN'T* trying to mandate what *CONNECTS* to *TiVO's*
> *NETWORK*.  *BOY* *ISN'T* *SHOUTING* *FUN*?  If Tivo wants to restrict
> what connects to their network, the GPL and FSF will not stop them.
> The major new (relative to GPLv2) things the FSF is trying to restrict
> are new technical and legal methods that software distributors have
> tried to use to convert free software into non-free software.
>
> Tivo's right to dictate or change aspects of devices connecting to
> their network is very much tied to copyright: Tivo needs appropriate
> license if they are modifying or distributing anyone else's work.  I
> am not sure why you think Tivo's right to free association trumps its
> freely entered obligations to copyright owners.

Okay. So they give everyone the right to change the software on the box, but 
on connection replace the modified stuff with the official versions. Is that 
still a copyright problem? Absolutely, positively no. Is the current 
situation any different? Not that I can tell - they've changed a reaction 
into a preemptive act.

> >> Neither law nor common sense give much common ground to the two,
> >> except in the general sense of two parties interacting.  One is a set
> >> of rules so the two can interact through some information channel.
> >> The other is a set of rules so that one can exploit a creative work
> >> developed by the other.
> >>
> >> I suppose that you think it is acceptable for someone to offer access
> >> to binary and source versions of GPLed software (with or without
> >> modifications from commonly available versions) -- but only on the
> >> condition that people never download the source versions?  That
> >> certainly corresponds to the idea that Tivo can keep proprietary
> >> extensions to the kernel if Tivo's customers want to connect to Tivo's
> >> network services.
> >
> > Nope. Because that isn't a right they have that is disconnected from
> > copyright law. Or did you not read the entire post and just decide to try
> > and make me look stupid?
>
> I read your post, but it was full of nonsense.  Tivo has every right
> to restrict what connects to their network.  Tivo does not have the
> right to infringe copyrights in order to make that restriction
> effective.

Okay - "nonsense" in this sense meaning "it proves me wrong, but I can't be 
wrong, so it has no real meaning." I can accept that.

> You have said -- using enough words that you probably deceived
> yourself -- that if Tivo distributes a specially mangled version of
> Linux in order to restrict what connects to their network, and they
> keep the mangling method proprietary, the GPL cannot shed light on
> whether that hoarding is allowed.  Wrong: it can and does.

But the mangling method isn't proprietary. What is proprietary is a number 
that is input to a step of the process. (AFAICT the signing process is done 
with proprietary tools, but the process itself isn't)

DRH

> Michael Poole



-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 17, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> Okay. So they give everyone the right to change the software on the
> box, but on connection replace the modified stuff with the official
> versions.

If I haven't modified it so as to stop them from doing so on my
computer, that is ;-)

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Michael Poole]

Daniel Hazelton writes:

> Okay. So they give everyone the right to change the software on the box, but 
> on connection replace the modified stuff with the official versions. Is that 
> still a copyright problem? Absolutely, positively no. Is the current 
> situation any different? Not that I can tell - they've changed a reaction 
> into a preemptive act.

Again, it would be useful to limit discussions to this universe rather
than one that contains fictions necessary to your religious views.

(Separately, hardware connecting to Tivo's network is hardly related
to why they have DRM at all, and _entirely_ unrelated to what the GPL
permits; basing your argument on that nexus is stupid, but you have
obviously given up on your previous idea that the signature is somehow
an independent work of authorship and need some -- any -- theory to
continue holding to your general faith.)

>> > Nope. Because that isn't a right they have that is disconnected from
>> > copyright law. Or did you not read the entire post and just decide to try
>> > and make me look stupid?
>>
>> I read your post, but it was full of nonsense.  Tivo has every right
>> to restrict what connects to their network.  Tivo does not have the
>> right to infringe copyrights in order to make that restriction
>> effective.
>
> Okay - "nonsense" in this sense meaning "it proves me wrong, but I can't be 
> wrong, so it has no real meaning." I can accept that.

No; it means your position is at odds to fact, law and logic.  The
rest of my paragraph summarized why your argument is garbage.

Have you ever heard of the lawyer joke that if law is on your side,
you should pound the law; that if fact is on your side, you should
pound the facts; and that if neither are on your side, you should
pound the table?  You seem to be in the third state.

You suggested that Tivo's right to set terms for their network gave
them the right to modify the Linux kernel.  I twice pointed out in my
email that it did not -- that only the GPL gives them that right.  Any
right to modify the Linux kernel is inherently connected to copyright.
Kindly either refute or acknowledge that point: then this discussion
can advance.

>> You have said -- using enough words that you probably deceived
>> yourself -- that if Tivo distributes a specially mangled version of
>> Linux in order to restrict what connects to their network, and they
>> keep the mangling method proprietary, the GPL cannot shed light on
>> whether that hoarding is allowed.  Wrong: it can and does.
>
> But the mangling method isn't proprietary. What is proprietary is a number 
> that is input to a step of the process. (AFAICT the signing process is done 
> with proprietary tools, but the process itself isn't)

That does not help them at all.  In fact, it is probably *worse* for
Tivo if they are intentionally withholding input to a tool, since that
is closer to the traditional idea of source code.

Michael Poole

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 17, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> On Sunday 17 June 2007 09:54:39 Michael Poole wrote:
>> What in the world makes you think there is a useful analogy
>> between communication standards and copyright licenses?

> I don't. I was *REPEATING* an example of how TiVO has a *RIGHT* to
> change the kernel or any other facet of the device connecting to
> their network. That right *ISN'T* tied to copyright - as you have
> stated. Since it isn't, why is the FSF trying to mandate that it is
> with the tivoization clauses in GPLv3?

Since you're talking about rights, and that's a legal term, and we've
(hopefully) already established that intent of license author, intent
of copyright holder and letter of the license are different concepts,
and only the last of the 3 has to do with legal terms, I'll excuse
myself from the plane of spirits ;-) and get down to legal terms to
shoot down your argument.


Let's see...  US law states that (paraphrasing), if you grant a
copyright license that says the person can do such and such, you can't
later turn to that person and say "oh, BTW, I have this patent, and it
means you couldn't do such and such in the first place, unless you pay
me a gazillion bucks"



Patents have nothing to do with copyrights.  Still, a copyright
license can (and does) limit the ways in which you can use the power
that patent law gives you.

You could try to argue that "you have a right to the patent, and to
use it however you like".  But the moment you accept a license such as
v1, v2, or any later version to be published by the FSF, you give up
the power to use that patent to stop users from fully enjoying the
freedoms that the license granted them and said you couldn't further
restrict.


s/patent/anti-circumvention measure/

s/patent/hardware/


See?


Now, why would we be revising the license, if it's all already there?

First of all, to make this all clear.

Second of all, because law does not operate this way.  While there is
case law that establishes that copyright law supersedes patent law in
this sense (or so I'm told, I don't have the references and IANAL),
it's not clear that the same would hold for the DMCA, or technical
measures, or even discriminatory agreements.

So, in order to provide users with a better defense against these
dangers for the freedoms, the newer revision clarifies them, such that
whoever attempts to deny users' freedoms has a weaker defense for such
attempts, in a copyright infringement lawsuit.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Sunday 17 June 2007 15:32:34 Alexandre Oliva wrote:
> On Jun 17, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> > On Sunday 17 June 2007 09:54:39 Michael Poole wrote:
> >> What in the world makes you think there is a useful analogy
> >> between communication standards and copyright licenses?
> >
> > I don't. I was *REPEATING* an example of how TiVO has a *RIGHT* to
> > change the kernel or any other facet of the device connecting to
> > their network. That right *ISN'T* tied to copyright - as you have
> > stated. Since it isn't, why is the FSF trying to mandate that it is
> > with the tivoization clauses in GPLv3?
>
> Since you're talking about rights, and that's a legal term, and we've
> (hopefully) already established that intent of license author, intent
> of copyright holder and letter of the license are different concepts,
> and only the last of the 3 has to do with legal terms, I'll excuse
> myself from the plane of spirits ;-) and get down to legal terms to
> shoot down your argument.
>
>
> Let's see...  US law states that (paraphrasing), if you grant a
> copyright license that says the person can do such and such, you can't
> later turn to that person and say "oh, BTW, I have this patent, and it
> means you couldn't do such and such in the first place, unless you pay
> me a gazillion bucks"
>
>
>
> Patents have nothing to do with copyrights.  Still, a copyright
> license can (and does) limit the ways in which you can use the power
> that patent law gives you.
>
> You could try to argue that "you have a right to the patent, and to
> use it however you like".  But the moment you accept a license such as
> v1, v2, or any later version to be published by the FSF, you give up
> the power to use that patent to stop users from fully enjoying the
> freedoms that the license granted them and said you couldn't further
> restrict.
>

Agreed.

> s/patent/anti-circumvention measure/
>
> s/patent/hardware/

Under GPLv2 the rights granted are to "copy, distribute and modify". Under v3 
they are, as near as I can tell, your "beloved" 'Four Freedoms'. See the 
difference? GPLv2 can be boiled down to mandating the "open exchange of 
software", GPLv3 does the same, but also attempts to force the philosophy of 
one group on everyone else.

>
> See?
>
>
> Now, why would we be revising the license, if it's all already there?
>
> First of all, to make this all clear.

First of all, because we want everyone to believe exactly the same things we 
believe.

> Second of all, because law does not operate this way.  While there is
> case law that establishes that copyright law supersedes patent law in
> this sense (or so I'm told, I don't have the references and IANAL),
> it's not clear that the same would hold for the DMCA, or technical
> measures, or even discriminatory agreements.
>
> So, in order to provide users with a better defense against these
> dangers for the freedoms, the newer revision clarifies them, such that
> whoever attempts to deny users' freedoms has a weaker defense for such
> attempts, in a copyright infringement lawsuit.

I agree with attempting to protect people from the DMCA. The simple fact is 
that I feel that the GPLv3 would be better if all the "philosophy" crap was 
removed from it and the language cleaned up and simplified. In fact, add 
GPLv3dd4's section 3 was added to the GPLv2 it'd do just that. 

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Al Viro]

On Sun, Jun 17, 2007 at 12:31:00AM -0300, Alexandre Oliva wrote:
 
> I'm not trying to say why Linus and others chose the GPLv2.
> 
> I'm not trying to determine what their motivations were.
> 
> I'm not trying to force them to change to GPLv3.
> 
> I'm not trying to convince them that tivozation is a bad thing.
> 
> I'm only trying to show that anti-tivozation is in line with the
> spirit of the GPL.
 
... and anti-tivoization section shows all symptoms of going in wrong
direction, *whether* *tivo* *is* *good* *or* *not*.  It's full of
kludges exactly because it tries to carve out a notion that can only
be determined on case-to-case basis and not by generic definition.
And no, that's not a matter of bad wording in that section.

I don't _care_ whether it breaks spirit, etc. - it's a fundamentally bad
idea for completely independent reasons.  Even if one thinks that tivo
in particular ought to be sued into the ground at some point.

Besides, it's fscking *pointless* for userland stuff.  If you insist that
e.g. glibc will infect by linking, you've just created a huge problem for
any GPLv2 userland code, which will make all bad blood about kernel look
trivial in comparison.  If you do not, then you've lost all leverage anyway;
kernel won't switch, libraries are OK, toolchain is obviously OK for building
code with any license... what's left?  The glorious /bin/cp?  Sorry, it would
work as usual, subject to open(2) not returning EACCES.  Just as on any
system.  Just what is it going to prevent?  Hell, they can slap selinux on
the box, protect what they want to protect, use crypto-loop to prevent offline
modifications of filesystem and hide the key in firmware.

Either GPLv2 is sufficient in given case (and e.g. Alan decides to go
after company in question), or you've at most created a moderate amount
of work rewriting the checks they are doing into a different form (if
that).  Good job.  In the meanwhile, you've got a load of ill-defined
verbiage around installation instructions.  I.e. a lovely fodder for
potential abusers.

Sod it.  GPLv3, with your involvement in its development or not, sucks
rocks, thanks to what you call anti-tivoization section.

-- 
How many GPL spirits can dance on the end of a pin?

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 17, 2007, Al Viro <viro@ftp.linux.org.uk> wrote:

> It's full of kludges exactly because it tries to carve out a notion
> that can only be determined on case-to-case basis and not by generic
> definition.

I agree it's a very difficult definition.  I'm not sure I'm happy with
the wording in place right now.  But I very much like and agree with
its purpose, and it is in line with the goal of respecting and
defending users' freedoms, which is what the FSF cares mostly about,
and must care about, because it's its official and public mission.

> I don't _care_ whether it breaks spirit, etc.

That's what I care about, and I've seen false claims that it does.

Can you please acknowledge that it doesn't, such that I can feel I've
fulfilled my goal of dispelling the myth that the GPLv3 changes the
spirit of the GPL?

> GPLv3, with your involvement in its development or not, sucks rocks,
> thanks to what you call anti-tivoization section.

Is it correct to say that you share Linus' opinion, that the only
problem with the GPLv3 is the anti-tivoization provision?


To make this more concrete, if there was a hypothetical GPLv2.9,
consisting of GPLv3dd4 minus the "installation information"
requirements for user products, (i) Would you consider it a better
license than GPLv2?  (ii) Better for Linux?  (iii) Enough to go
through the trouble of switching?

I'd love answers to these 3 questions from others too.

Just in case, I shall point out, one more time, that I'm speaking for
myself, not for the FSF, not for FSFLA, not for Red Hat.  The
questions above are to satisfy my personal curiosity.  I don't make
any commitment whatsoever to take the answers up to the FSF, and I
don't want to set any expectations as to whether they could or would
make any difference, at this point, about the outcome of GPLv3.

If you want your opinions to stand a chance to make a difference, the
right place to provide them is gplv3.fsf.org/comments, and time is
running short.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Greg KH]

On Sun, Jun 17, 2007 at 02:56:24AM -0300, Alexandre Oliva wrote:
> 
> If you want your opinions to stand a chance to make a difference, the
> right place to provide them is gplv3.fsf.org/comments, and time is
> running short.

If you honestly think that the "anti-tivo" clause in GPLv3 will be
removed just because we start to add more comments to that page, then
you are sorely mistaken.  From the very _beginning_ of the v3 process
the kernel developers have showed their objection to that section of the
license, and we were told, to our face, with no uncertian terms, that it
was going to stay, in one form or another, no matter what we thought or
said about it.

So, why would we want to waste our time filling out web forms after
that?

greg k-h

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 18, 2007, Greg KH <greg@kroah.com> wrote:

> On Sun, Jun 17, 2007 at 02:56:24AM -0300, Alexandre Oliva wrote:
>> 
>> If you want your opinions to stand a chance to make a difference, the
>> right place to provide them is gplv3.fsf.org/comments, and time is
>> running short.

> If you honestly think that the "anti-tivo" clause in GPLv3 will be
> removed just because we start to add more comments to that page, then
> you are sorely mistaken.

I agree that adding comments wouldn't accomplish this.

But many objections were about the wording, about cases that perhaps
shouldn't be covered, and these could make a difference.

> From the very _beginning_ of the v3 process the kernel developers
> have showed their objection to that section of the license, and we
> were told, to our face, with no uncertian terms, that it was going
> to stay, in one form or another, no matter what we thought or said
> about it.

This sounds about right.

> So, why would we want to waste our time filling out web forms after
> that?

If you're adamantly favorable to permitting any form of Tivoization
whatsoever, don't bother.

Others who aren't so fundamentalist as to reject anti-tivozation on
ideological grounds, in spite of evidence that such measures would
advance the very pragmatic interests they claim to place above
ideology, might be willing to help shape these provisions so
that they don't hurt those who respect users' freedoms, but accomplish
the goal of keeping Free Software Free.


Seriously, looking only at the downside of anti-tivoization (tivoizer
might turn us down), without even acknowledging that, should the
tivoizer change practice and respect users' freedoms, you'd be able to
get far more contributions from all those users, is typical minimax
strategy.  That's the worst case for the prisoner's dilemma.  That's
not pareto optimal.  It may not be a losing strategy, but it's not the
best strategy for everyone.


Every time you enable someone to disrespect other users' freedoms WRT
your software, you cut yourself out of some contributions that user
could make.  Even if you completely disregard the moral and ethical
aspects of software freedom, the open source mentality inherently
depends on the notion of respect for others' freedoms.  You only reap
the benefits of open source when the user gets the freedoms respected.
That's why preventing people from hiding source code, from using other
technical measures, and from using copyright, patents and
anti-circumvention laws, to stop or decrease the possibility or the
incentive for a user to contribute to your community is not only a
great ethical and moral stance, it is also self-beneficial, in the
very sense that Linus and others claim.


So although I like to highlight the moral and ethical aspects, and
others like to highlight the self-beneficial mechanics of the system,
they are really two sides of the same coin.

And if you didn't think so, if you didn't believe in increasing the
incentive and enablement for a user to cooperate with you by means of
stopping others from removing such incentive or possibilities, you
wouldn't be using a license that established such conditions, you'd be
using a more liberal license.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[david]

do you realize that redhat uses checksums or signatures to check the 
validity of their CD's?

try to burn a redhat image with the -pad option which adds a chunk of 0's 
to the end of the image and try to boot it.

I seriously doubt if redhat tells you how to how to generate such a 
checksum/signature.

in addition to the problem that Linus points out about being unable to 
change the contents of the write-only CD this would seem to conflict with 
what you are claiming the GPL is supposed to allow.

David Lang

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Chris Adams]

Once upon a time,  <david@lang.hm> said:
>do you realize that redhat uses checksums or signatures to check the 
>validity of their CD's?
>
>try to burn a redhat image with the -pad option which adds a chunk of 0's 
>to the end of the image and try to boot it.

It'll boot and run just fine.  The checksum is simply an optional
integrity check (you can bypass it by choosing "Skip" when it prompts to
test your CD/DVD).

>I seriously doubt if redhat tells you how to how to generate such a 
>checksum/signature.

You are seriously wrong then.  The tool is in the anaconda package, and
there are docs on using it both there and many places on the web.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 18, 2007, david@lang.hm wrote:

> do you realize that redhat uses checksums or signatures to check the
> validity of their CD's?

Yes.  How does this stop the users from enjoying any of the freedoms?

> I seriously doubt if redhat tells you how to how to generate such a
> checksum/signature.

Mixing two different issues here.

The checksums embedded in the ISO images are implemented in anaconda
itself, provided along with the distro, so anyone can generate them,
even though I don't know the precise algorithm.  The checksums over
the ISO images themselves are implemented with sha1sum (earlier,
md5sum), so anyone can generate them too.

As for the GPG signatures in the RPMs and in the SHA1SUM file, these
are indeed generated using public algorithms but private keys.  But
these signatures are not functional, and they don't in any way stop
anyone from enjoying freedoms.

> in addition to the problem that Linus points out about being unable to
> change the contents of the write-only CD this would seem to conflict
> with what you are claiming the GPL is supposed to allow.

It looks like you read only Linus' messages, not my responses.  A
write-only CD is like ROM.  It's not the distributor who's imposing
restrictions on its modification.  Nobody can modify it because nature
says so.  It's not like the software is being recorded in a CD as a
means to prevent you from modifying it.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Dave Neuer]

On 6/18/07, Alexandre Oliva <aoliva@redhat.com> wrote:
>
> Seriously, looking only at the downside of anti-tivoization (tivoizer
> might turn us down), without even acknowledging that, should the
> tivoizer change practice and respect users' freedoms, you'd be able to
> get far more contributions from all those users, is typical minimax
> strategy.

<snip>

> Every time you enable someone to disrespect other users' freedoms WRT
> your software, you cut yourself out of some contributions that user
> could make.  Even if you completely disregard the moral and ethical
> aspects of software freedom, the open source mentality inherently
> depends on the notion of respect for others' freedoms.  You only reap
> the benefits of open source when the user gets the freedoms respected.

Alexandre, while I backed you up on the whole "spirit of the GPL
hasn't changed" thing, I think you are wrong here. As Jesper, Johannes
and others have already pointed out (in a couple of the very few
cogent non-flames since this thread started), we undoubtedly get more
back in the form of software contributions from paid developers of
TiVO-like companies than from the very few end users with the skill to
hack the software or the inclination to aquire said skill. The simple
fact is that most end users of most electronic devices don't care
about the Freeness of the software, they care if their device works.

If you disregard the ethical dimension, I think it's hard to argue w/
a straight face that Linus' stance is wrong from a pragmatic
standpoint.

The problem is that the people saying "just don't buy TiVOs" know full
well that because the number of end users who care is so small, they
have _no_ economic power to change the situation, and that's why
people who do consider this an ethical issue want to leverage the
power the copyright holders have both legally and by virtue of their
expertise to force the hardware vendors to cooperate w/ end users.

Dave

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 18, 2007, "Dave Neuer" <mr.fred.smoothie@pobox.com> wrote:

> On 6/18/07, Alexandre Oliva <aoliva@redhat.com> wrote:

>> Seriously, looking only at the downside of anti-tivoization (tivoizer
>> might turn us down), without even acknowledging that, should the
>> tivoizer change practice and respect users' freedoms, you'd be able to
>> get far more contributions from all those users, is typical minimax
>> strategy.

>> Every time you enable someone to disrespect other users' freedoms WRT
>> your software, you cut yourself out of some contributions that user
>> could make.  Even if you completely disregard the moral and ethical
>> aspects of software freedom, the open source mentality inherently
>> depends on the notion of respect for others' freedoms.  You only reap
>> the benefits of open source when the user gets the freedoms respected.

> Alexandre, while I backed you up on the whole "spirit of the GPL
> hasn't changed" thing, I think you are wrong here.

Thank you, this is the first time someone voices a coherent argument
that doesn't completely dismiss the potential benefits of prohibiting
tivoization, and doesn't focus only on the potential downsides of such
measures.  Even though you come to a different conclusion than I did,
I thank you very much for not pretending the point does not exist at
all, or phrasing arguments that don't counter the point in any way.

> we undoubtedly get more back in the form of software contributions
> from paid developers of TiVO-like companies than from the very few
> end users with the skill to hack the software or the inclination to
> aquire said skill. The simple fact is that most end users of most
> electronic devices don't care about the Freeness of the software,
> they care if their device works.

This is also true of computer users in general.  So what is it that
makes this case different?

Why should we ask vendors to pass on the corresponding sources to
users, if most users don't care?

Why should we ask vendors to not use patents to stop a user from
modifying and sharing the software they distribute, if most users
don't care?

What is it that makes the reasoning different for the particular case
of vendors using technical tricks to stop a user from enjoying the
benefits of modifications she could make herself, which could then
very well benefit the entire community and even the vendor?

> If you disregard the ethical dimension, I think it's hard to argue
> w/ a straight face that Linus' stance is wrong from a pragmatic
> standpoint.

Maybe I'm wrong, but I hadn't seen arguments that even pointed in this
direction, so the complete dismissal of the upsides came off as
selective attention.

Let's see how this proceeds.

> The problem is that the people saying "just don't buy TiVOs" know full
> well that because the number of end users who care is so small, they
> have _no_ economic power to change the situation, and that's why
> people who do consider this an ethical issue want to leverage the
> power the copyright holders have both legally and by virtue of their
> expertise to force the hardware vendors to cooperate w/ end users.

Yes, your reasoning makes sense to me.  I can't speak for the FSF's
intentions, but I have no reason to disbelieve it's moved by ethics,
rather than by some plot to hurt TiVo.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Greg KH]

On Mon, Jun 18, 2007 at 03:20:39PM -0300, Alexandre Oliva wrote:
> On Jun 18, 2007, Greg KH <greg@kroah.com> wrote:
> > From the very _beginning_ of the v3 process the kernel developers
> > have showed their objection to that section of the license, and we
> > were told, to our face, with no uncertian terms, that it was going
> > to stay, in one form or another, no matter what we thought or said
> > about it.
> 
> This sounds about right.
> 
> > So, why would we want to waste our time filling out web forms after
> > that?
> 
> If you're adamantly favorable to permitting any form of Tivoization
> whatsoever, don't bother.

For the record, I completely feel that what Tivo did was both legally
correct[1], and the correct thing to do for their system, and would
fight _very_ hard any attempt to change the Linux kernel's license that
would prevent this usage model.

So I will not bother anymore.

greg k-h

[1] The FSF lawyers also agree with this.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Theodore Tso]

On Mon, Jun 18, 2007 at 12:48:13PM -0700, Greg KH wrote:
> > > So, why would we want to waste our time filling out web forms after
> > > that?
> > 
> > If you're adamantly favorable to permitting any form of Tivoization
> > whatsoever, don't bother.
> 
> For the record, I completely feel that what Tivo did was both legally
> correct[1], and the correct thing to do for their system, and would
> fight _very_ hard any attempt to change the Linux kernel's license that
> would prevent this usage model.
> 
> So I will not bother anymore.

Linus has spoken, Greg K-H has spoken, many other people spoken ---
and yet Alexandre keeps on speaking, and speaking, and speaking....

It's pretty clear no one is convincing anyone, and that everyone
understands their position, and are happy with it, and so all we're
doing now is wasting bandwidth.

Can we please end this thread?  

Please?

						- Ted

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bernd Schmidt]

Greg KH wrote:
> On Sun, Jun 17, 2007 at 02:56:24AM -0300, Alexandre Oliva wrote:
>> If you want your opinions to stand a chance to make a difference, the
>> right place to provide them is gplv3.fsf.org/comments, and time is
>> running short.
[...]
> So, why would we want to waste our time filling out web forms after
> that?

In case anyone was wondering if the FSF is genuinely interested in 
feedback - I went and made some comments on the draft, and they appear 
to no longer be there a few days later.  Thanks for the invitation Alex.


Bernd

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 21, 2007, Bernd Schmidt <bernds_cb1@t-online.de> wrote:

> I went and made some comments on the draft, and they appear to no
> longer be there a few days later.

This would be very bad.  Please let me know what they were about and
I'll try to figure out what happened.

Did you by any chance file them against an earlier draft?  Those (for
obvious reasons) no longer appear against the current draft, but
they're still accessible by other means.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Al Viro]

On Sun, Jun 17, 2007 at 02:56:24AM -0300, Alexandre Oliva wrote:
> Can you please acknowledge that it doesn't, such that I can feel I've
> fulfilled my goal of dispelling the myth that the GPLv3 changes the
> spirit of the GPL?

No.  I don't do metaphysics.  This thread alone has shown that the
notion is not well-defined *at* *all*, to the point of being useless
and seriously misleading.  I.e. the phrase about similar spirit
should be replaced with something far more explicit and very, very
hard to miss.  I don't think you need more proof that people *do*
interpret it in very different ways, with quite unpleasant results.
 
> > GPLv3, with your involvement in its development or not, sucks rocks,
> > thanks to what you call anti-tivoization section.
> 
> Is it correct to say that you share Linus' opinion, that the only
> problem with the GPLv3 is the anti-tivoization provision?

No.  If you want a basic splitup by sections compared to GPLv2,
	1	-	at least not better; attempts at being precise
			end up creating a no-common-sense-land *and*
			turn out to leave serious unanswered questions
			in that area.
	2	-	no opinion on actual changes
	3	-	more or less an improvement
	4,5	-	about on par with v2, modulo wording in (5)
	6	-	much worse
	7	-	if I want to give additional permissions, I don't
			want them stripped, for fsck sake!  There is a
			bog-standard mechanism for _that_ (dual-licensing),
			thank you very much.  I.e. that section looks like				a pile of dishonest PR games, pardon the redundance.
	8	-	on par
	9	-	on par, modulo piss-poor attempt to define "modify"
			backfiring here (e.g. prelinking constitutes
			modification according to it, so does running rdev(8),
			etc., etc.)
	10	-	no opinion on actual changes
	11	-	improvement
	12	-	on par (aside of basic bad writing, but there are
			much worse problems *not* with wording, so that's
			not interesting)
	13	-	special-case kludges are fun, aren't they (specifically
			"linking"?), but in any case, that's secondary.
			FWIW, I'm not fond of ideas behind Affero, so if
			anything, that's a point against v3.
	14	-	... and thank you very much for keeping such a lovely
			source of periodic clusterfucks in v3 as well.
			I think it's painfully obvious for everyone in this
			thread that reference to "spirit" is a recipe for
			massive disagreements down the road.  If you want the
			words you are using to be interpreted your way, use
			ones that have commonly agreed upon meaning.  The
			measure is "do other people read it differently?",
			not "how sure I am in deriving the meaning I want from
			the words I've used?".  Related problem is that
			version choice rules _must_ be stated in maximally
			unambiguous and hard to miss way.  Look through
			Bernd-produced parts of this thread and you'll see
			the reason why it is needed.
			Moving that into terms and conditions is a good step,
			but it's still not enough.  E.g. you really want
			to be explicit on the form (in)sufficient to specify
			the version of license.
	the rest	on par.

Overall: definitely worse than v2.  v2 + (3) + (11) would be an improvement,
provided that v2 section 9 is cleaned up.

> To make this more concrete, if there was a hypothetical GPLv2.9,
> consisting of GPLv3dd4 minus the "installation information"
> requirements for user products, (i) Would you consider it a better
> license than GPLv2?
	Negative, see above
 (ii) Better for Linux?
	Negative, for kernel as well as for userland
(iii) Enough to go through the trouble of switching?
	See above.

In other words, I don't see any chance for v3 to be a good choice
for anything I write, kernel or userland.  If I end up sending patches
to v3 projects, I'll put the patches under BSDL and let them convert
on merge.

Note that this is *not* about the problems with wording; those also exist,
of course (_that_ is a final draft?), but that's a separate story and it
interests me only inasmuch as it is caused by inherent problems with meaning
of section in question.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 18, 2007, Al Viro <viro@ftp.linux.org.uk> wrote:

> I.e. the phrase about similar spirit should be replaced with
> something far more explicit and very, very hard to miss.

This is a very good idea.  Would you please file it at
http://gplv3.fsf.org/comments?

> I don't think you need more proof that people *do* interpret it in
> very different ways, with quite unpleasant results.
 
Agreed.

>> Is it correct to say that you share Linus' opinion, that the only
>> problem with the GPLv3 is the anti-tivoization provision?

> No.

Thanks for your detailed analysis.  I wish I knew what to do with
it.  (I decline impolite suggestions, thanks ;-)

Would you like me to put you in touch with Richard Fontana, one of the
lawyers involved in GPLv3, that's regarded as a legalese compiler, to
discuss your issues about wording with him?  Or would you rather file
them (with a bit more detail) at gplv3.fsf.org/comments?

Aside from wording issues, which appear to dominate your comments, is
it fair to characterize that your objections to GPLv3 are
anti-tivoization provisions (strong) and Affero compatibility (weak?)?


I'm setting wording issues aside because these are easier to fix once
the problem is understood, rather than ideological differences, that
would probably be pointless to attempt to fix.

> 	7	-	if I want to give additional permissions, I don't
> 			want them stripped, for fsck sake!  There is a
> 			bog-standard mechanism for _that_ (dual-licensing),
> 			thank you very much.

additional permissions are indeed a form of dual-licensing, but one
that doesn't require one to create a copy of the GPL and add the
additional permissions to that copy.  Yes, it could be accomplished
with dual-licensing terms such as "you can follow the terms of the
GPL, with the following additional permissions".

I don't quite see the point of criticizing this.  This is more
informative than anything else.

The meat here is really in the few additional restrictions, and the
provisions to combat the practice of adding restrictions on top of the
GPL and claiming the software is available under the GPL, which has
made for a lot of confusion over time.


Thanks a lot for your feedback.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 17, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:

> On Sun, 17 Jun 2007, Alexandre Oliva wrote:
>> 
>> They're based on the Free Software definition, that establishes the
>> four freedoms that the GPL was designed to respect and defend.

> The GPL is a software license, *independent* of that thing.

One more time, I'm not talking about the license (the legal terms).
I'm talking about its spirit.  It's encoded in the preamble, that
refers to "free software", which can't possibly be defended as meaning
anything but what the FSF itself defined in the Free Software
Definition.

Is this some form of mental block that stops you from realizing that
the spirit of the license is pretty much all I've been talking about
here, and that I've already said that at least 20 times in this
thread?

Why do you insist in bringing the legal terms back into this
discussion about the spirit of the license?

What are you trying to accomplish, other than generating more
confusion and pretending that you have a strong point?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Sun, 17 Jun 2007, Alexandre Oliva wrote:
> 
> One more time, I'm not talking about the license (the legal terms).

Ok. Then go away.

Everybody else just cares about the legal reasons.

The "legal terms" is the only reason a license *exists*. That's what a 
license *is*, for crying out loud!

If you don't care about the legal side, go and read the free software 
manifesto. That's the paper you're really arguing about.

If you want to argue about the GPLv2 *license*, then you'd better start 
caring about the legal issues. Because that is what the license is: a 
_legal_ document.

		Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 17, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:

> On Sun, 17 Jun 2007, Alexandre Oliva wrote:
>> 
>> One more time, I'm not talking about the license (the legal terms).

> Ok. Then go away.

> Everybody else just cares about the legal reasons.

That's false, and the reason I know it is that, if this was true, I
probably wouldn't even have got into this debate.

The reason I got into it was that there were false claims about
changes in the spirit of the license, implied in Greg KH's e-mail that
refers to the long thread and position paper that made these false
claims.  I got into this debate in response to this reference:

http://www.ussg.iu.edu/hypermail/linux/kernel/0706.1/1904.html

to a thread about an article that stated

  [...] solemn trust, as stated in article 9 of GPLv2, only to licence
  the code under versions of the GPL that "... will be similar in
  spirit to the present version". We, like all the individual
  contributors to GNU projects, have taken that trust at face value
  and accorded the FSF a special role in the Open Source Universe
  because of it. It goes without saying that any updates to GPLv2 must
  be completely in accord with the execution of that trust.

before it went on to misleading references to a DRM clause (*), trying
to frame anti-tivoization as end-use restrictions and how unacceptable
that is in a license, while blessing tivoization itself (an end-use
restriction, no less).


(*) meaning the anti-tivoization provisions, in spite of the existence
of an unrelated section named DRM in a draft available at that time


And then, the article goes on about FUD of jeopardising patent
portfolios, which is *obviously* a misunderstanding, as I wrote in my
e-mail.


In response to that intervention, my first one in this subject, you
exploded.  And then, shortly after your explosion and all the name
calling, there was this:

http://www.ussg.iu.edu/hypermail/linux/kernel/0706.1/2235.html


And you say people don't care about the spirit?


Seriously, I'm not even asking for an apology for the name calling.
All I'm asking for is that you don't lie about GPLv3 and about the
spirit of the GPL.

I know it's hard when you're trying to frame yourself as a victim of a
tyranny, but since FSF's offer of a new license, and switching its own
projects to it, framing it as "suborn or coerce others to go along
with them" (from the article) is utterly false, and an evident
misunderstanding of the intent.

And that's not even unusual.  See, when Free Software advocates *ask*
people to refer to the GNU operating system as such, even when it's
combined with the kernel Linux, a lot of people over-react and
complain that the FSF is trying to *force* anyone to rename Linux.
This is false on two accounts.  Linux is the name of the kernel, and
the FSF has never tried to rename that.  GNU is the name of the
operating system, and it's people who confusingly call that operating
system Linux that renamed it.  Free Software advocates merely ask that
people call it by the right name.

I can appreciate that some FSFers take positions that may come off as
unfriendly and trigger this kind of negative feeling of being forced
to do something.  I suspect this is the same case.

But taking it as forcing is just as inappropriate.  There's no
coercive force to be applied, and no moral grounds to coerce.  In both
cases, it's a request for people to make the choice of helping defend
users' freedoms.


I know you're not going to believe this, because it doesn't fit in
your world-view.

My intent in participating in this discussion was really not to
convince you.  I had a feeling, all the way from the beginning, that
this was probably hopeless, even though there was thin hope.

My intent was to enlighten others who were still open to listening and
to accepting different viewpoints.  From personal feedback I got, I
know I've accomplished a lot of that, and I thank deeply all of you
who encouraged me to proceed, and who thanked me for the information I
added to the debate.


I wish I could say I'm going to step out of this debate now, because
it's taken me far more time than I could afford.  But since my goal
was to counter the spreading of mis-information about GPLv3, and since
I am indeed addicted to this kind of discussion, I may end up failing
to stay away from the discussion.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bron Gondwana]

[note: I'm writting this while offline and likely to remain so for the
next 8 hours or so, so I'll probably miss a bunch of other replies]

On Sat, Jun 16, 2007 at 02:14:29PM -0300, Alexandre Oliva wrote:
> On Jun 16, 2007, Bron Gondwana <brong@fastmail.fm> wrote:
> 
> > On Sat, Jun 16, 2007 at 05:22:21AM -0300, Alexandre Oliva wrote:
> >> On Jun 15, 2007, Bron Gondwana <brong@fastmail.fm> wrote:
> >> 
> >> > because it could easily be argued that they linked the BIOS with the
> >> > Linux kernel
> >> 
> >> How so?
> 
> > Er, they installed it in the same piece of equipment, and the kernel
> > couldn't function without it in that work.
> 
> I see what you're getting at.  You're thinking of a license that
> doesn't respect the idea of "mere aggregation", right?

No, I'm arguing that it's not "mere aggregation" - the kernel is useless
on that machine unless the BIOS is present or replaced with something
else with equivalent functionality.  I suspect any decent lawyer could
make the theory that this made the kernel as compiled on to that machine
with specific chipset support selected for that hardware into a "derived
work" of the BIOS - especially if the vendor had contributed GPLed code
for drivers which interact with their hardware into said kernel.

In fact, particularly if the hardware vendor has also contributed GPL
code that interacts on one side of the software/(firmware, hardware)
boundard which worked around bugs in said firmware/hardware which they
also had the ability to change.  The two really are a combined work of
which only one part is GPLed.

Ringing any binary kernel module video card driver bells yet?  It's
really the same thing from the opposite direction - the only criteria
is where you fit in the pecking order - hardware manufacturers work
around Windows bugs, Linux kernel drivers work around hardware bugs -
it's all about who has more to lose if they aren't compatible.
 
> For starters, this wouldn't evidently not qualify as an Open Source
> license, and I'm pretty sure it wouldn't qualify as a Free Software
> license either.

Strawman licence?

> > By using GPLix as part of their boot process along with their
> > non-GPL BIOS, they're subverting the freedoms that the user should
> > have in being able to control the entire boot process.
> 
> You're pushing the "freedom to change" too far.  Sure, I'd like to be
> able to do that, and I prefer hardware that lets me do it, but it's
> not like this BIOS in the scenario you described is being used as a
> means to stop me from modifying the GPLed software.

Well, yeah - except this is the direction GPL3 takes us, and it's a
theory that GPL3 makes more likely to fly in court than GPL2 does -
meaning that hardware vendor lawyers lie awake at night worrying about
stuff (I'd hate to be a good lawyer - I'd never get any sleep!)

> I have never said that including a GPLed piece of software should
> grant users the right to modify anything whatsoever in the system, or
> grant them control over the entire system.  Others have, but it's not
> true, it just shows how much mis-information is floating around.

No, but your interactions with Linus (lazy bums 'r' us) have shown that
the logical result of what you do want includes this.  It's a lot harder
to objectively judge one of these than the other:

a) have they provided the source code to this binary to anyone who asks.

b) have any of the limitations of this piece of hardware been created
   with the intent of making it more difficult for J. Random Enduser to
   build modified binaries from said source and have them function
   correctly.

(b) has much more scope for shenanigans by bad apples on the copyright
owner side - and don't pretend that only the hardware vendors are bad
guys - it takes all sorts and the idea of a licence is to protect both
parties.

> All the GPL stands for is to defend the freedom of the users over the
> particular program it applies to.  You can't impose further
> restrictions on the user's ability to modify what *that* software
> does.

Except where they run into limitations of the platform itself, or just
plain bugs.  Oops.  The lawyers will have a field day discovering intent
every time J. Random's kernel doesn't do what he wants after he fiddles
the code a bit.

> If you wanted to change something else, but this something else is not
> covered by the license, and is not being used to contradict the terms
> of the license, well, too bad, you lose.
> 
> > b) deny themselves the ability to every offer a patch to said BIOS if
> >    bugs are found 
> 
> > Point (b) is also exactly on topic for the discussion of enforcing
> > legal safety obligations in hardware on a peripheral rather than the
> > software drivers.
> 
> > It's requiring that these limitations be placed in a technically
> > inferior location to hack around a legal "bug".
> 
> I don't think this last sentence is true.  If you implement hardware
> locks that prevent modification of the software even by yourself, then
> you're in compliance with the terms of the GPLv3dd4.  But IANAL.

I obviously wasn't clear enough.  The only way to come into complience
with GPL3dd4 is to reduce your ability to fix things or grant everyone
else the ability to mess with things.  This severely restricts you from
doing _anything_ in certain problem spaces due to local laws on the
topic, even if you're an otherwise good actor who is making worthwhile
source code contributions to the rest of the community.

This would be a lot less of an issue if Linux was a modular kernel
(don't shoot me Linus) and you could be allowed to change the bits that
didn't touch the regulated hardware's access paths.  Messy to control if
you're running in ring0 though - you need hardware managed restrictions
at some level, and a barrier around the entire kernel is certainly the
easiest way to do that.

Bron.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Sun, 17 Jun 2007, Bron Gondwana wrote:
> 
> No, I'm arguing that it's not "mere aggregation" - the kernel is useless
> on that machine unless the BIOS is present or replaced with something
> else with equivalent functionality.

That's *not* a valid argument!

I know, I know, it's a common one, but it is *nonsense*.

The thing is, "mere aggregation" doesn't mean what you think it means.

"Mere aggregation" doesn't mean that they cannot depend on each other. It 
means that they are not *based* on each other in the sense of GPLv2.

In other words, "mere aggregation" is about two pieces that are not 
derivative works under copyright law.

For example, on a Red Hat DVD, *every*single*binary* on that DVD requires 
a kernel to run on. And the kernel image itself "depends" on the user 
programs to actually do something _useful_. 

But it's all still "mere aggregation", because they are not related to 
each other in the sense of being derived works!

So "mere aggregation" is not about intimacy. OF COURSE high-tech products 
depend intimately on each other. The Linux kernel cannot boot on a PC 
without a BIOS or something equivalent. You cannot run your graphical 
environment without a kernel, an X server, the CPU, the memory, the 
display, the BIOS, the power company (or an equivalent hand-crank) etc etc 
etc, and these things are all very much dependent on each other to make a 
"usable system", that has absolutely _zero_ relevance to whether they are 
"mere aggregation" or not.

So the only thing the "mere aggregation" phrase in the GPLv2 means is 
simply: "putting together two or more pieces that are not derived works of 
each other". That's what that

	mere aggregation of another work not based on the Program
	                 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

part of the license is all about.  The BIOS is "another work", and it 
clearly is "not based on the Program". The "aggregation" just means 
"putting them together", and the "mere" is there just to show that as far 
as the license is concerned, that is not even interesting!

In other words, the sentence in section 2 about mere aggregation says one 
thing only: the fact that you lump things together storage-wise does not 
matter *at*all* for the license. The _only_ thing that matters for the 
GPLv2 is whether something is a derived work, _not_ whether it is lumped 
together, and depends on something else.

This is why the BIOS is irrelevant. You can put it in the same flash chip 
as the kernel: it still isn't a derived work of Linux (and Linux is not a 
derived work of the BIOS), and the GPLv2 explicitly makes it clear that 
the license holds no sway over the BIOS, even though it's physically on 
the same chip.

The BIOS can do whatever it wants to, and the GPLv2 has *nothing* to say 
about it. The GPLv2 makes that very clear indeed. It is just "mere 
aggregation", and the fact that they  are in the same machine, on the same 
harddisk, on the same flash rom, or on the same DVD doesn't bring them any 
closer from a *copyright* angle.

Copyright in general (and the GPLv2 in the specific) isn't about how 
things are *physically* tied together. It's not even about how things 
interconnect, and how "important" they are for each other. It is _purely_ 
about whether they are derived works.

So when the GPLv2 talks about "mere aggregation", it talks about it 
specifically to say that it does not matter, and that the license _only_ 
affects the actual derived work!

This is why the BIOS, the hardware, the programs you run under Linux etc 
etc are all totally irrelevant. The GPLv2 explicitly says that they are 
irrelevant as far as that license is concerned.

			Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Jeffrey V. Merkey]

Linus Torvalds wrote:

>On Sun, 17 Jun 2007, Bron Gondwana wrote:
>  
>
>>No, I'm arguing that it's not "mere aggregation" - the kernel is useless
>>on that machine unless the BIOS is present or replaced with something
>>else with equivalent functionality.
>>    
>>
>
>That's *not* a valid argument!
>
>  
>
Linus,

Just take a vote and start tagging files and ignore this needless 
diatribe. It is was it is, I seriously doubt you will
get all of Linux moved to GPL3 as a monolith, since you will never get 
concensus. You should fork a GPL3 kernel and let people decide
whether their code goes in or not. If they don't want to move to it, new 
people can contribute new code. Start a 2.8 tree or whatever that is 
GPL3 only.

Jeff

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Carlo Wood]

On Sat, Jun 16, 2007 at 08:21:22PM -0600, Jeffrey V. Merkey wrote:
> Linus,
> 
> Just take a vote and start tagging files and ignore this needless 
> diatribe. It is was it is, I seriously doubt you will
> get all of Linux moved to GPL3 as a monolith, since you will never get 
> concensus. You should fork a GPL3 kernel and let people decide
> whether their code goes in or not. If they don't want to move to it, new 
> people can contribute new code. Start a 2.8 tree or whatever that is 
> GPL3 only.

*snicker*

Yeah, Linus - when do you finally understand that nobody WANTS
your GPL3 kernel?  :))

ROFL

-- 
Carlo Wood <carlo@alinoe.com>

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bron Gondwana]

On Sat, Jun 16, 2007 at 05:58:11PM -0700, Linus Torvalds wrote:
> On Sun, 17 Jun 2007, Bron Gondwana wrote:
> > 
> > No, I'm arguing that it's not "mere aggregation" - the kernel is useless
> > on that machine unless the BIOS is present or replaced with something
> > else with equivalent functionality.
> 
> That's *not* a valid argument!
> 
> But it's all still "mere aggregation", because they are not related to 
> each other in the sense of being derived works!
> 
> So "mere aggregation" is not about intimacy. OF COURSE high-tech products 
> depend intimately on each other. The Linux kernel cannot boot on a PC 
> without a BIOS or something equivalent. You cannot run your graphical 
> environment without a kernel, an X server, the CPU, the memory, the 
> display, the BIOS, the power company (or an equivalent hand-crank) etc etc 
> etc, and these things are all very much dependent on each other to make a 
> "usable system", that has absolutely _zero_ relevance to whether they are 
> "mere aggregation" or not.

Ok, can I please rewrite my argument to:

"The hardware manufacturer has built a custom BIOS and also written
Linux kernel support for said BIOS.  They have released the kernel
drivers under GPL as required, but have not released the code to the
BIOS, instead just releasing the interface documentation.  The BIOS
didn't exist before, and as they only intend to run Linux on the
device, the BIOS design was heavily influenced around working well
with Linux."

Actually, we don't know that last bit, maybe they created the BIOS
in a total vacuum and then wrote the Linux kernel driver later.
Maybe not.

Anyway, I think I've wound up arguing two sides of the same argument,
oops.

Bron.

RE: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Schwartz]

> Ok, can I please rewrite my argument to:
>
> "The hardware manufacturer has built a custom BIOS and also written
> Linux kernel support for said BIOS.  They have released the kernel
> drivers under GPL as required, but have not released the code to the
> BIOS, instead just releasing the interface documentation.  The BIOS
> didn't exist before, and as they only intend to run Linux on the
> device, the BIOS design was heavily influenced around working well
> with Linux."
>
> Actually, we don't know that last bit, maybe they created the BIOS
> in a total vacuum and then wrote the Linux kernel driver later.
> Maybe not.
>
> Anyway, I think I've wound up arguing two sides of the same argument,
> oops.
>
> Bron.

Why does it matter whether the BIOS was "heavily influenced around working
well" with Linux? Your argument is arguing about things that have nothing to
do with anything. Your comparing things before you have any idea what the
right criteria for the comparison is.

Do you understand that all that matters is whether the BIOS contains
significant portions of the Linux code base? *Nothing* else matters.
Everything else will leave it an independent work and one that the authors
of Linux have nor right to claim any ownership of or control over.

I maintain a private application that has huge amounts of code that are
heavily influenced around working well with Linux. All the epoll code, for
example, in this code base meets that criteria. That doesn't mean it
*resembles* the Linux code in any way. It doesn't mean the Linux folks have
any right to tell me what I can and can't do with that code.

So you're arguing two sides of no argument at all.

DS

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bron Gondwana]

On Mon, Jun 18, 2007 at 01:03:40PM -0700, David Schwartz wrote:
> So you're arguing two sides of no argument at all.

Yeah, pretty much.  I take back my arguments in the previous
couple of my posts up this thread.  They don't actually hold
together!  Sorry for wasting your time correct me.

Bron.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bron Gondwana]

On Sat, Jun 16, 2007 at 05:58:11PM -0700, Linus Torvalds wrote:
> 
> 
> On Sun, 17 Jun 2007, Bron Gondwana wrote:
> > 
> > No, I'm arguing that it's not "mere aggregation" - the kernel is useless
> > on that machine unless the BIOS is present or replaced with something
> > else with equivalent functionality.
> 
> That's *not* a valid argument!
> 
> I know, I know, it's a common one, but it is *nonsense*.

Further to my other response on this (yeah, I know, I should think
first).

Where the BIOS author and 'work aggregator' are different organisation
with no shady backroom links (other than the usual industry cabal(TINC)
of course) then it's clear.

When they are the same organisation then the derivedness state is a lot
less clear and more "discoverable", leading to a higher risk of ambush
by litigation.

This isn't specific to any particular licence, but it's something that
the "intent" theory of hardware limitations being a GPL3 violation makes
extra dangerous, because that clause can be used as a hook to drag a
claim through summary judgement (IMHO, IANAL, etc)

Bron ( mostly arguing the same things that you are Linus, I think, but
       I didn't clarify that I was writing from a devil's advocate
       position in an alternative reality where Tivo was illegal
       purely due to "intent" )

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 16, 2007, Bron Gondwana <brong@fastmail.fm> wrote:

> I obviously wasn't clear enough.  The only way to come into complience
> with GPL3dd4 is to reduce your ability to fix things or grant everyone
> else the ability to mess with things.  This severely restricts you from
> doing _anything_ in certain problem spaces due to local laws on the
> topic, even if you're an otherwise good actor who is making worthwhile
> source code contributions to the rest of the community.

I don't know any law that requires tivoization.

There may be laws that require certification or limitations on the
user.  Manufacturer giving up the ability to make modifications would
address this, or *perhaps* arranging for user and manufacturer to each
hold half of the key needed to run a modification (which might comply
with the GPLv3dd4 terms, IANAL).

There may be business models that require the ability to make changes.
Then it's fair to enable the user to make changes as well, such that
they don't become dependent on the vendor, or even have their
1st-generation TiVo boxes left out in the cold for a while when the US
changes the DST rules again ;-)

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Saturday 16 June 2007 21:54:56 Alexandre Oliva wrote:
> On Jun 16, 2007, Bron Gondwana <brong@fastmail.fm> wrote:
> > I obviously wasn't clear enough.  The only way to come into complience
> > with GPL3dd4 is to reduce your ability to fix things or grant everyone
> > else the ability to mess with things.  This severely restricts you from
> > doing _anything_ in certain problem spaces due to local laws on the
> > topic, even if you're an otherwise good actor who is making worthwhile
> > source code contributions to the rest of the community.
>
> I don't know any law that requires tivoization.
>
> There may be laws that require certification or limitations on the
> user.  Manufacturer giving up the ability to make modifications would
> address this, or *perhaps* arranging for user and manufacturer to each
> hold half of the key needed to run a modification (which might comply
> with the GPLv3dd4 terms, IANAL).

It doesn't. The GPLv3 (dd4) makes that very clear. See the quote below.

"Installation Information" for a User Product means any methods, procedures, 
authorization keys, or other information required to install and execute 
modified versions of a covered work in that User Product from a modified 
version of its Corresponding Source. The information must suffice to ensure 
that the continued functioning of the modified object code is in no case 
prevented or interfered with solely because modification has been made.

DRH

> There may be business models that require the ability to make changes.
> Then it's fair to enable the user to make changes as well, such that
> they don't become dependent on the vendor, or even have their
> 1st-generation TiVo boxes left out in the cold for a while when the US
> changes the DST rules again ;-)



-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 17, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> On Saturday 16 June 2007 21:54:56 Alexandre Oliva wrote:

>> There may be laws that require certification or limitations on the
>> user.  Manufacturer giving up the ability to make modifications would
>> address this, or *perhaps* arranging for user and manufacturer to each
>> hold half of the key needed to run a modification (which might comply
>> with the GPLv3dd4 terms, IANAL).

> It doesn't. The GPLv3 (dd4) makes that very clear. See the quote below.

You left out the relevant bit:

  this requirement does not apply if neither you nor any third party
  retains the ability to install modified object code on the User
  Product (for example, the work has been installed in ROM).

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Sunday 17 June 2007 00:19:49 Alexandre Oliva wrote:
> On Jun 17, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> > On Saturday 16 June 2007 21:54:56 Alexandre Oliva wrote:
> >> There may be laws that require certification or limitations on the
> >> user.  Manufacturer giving up the ability to make modifications would
> >> address this, or *perhaps* arranging for user and manufacturer to each
> >> hold half of the key needed to run a modification (which might comply
> >> with the GPLv3dd4 terms, IANAL).
> >
> > It doesn't. The GPLv3 (dd4) makes that very clear. See the quote below.
>
> You left out the relevant bit:
>
>   this requirement does not apply if neither you nor any third party
>   retains the ability to install modified object code on the User
>   Product (for example, the work has been installed in ROM).

Ah, but giving the user half the key doesn't mean they still don't have access 
to the entire key. QED: Giving people half the key won't cut it under the 
GPLv3 (dd4)

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 17, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> On Sunday 17 June 2007 00:19:49 Alexandre Oliva wrote:
>> On Jun 17, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
>> > On Saturday 16 June 2007 21:54:56 Alexandre Oliva wrote:
>> >> There may be laws that require certification or limitations on the
>> >> user.  Manufacturer giving up the ability to make modifications would
>> >> address this, or *perhaps* arranging for user and manufacturer to each
>> >> hold half of the key needed to run a modification (which might comply
>> >> with the GPLv3dd4 terms, IANAL).
>> >
>> > It doesn't. The GPLv3 (dd4) makes that very clear. See the quote below.
>> 
>> You left out the relevant bit:
>> 
>> this requirement does not apply if neither you nor any third party
>> retains the ability to install modified object code on the User
>> Product (for example, the work has been installed in ROM).

> Ah, but giving the user half the key doesn't mean they still don't have access 
> to the entire key. QED: Giving people half the key won't cut it under the 
> GPLv3 (dd4)

I meant really giving, rather than giving a copy, or giving the
original and keeping a copy.

You could make it require a pair of signatures, one from the vendor,
that the vendor keeps, one from the user, that the vendor never sees,
too.  Like some bank PINs, it gets generated, used to generate some
hash (the signature for the initial installation), printed in an
envelope for you and stored in the package along with the machine.  Or
something like that.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Bron Gondwana]

On Sun, Jun 17, 2007 at 02:38:43AM -0300, Alexandre Oliva wrote:
> On Jun 17, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> 
> > Ah, but giving the user half the key doesn't mean they still don't have access 
> > to the entire key. QED: Giving people half the key won't cut it under the 
> > GPLv3 (dd4)
> 
> I meant really giving, rather than giving a copy, or giving the
> original and keeping a copy.
> 
> You could make it require a pair of signatures, one from the vendor,
> that the vendor keeps, one from the user, that the vendor never sees,
> too.  Like some bank PINs, it gets generated, used to generate some
> hash (the signature for the initial installation), printed in an
> envelope for you and stored in the package along with the machine.  Or
> something like that.

Wow, and I thought losing a microsoft "certificate of authenticity" and
associated key was a pain.  Ouch.  Talk about your paper "dongle".

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alan Cox]

> I don't know any law that requires tivoization.

In the USSA it is arguable that wireless might need it (if done in
software) for certain properties. (The argument being it must be
tamperproof to random end consumers). Obviously an electronics graduate
can tamper with hardware ones just as well

The USSA also has interesting rules about some other technologies
including GPS on national security grounds.

Alan

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 17, 2007, Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:

>> I don't know any law that requires tivoization.

> In the USSA it is arguable that wireless might need it (if done in
> software) for certain properties. (The argument being it must be
> tamperproof to random end consumers).

But this is not tivoization.

Tivoization is a manufacturer using technical measures to prevent the
user from tampering (*) with the device, *while* keeping the ability
to tamper with it changes itself.

(*) tampering brings in negative connotations that I'd rather avoid,
but since that was the term you used, and the term "modifying" might
bring in legal-based technicalities such as that replacing isn't
modification, I just went with it.


So, given a proper definition, do you know any law that requires
tivoization?

Taking it further, do you know whether any such law requires
*worldwide* tivoization, as in, applying the restrictions in the law
even outside its own jurisdiction?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alan Cox]

On Sun, 17 Jun 2007 15:33:33 -0300
Alexandre Oliva <aoliva@redhat.com> wrote:

> On Jun 17, 2007, Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:
> 
> >> I don't know any law that requires tivoization.
> 
> > In the USSA it is arguable that wireless might need it (if done in
> > software) for certain properties. (The argument being it must be
> > tamperproof to random end consumers).
> 
> But this is not tivoization.

It most definitely is

> Tivoization is a manufacturer using technical measures to prevent the
> user from tampering (*) with the device, *while* keeping the ability
> to tamper with it changes itself.

That accurately describes the FCC wireless rules. As a vendor I am
perfectly allowed to tinker, but the product must not be "end user"
modifiable. (and some vendors take this to mean binary only, despite the
fact end users generally aren't able to modify code and EE students are
just as able to modify electronics as we are code).

Alan

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 17, 2007, Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:

> On Sun, 17 Jun 2007 15:33:33 -0300
> Alexandre Oliva <aoliva@redhat.com> wrote:

>> On Jun 17, 2007, Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:
>> 
>> >> I don't know any law that requires tivoization.
>> 
>> > In the USSA it is arguable that wireless might need it (if done in
>> > software) for certain properties. (The argument being it must be
>> > tamperproof to random end consumers).
>> 
>> But this is not tivoization.

> It most definitely is

>> Tivoization is a manufacturer using technical measures to prevent the
>> user from tampering (*) with the device, *while* keeping the ability
>> to tamper with it changes itself.

> That accurately describes the FCC wireless rules.

AFAIK the FCC mandates not permitting the user to tinker.  It doesn't
mandate the vendor to retain this ability to itself.

Therefore, per the above, FCC doesn't mandate tivoization.

Is there anything else I'm missing that would show it does?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alan Cox]

> > That accurately describes the FCC wireless rules.
> 
> AFAIK the FCC mandates not permitting the user to tinker.  It doesn't
> mandate the vendor to retain this ability to itself.

In practical terms it does since a recall/replacement in the event of
rule changes is a bit impractical

> Therefore, per the above, FCC doesn't mandate tivoization.

I'm sure you can find a definition to sort your goals whatever.

mea culpa on the meaning of Tivoization (was: Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3)

[Alexandre Oliva]

On Jun 17, 2007, Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:

>> > That accurately describes the FCC wireless rules.
>> 
>> AFAIK the FCC mandates not permitting the user to tinker.  It doesn't
>> mandate the vendor to retain this ability to itself.

> In practical terms it does since a recall/replacement in the event of
> rule changes is a bit impractical

Indeed.  But that's not a legal requirement, it's an economic reason.

"But I need to make a profit" or "But I need to reduce costs" is no
excluse to disrespect the GPL.


>> Therefore, per the above, FCC doesn't mandate tivoization.

> I'm sure you can find a definition to sort your goals whatever.

Are you per chance implying that I'm twisting the definition of
tivoization?


You know...  I now believe that would be correct.  I have indeed
twisted the definition of tivoization, and I'm sorry about that.
Which is not to say that I agree that the FCC or any other law
mandates tivoization, or that tivozation is a good thing or that it is
permitted by GPLv2.  Please read on.


After long conversations with RMS about the section on poisoned apples
and tivoization in my draft article about GPLv3 (Corresponding Sources
is the name of the section in
http://fsfla.org/svnwiki/blogs/lxo/draft/gplv3-snowwhite) I had come
to the conclusion that Tivoization amounted to:

  denying the user of the computer the freedom to run modified
  versions of the Free Software in it, while retain this ability to
  oneself.

This understanding of mine had been strengthened by my understanding
of the wording and the rationale of GPLv3dd3, the wording about
technical restrictions in the rationales published along with
GPLv2dd2, and the various speeches in which the term was presented.

Nevertheless, I consulted with him and others highly involved in the
development of GPLv3 about some of the discussions going on here, and
got responses over the past few hours that surprised me.  A lot.

So I've just went back to that discussion about my article, and to
various other cases in which RMS, Eben Moglen and others presented
Tivoization, the rationales, and so on, and I came to the conclusion
that I had experienced a subtle but very significant misunderstanding.

I'm now convinced that a more appropriate definition would be:

  denying the user of the computer the freedom to run modified
  versions of the Free Software in it, by not sharing information as
  to how it could be accomplished.


This difference is very significant, and even more so for this
discusion, because it contradicts some of what I claimed before about
forms to use GPLed software where regulations require the user to be
unable to modify the software.


Let me start with an example: I bought a wireless router some time
ago, and it had a GNU+Linux distribution installed in it.  No source
code or written offer for source code, though.

Now, if I called the vendor next day and asked for the source code,
and they responded "sorry, I can't give you that.  I threw it all
away, such that I wouldn't be able to give it to you.", they would
still be disrespecting my freedoms, as well as the license, right?

You see where I'm going?  Now, if they gave me the source code, but I
still couldn't install modified versions, because they introduced
technical measures with the purpose of denying me this possibility,
then the inability to modify the program wouldn't be caused by
something like a physical impossibility (something like ROM), but
rather by an active measure to trample my ability to adapt the program
and run it for any purpose.

So, if I called them to ask how to install and run modified versions
of the GPLed programs, and they responded "sorry, I can't give you
that.  I threw it all away, such that I wouldn't be able to give it to
you.", they would still be disrespecting my freedoms, as well as the
license.

The reasons as to why they'd want to disrespect the freedoms don't
matter.  It could be "making a profit", "complying with the law",
"abiding by contractual restrictions", anything.  Imposing
restrictions to the exercise of the freedoms is not in line with the
spirit of the GPL, as such restrictions render the Software non-Free.


The conclusion?  Throwing keys away, or using split-key techniques, as
I have suggested as potential alternatives to ROM for compliance with
GPLv3 are not meant to be permitted by GPLv3.  There might be
practical advantages to compromising and permitting these techniques,
but that would amount to endorsing disrespect for users' freedoms, and
more, betraying those who licensed their works under GPLv1+ or v2+
with an intent to not permit these practices.  I don't think FSF is
willing to be part of this, and this is how it should be.

As for those who didn't mean the GPL this way, they can always grant
additional permissions, or simply refrain from enforcing the license
in these cases.


I apologize for my terrible misunderstanding, and for spreading it.

Hopefully this message will reach everyone I have misled.

I've tried to Cc: everyone who'd received copies of my mistaken claims
directly from me.  If I left you out by accident, please holler ;-)

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: mea culpa on the meaning of Tivoization (was: Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3)

[Alan Cox]

> > In practical terms it does since a recall/replacement in the event of
> > rule changes is a bit impractical
> 
> Indeed.  But that's not a legal requirement, it's an economic reason.

Cynical Economists would argue 'legal requirements' are just changes to
the cost of the various economic options. Sometimes when I look at
Microsoft's approach to various cases it seems they think that way too.

>   denying the user of the computer the freedom to run modified
>   versions of the Free Software in it, by not sharing information as
>   to how it could be accomplished.

That's a nice definition but one I can see being sort of abusable
depending how you read it. We head ever more into the disposable computer
era where as a vendor putting the code on ROM is cheap and upgrades don't
matter (throw it away get a new one).

I can tell you how to upgrade it ("you can't") yet I as the manufacturer
can issue new units with modified code so I still control it even though
it is meant to be "free"

Re: mea culpa on the meaning of Tivoization

[Alexandre Oliva]

On Jun 17, 2007, Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:

>> > In practical terms it does since a recall/replacement in the event of
>> > rule changes is a bit impractical
>> 
>> Indeed.  But that's not a legal requirement, it's an economic reason.

> Cynical Economists would argue 'legal requirements' are just changes to
> the cost of the various economic options. Sometimes when I look at
> Microsoft's approach to various cases it seems they think that way too.

I guess this depends to some point on the kind of penalties you face.
If they're only economic, then yes.  If you may end up going to jail
or some such, I think the picture gets different.  But yes, that's a
way to see it, and I know we're not alone in perceiving some behaviors
that way.

> I can tell you how to upgrade it ("you can't") yet I as the manufacturer
> can issue new units with modified code so I still control it even though
> it is meant to be "free"

The GPL has never prohibited the distribution of software in ROM, just
like it's never prohibited the fixation of software in CD-ROMs.  So
explicitly permitting is not a step back in terms of defending
freedoms, even if there might be something to do that would advance
freedoms in this field.

Anyhow, AFAIK software in ROM is not non-Free Software.  That it's
impossible to modify/replace/whathaveyou it is not the result of a
restriction that someone is imposing on you.

It's the difference between "you can't fly because you don't have
wings" and "you won't fly because I've tied your wings".  With tied
wings, you're evidently not free to fly any more.  But if the problem
is that you don't have wings, if you're free and sufficiently
creative, you may be able to invent baloons, airplanes, rockets et al
and overcome the barriers that nature poses for you.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: mea culpa on the meaning of Tivoization

[Hans-Jürgen Koch]

Am Montag 18 Juni 2007 02:56 schrieb Alexandre Oliva:

> 
> Anyhow, AFAIK software in ROM is not non-Free Software.  That it's
> impossible to modify/replace/whathaveyou it is not the result of a
> restriction that someone is imposing on you.
> 
> It's the difference between "you can't fly because you don't have
> wings" and "you won't fly because I've tied your wings".  With tied
> wings, you're evidently not free to fly any more.  But if the problem
> is that you don't have wings, if you're free and sufficiently
> creative, you may be able to invent baloons, airplanes, rockets et al
> and overcome the barriers that nature poses for you.
> 

So, if a manufacturer used a ROM instead of a flash memory with the
intention to make software modifications impossible, then it is bad,
and when he did it for economical reasons, then it is a "natural barrier"?

Your tied-up-wings comparison is simply not valid. Or, more precisely,
you will usually not be able to tell whether you don't have wings or
if they're tied. Hardware design decisions of a manufacturer should
never be the subject of a software license.

Hans

Re: mea culpa on the meaning of Tivoization

[Alexandre Oliva]

On Jun 18, 2007, Hans-Jürgen Koch <hjk@linutronix.de> wrote:

> So, if a manufacturer used a ROM instead of a flash memory with the
> intention to make software modifications impossible, then it is bad,
> and when he did it for economical reasons, then it is a "natural barrier"?

This sounds about right to me.

Intent is very significant, but then, what vendor would justify the
choice of ROM as "intent to prevent modifications", if this amounted
to copyright infringement?

Vendor would be entitled to the benefit of the doubt as to the
motivations in this case, so it would likely be unenforceable anyway.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: mea culpa on the meaning of Tivoization

[Hans-Jürgen Koch]

Am Montag 18 Juni 2007 20:55 schrieb Alexandre Oliva:
> On Jun 18, 2007, Hans-Jürgen Koch <hjk@linutronix.de> wrote:
> 
> > So, if a manufacturer used a ROM instead of a flash memory with the
> > intention to make software modifications impossible, then it is bad,
> > and when he did it for economical reasons, then it is a "natural barrier"?
> 
> This sounds about right to me.
> 
> Intent is very significant, but then, what vendor would justify the
> choice of ROM as "intent to prevent modifications", if this amounted
> to copyright infringement?

Indeed. That nicely shows how useless any licensing discussion is
when it comes to hardware design issues, including "Tivoization". 

> 
> Vendor would be entitled to the benefit of the doubt as to the
> motivations in this case, so it would likely be unenforceable anyway.
>

Right. If GPL v3 comes out, there'll probably be a new task for
hardware development engineers: How to find excuses for hardware that
prevents software modifications and how to conceal the true intent.

Hans
 

Re: mea culpa on the meaning of Tivoization

[Alexandre Oliva]

On Jun 18, 2007, Hans-Jürgen Koch <hjk@linutronix.de> wrote:

>> Vendor would be entitled to the benefit of the doubt as to the
>> motivations in this case, so it would likely be unenforceable anyway.

> Right. If GPL v3 comes out, there'll probably be a new task for
> hardware development engineers: How to find excuses for hardware that
> prevents software modifications and how to conceal the true intent.

Yup.  And then GPLv4 will have to plug whatever holes they find to
disrespect users' freedoms.  That's how I expect the game to be
played.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: mea culpa on the meaning of Tivoization

[Hans-Jürgen Koch]

Am Montag 18 Juni 2007 23:18 schrieb Alexandre Oliva:
> On Jun 18, 2007, Hans-Jürgen Koch <hjk@linutronix.de> wrote:
> 
> >> Vendor would be entitled to the benefit of the doubt as to the
> >> motivations in this case, so it would likely be unenforceable anyway.
> 
> > Right. If GPL v3 comes out, there'll probably be a new task for
> > hardware development engineers: How to find excuses for hardware that
> > prevents software modifications and how to conceal the true intent.
> 
> Yup.  And then GPLv4 will have to plug whatever holes they find to
> disrespect users' freedoms.  That's how I expect the game to be
> played.
>

If you were right and it turned out that way, the whole GPL would
become so ridiculous that it won't have any of its intended effects.
GPLv17 would contain so many #ifdefs that it'll simply not be readable
or understandable anymore. 

As far as the kernel is concerned, I expect the game's played by
simply keeping GPLv2. And I like it that way.

Hans

Re: mea culpa on the meaning of Tivoization

[Alexandre Oliva]

On Jun 18, 2007, Hans-Jürgen Koch <hjk@linutronix.de> wrote:

> Am Montag 18 Juni 2007 23:18 schrieb Alexandre Oliva:
>> On Jun 18, 2007, Hans-Jürgen Koch <hjk@linutronix.de> wrote:
>> 
>> >> Vendor would be entitled to the benefit of the doubt as to the
>> >> motivations in this case, so it would likely be unenforceable anyway.
>> 
>> > Right. If GPL v3 comes out, there'll probably be a new task for
>> > hardware development engineers: How to find excuses for hardware that
>> > prevents software modifications and how to conceal the true intent.
>> 
>> Yup.  And then GPLv4 will have to plug whatever holes they find to
>> disrespect users' freedoms.  That's how I expect the game to be
>> played.

> If you were right and it turned out that way, the whole GPL would
> become so ridiculous that it won't have any of its intended effects.

How so?  The intended effects are to protect users' freedoms, by
requiring them to be respected.  If we keep on plugging holes as they
appear, it will keep close to achieving its intended effects.  It's
earlier versions of the license that will get more and more distant
from it.

> As far as the kernel is concerned, I expect the game's played by
> simply keeping GPLv2. And I like it that way.

Just think about it...  What if, today, some law passed, or some court
decision came up, that rendered a significant defense provision of
GPLv2 or GPLv3 ineffective?

GPLv4 could plug that, and anyone using GPLvN+ would be able to switch
to it immediately.  This wouldn't revoke previous licenses, of course,
but further developments could be made under the newer license, and at
least those could still be defended, and, as time elapsed, earlier
versions of the software would become less and less relevant, to the
point that the holes in their license also become less and less
relevant, until copyright finally expires and they enter the public
domain.


The distrust for the FSF led to this very short-sighted decision of
painting the Linux community into a corner from which it is very
unlikely to be able to ever leave, no matter how badly it turns out to
be needed.  Let's just hope it never is, or that some influx of
long-sighted comes in and introduces mechanisms for the license of
Linux to be patched, should this ever be needed.  I'm not even talking
about GPLv2+, there are many other ways to accomplish this, that I've
already mentioned in another posting in another recent huge thread.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: mea culpa on the meaning of Tivoization

[Hans-Jürgen Koch]

Am Dienstag 19 Juni 2007 04:46 schrieb Alexandre Oliva:
> On Jun 18, 2007, Hans-Jürgen Koch <hjk@linutronix.de> wrote:
> 
> > Am Montag 18 Juni 2007 23:18 schrieb Alexandre Oliva:
> >> On Jun 18, 2007, Hans-Jürgen Koch <hjk@linutronix.de> wrote:
> >> 
> >> >> Vendor would be entitled to the benefit of the doubt as to the
> >> >> motivations in this case, so it would likely be unenforceable anyway.
> >> 
> >> > Right. If GPL v3 comes out, there'll probably be a new task for
> >> > hardware development engineers: How to find excuses for hardware that
> >> > prevents software modifications and how to conceal the true intent.
> >> 
> >> Yup.  And then GPLv4 will have to plug whatever holes they find to
> >> disrespect users' freedoms.  That's how I expect the game to be
> >> played.
> 
> > If you were right and it turned out that way, the whole GPL would
> > become so ridiculous that it won't have any of its intended effects.
> 
> How so?  The intended effects are to protect users' freedoms, by
> requiring them to be respected.  If we keep on plugging holes as they
> appear, it will keep close to achieving its intended effects.  

No. Credible licenses should be simple like physical laws. Newton's law
is expressed in terms of a single mathematical equation. That's why it's
still valid, and you still learn it at school although meanwhile people
know that there are limitations to it.

If you come up with a new version of a license every year, you will only
weaken it. Please note that quantum mechanics is _not_ such a hole-plugging
addition to Newton's law. It's a new simple physical law, expressed in terms
of a single simple mathematical equation that contains the old law as a
border case. If that were not the case, it would have never been accepted.

If you want a GPLv3, please make it simple and make it contain GPLv2 as
a border case. The current draft isn't like that.

> It's 
> earlier versions of the license that will get more and more distant
> from it.

No. GPLv2 is a simple set of rights and restrictions that's easy to
understand and therefore accepted by many courts all over the world.
I cannot see any danger for the code I put under it, at least none
that would be mitigated by GPLv3.

> 
> > As far as the kernel is concerned, I expect the game's played by
> > simply keeping GPLv2. And I like it that way.
> 
> Just think about it...  What if, today, some law passed, or some court
> decision came up, that rendered a significant defense provision of
> GPLv2 or GPLv3 ineffective?

The best way to prevent that is to make the license simple and easy to 
follow. If many important open source software developers have problems
with the wording of a license, lawyers and judges will have them, too.

> 
> GPLv4 could plug that, and anyone using GPLvN+ would be able to switch
> to it immediately.  This wouldn't revoke previous licenses, of course,
> but further developments could be made under the newer license, and at
> least those could still be defended, and, as time elapsed, earlier
> versions of the software would become less and less relevant, to the
> point that the holes in their license also become less and less
> relevant, until copyright finally expires and they enter the public
> domain.

Ah, now I get the point. Yes, that's one in favor of a complicated and
confusing license. You can lengthen court proceedings until copyright
expires...

> 
> 
> The distrust for the FSF led to this very short-sighted decision of
> painting the Linux community into a corner from which it is very
> unlikely to be able to ever leave, no matter how badly it turns out to
> be needed.  

I'm neither in a corner nor do I feel the need for a different license.
I've got some code in the kernel, and I've got it under GPLv2, and I'm
happy with it. It's the FSF that thinks I should see myself in a corner. 


> Let's just hope it never is, or that some influx of 
> long-sighted comes in 

Kernel programmers are short-sighted? What kind of arrogance is that?

> and introduces mechanisms for the license of 
> Linux to be patched, should this ever be needed.  

You know pretty well that Linus clearly said he would change the license
when _he_ thinks it's needed. The point is that _you_ want him to change
the license to support _your_ political ideas.

> I'm not even talking 
> about GPLv2+, there are many other ways to accomplish this, that I've
> already mentioned in another posting in another recent huge thread.
>

I partly read that "recent huge thread". Linus elaborated his point of 
view in detail, and I very much share his opinion.

Hans

Re: mea culpa on the meaning of Tivoization

[Michael Poole]

Hans-Jürgen Koch writes:

> No. Credible licenses should be simple like physical laws. Newton's law
> is expressed in terms of a single mathematical equation. That's why it's
> still valid, and you still learn it at school although meanwhile people
> know that there are limitations to it.
>
> If you come up with a new version of a license every year, you will only
> weaken it. Please note that quantum mechanics is _not_ such a hole-plugging
> addition to Newton's law. It's a new simple physical law, expressed in terms
> of a single simple mathematical equation that contains the old law as a
> border case. If that were not the case, it would have never been accepted.

This is an excellent example of how engineers tend to mis-analyze
legal issues.  In law, neither simple wording nor interpretation is so
simple or so mechanical as the things engineers prefer to work with.

Take an example: "Thou shalt not kill".  Very clear, but also very
problematic in that it does not address military conquests (which were
apparently approved by that law's drafter), self-defense, or a number
of other cases.

There are always grey areas between what is explicitly addressed and
what is not.  Courts interpret laws and precedent in ways that make
life (and license or contract writing) more unpredictable.  Wishing
otherwise will not make a simple license unambiguous.  If those areas
of ambiguity are exploited enough, the perceived cost of having a hole
will exceed the perceived cost of plugging it.

Michael Poole

Re: mea culpa on the meaning of Tivoization

[Hans-Jürgen Koch]

Am Dienstag 19 Juni 2007 13:50 schrieb Michael Poole:
> Hans-Jürgen Koch writes:
> 
> > No. Credible licenses should be simple like physical laws. Newton's law
> > is expressed in terms of a single mathematical equation. That's why it's
> > still valid, and you still learn it at school although meanwhile people
> > know that there are limitations to it.
> >
> > If you come up with a new version of a license every year, you will only
> > weaken it. Please note that quantum mechanics is _not_ such a hole-plugging
> > addition to Newton's law. It's a new simple physical law, expressed in terms
> > of a single simple mathematical equation that contains the old law as a
> > border case. If that were not the case, it would have never been accepted.
> 
> This is an excellent example of how engineers tend to mis-analyze
> legal issues.  In law, neither simple wording nor interpretation is so
> simple or so mechanical as the things engineers prefer to work with.

Don't make fast assumptions about how _I_ think. I intentionally chose
examples from science, not from engineering. There are grey areas in
Newton's Law and in quantum mechanics as well.

> 
> Take an example: "Thou shalt not kill".  Very clear, but also very
> problematic in that it does not address military conquests (which were
> apparently approved by that law's drafter), self-defense, or a number
> of other cases.

Pretty clear. I accepted that about thirty years ago.

> 
> There are always grey areas between what is explicitly addressed and
> what is not.  

Yes, you named it: _always_ !! We have to accept these grey areas.
Tivoization (to come back to our real topic) is one of those.
Alexandre Oliva sees grey areas as holes he needs to plug. That's
wrong, and that's what I'm criticizing. You cannot get rid of grey
areas by modifying licenses. Maybe you can temporarily shift the grey 
areas to some other place. But at the same time, you get negative
side effects because your license becomes more and more complicated
and confusing. I consider this worse than some grey areas.

> Courts interpret laws and precedent in ways that make 
> life (and license or contract writing) more unpredictable.  Wishing
> otherwise will not make a simple license unambiguous.  If those areas
> of ambiguity are exploited enough, the perceived cost of having a hole
> will exceed the perceived cost of plugging it.

I completely disagree. A very important factor for the success of a license
is that people are familiar with it. Only then can they start obeying it.
Have you ever been out there in industry, trying to tell programmers what
Linux, Free Software, and the GPL is? You'd notice that they hardly know
what the GPLv2 is all about. And now, after I explained it to them, should
I tell them that there's a much more complicated license about to come, and
that it's going to be changed whenever some FSF people find a new hole?

All this hole-plugging just leads to a license that is so far away from
being useful in industrial reality that nobody will voluntarily obey it.
This Tivoization stuff is a good example. I was a hardware developer for
more than a decade. When I developed _hardware_, I made my design decisions
without having to read the licenses of the _software_ I want to run on
that hardware. And if you go today and tell a hardware developer that there
are some people in the world who want to achieve exactly that, he will
probably laugh at you and think you're joking.

Hans

Re: mea culpa on the meaning of Tivoization

[Alexandre Oliva]

On Jun 19, 2007, Hans-Jürgen Koch <hjk@linutronix.de> wrote:

> Am Dienstag 19 Juni 2007 04:46 schrieb Alexandre Oliva:
>> The distrust for the FSF led to this very short-sighted decision of
>> painting the Linux community into a corner from which it is very
>> unlikely to be able to ever leave, no matter how badly it turns out to
>> be needed.  

> I'm neither in a corner nor do I feel the need for a different license.

Yes.  Some day you may.  And then what will you or anyone else be able
to do about it?

>> Let's just hope it never is, or that some influx of 
>> long-sighted comes in 

> Kernel programmers are short-sighted? What kind of arrogance is that?

It's just stating the obvious.  The upgrade path is a nightmare.

A long-sighted decision should have established *some* means for a
quick fix should it be needed.  It didn't have to be GPLv2+.  In fact,
per the stated goals and general feelings, it probably *shouldn't* be
GPLv2+.  But cutting any reasonable possibility of fixing a legal
problem in the license is short-sighted, yes.  It's putting too much
trust in the perfection of the license *and* the worldwide legal
systems *and* legislators.

>> and introduces mechanisms for the license of 
>> Linux to be patched, should this ever be needed.  

> You know pretty well that Linus clearly said he would change the license
> when _he_ thinks it's needed.

And what makes you think even *he* can change the license?

> The point is that _you_ want him to change the license to support
> _your_ political ideas.

I would like him to, yes.  But this is besides the point.  That I see
reasons for an upgrade, and that I'd like such an upgrade, doesn't
make any difference whatsoever about the plain fact that relicensing
Linux today, to any other license and for whatever reason it was,
would be a nightmare, and that this is a consequence of the
short-sighted decision of not establishing a relicensing procedure.

At this point, the situation is very much like a kernel installed in
ROM.  Who knows that nobody will ever find security bugs in it?  How
would you go about fixing them?

>> I'm not even talking 
>> about GPLv2+, there are many other ways to accomplish this, that I've
>> already mentioned in another posting in another recent huge thread.

> I partly read that "recent huge thread". Linus elaborated his point of 
> view in detail, and I very much share his opinion.

Huh?  It looks like you're talking about something unrelated with
license patching procedures.  I don't think Linus ever responded in
that thread to my suggestions of various means to establish a license
patching procedure.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: mea culpa on the meaning of Tivoization

[Hans-Jürgen Koch]

Am Dienstag 19 Juni 2007 20:32 schrieb Alexandre Oliva:
> On Jun 19, 2007, Hans-Jürgen Koch <hjk@linutronix.de> wrote:
> 
> > Am Dienstag 19 Juni 2007 04:46 schrieb Alexandre Oliva:
> >> The distrust for the FSF led to this very short-sighted decision of
> >> painting the Linux community into a corner from which it is very
> >> unlikely to be able to ever leave, no matter how badly it turns out to
> >> be needed.  
> 
> > I'm neither in a corner nor do I feel the need for a different license.
> 
> Yes.  Some day you may.  And then what will you or anyone else be able
> to do about it?

A lot of awful things could happen to you or me _tomorrow_. Are you prepared
for everything? I'm not. If I'd try to be prepared for every possible 
disaster, you'd rightly call me mentally ill.

Even if my code is used on a Tivo-like device, I can sleep well and don't
need a different license. These are the _normal_ grey areas that _every_
license or law contains. 

> 
> >> Let's just hope it never is, or that some influx of 
> >> long-sighted comes in 
> 
> > Kernel programmers are short-sighted? What kind of arrogance is that?
> 
> It's just stating the obvious.  The upgrade path is a nightmare.

Well, maybe. Maybe this is a topic that needs further discussion.
But I don't find it very important as we're not in a situation where
we urgently need a new license.

[...]

Hans

Re: mea culpa on the meaning of Tivoization

[Alexandre Oliva]

On Jun 19, 2007, Hans-Jürgen Koch <hjk@linutronix.de> wrote:

>> >> Let's just hope it never is, or that some influx of 
>> >> long-sighted comes in 
>> 
>> > Kernel programmers are short-sighted? What kind of arrogance is that?
>> 
>> It's just stating the obvious.  The upgrade path is a nightmare.

> Well, maybe. Maybe this is a topic that needs further discussion.
> But I don't find it very important as we're not in a situation where
> we urgently need a new license.

Agreed.  It could have been discussed years ago, when the
clarification on GPLv2-only came up, but it's still not urgent, and
hopefully it never will be.  It wouldn't be a bad idea to think about
it, though.

Just in case it's not clear, this is in no way related with GPLv3.
It's just that GPLv3 discussions appear to get more people thinking
about relicensing, and then the "impossibilities" of doing it come up.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: mea culpa on the meaning of Tivoization (was: Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3)

[Rob Landley]

On Sunday 17 June 2007 19:50:41 Alan Cox wrote:
> That's a nice definition but one I can see being sort of abusable
> depending how you read it. We head ever more into the disposable computer
> era where as a vendor putting the code on ROM is cheap and upgrades don't
> matter (throw it away get a new one).

I'm waiting for cereal boxes in the store to have a display covering the 
entire front side which changes every 30 seconds, with all the circuitry to 
drive it taking up a few square milimeters and a small watch battery to power 
it for six months ala the blinky LEDs of today.

This should be economicaly feasible in what, 10 years?  15?  (There are all 
sorts of weird problems to solve like coming up with a battery that's not 
only cheap enough but which you won't get in trouble putting millions of in 
landfills.  Fuel cell might be more environmentally friendly depending on 
your catalysts...)

We haven't even brushed against the "disposable computer era" compared to 
what's coming.  Most of us will probably live to see happy meal toys capable 
of running Linux, and that's just a _start_.

Rob
-- 
"One of my most productive days was throwing away 1000 lines of code."
  - Ken Thompson.

Re: mea culpa on the meaning of Tivoization (was: Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3)

[Daniel Hazelton]

On Sunday 17 June 2007 19:11:13 Alexandre Oliva wrote:
> On Jun 17, 2007, Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:
> >> > That accurately describes the FCC wireless rules.
> >>
> >> AFAIK the FCC mandates not permitting the user to tinker.  It doesn't
> >> mandate the vendor to retain this ability to itself.
> >
> > In practical terms it does since a recall/replacement in the event of
> > rule changes is a bit impractical
>
> Indeed.  But that's not a legal requirement, it's an economic reason.
>
> "But I need to make a profit" or "But I need to reduce costs" is no
> excluse to disrespect the GPL.
>
> >> Therefore, per the above, FCC doesn't mandate tivoization.
> >
> > I'm sure you can find a definition to sort your goals whatever.
>
> Are you per chance implying that I'm twisting the definition of
> tivoization?
>
>
> You know...  I now believe that would be correct.  I have indeed
> twisted the definition of tivoization, and I'm sorry about that.
> Which is not to say that I agree that the FCC or any other law
> mandates tivoization, or that tivozation is a good thing or that it is
> permitted by GPLv2.  Please read on.
>
>
> After long conversations with RMS about the section on poisoned apples
> and tivoization in my draft article about GPLv3 (Corresponding Sources
> is the name of the section in
> http://fsfla.org/svnwiki/blogs/lxo/draft/gplv3-snowwhite) I had come
> to the conclusion that Tivoization amounted to:
>
>   denying the user of the computer the freedom to run modified
>   versions of the Free Software in it, while retain this ability to
>   oneself.
>
> This understanding of mine had been strengthened by my understanding
> of the wording and the rationale of GPLv3dd3, the wording about
> technical restrictions in the rationales published along with
> GPLv2dd2, and the various speeches in which the term was presented.
>
> Nevertheless, I consulted with him and others highly involved in the
> development of GPLv3 about some of the discussions going on here, and
> got responses over the past few hours that surprised me.  A lot.
>
> So I've just went back to that discussion about my article, and to
> various other cases in which RMS, Eben Moglen and others presented
> Tivoization, the rationales, and so on, and I came to the conclusion
> that I had experienced a subtle but very significant misunderstanding.
>
> I'm now convinced that a more appropriate definition would be:
>
>   denying the user of the computer the freedom to run modified
>   versions of the Free Software in it, by not sharing information as
>   to how it could be accomplished.
>
>
> This difference is very significant, and even more so for this
> discusion, because it contradicts some of what I claimed before about
> forms to use GPLed software where regulations require the user to be
> unable to modify the software.
>
>
> Let me start with an example: I bought a wireless router some time
> ago, and it had a GNU+Linux distribution installed in it.  No source
> code or written offer for source code, though.

Just want to point out that, when I read this, my reaction was "But... That is 
a direct violation of the GPLv2. No specific reading of the license needed."

> Now, if I called the vendor next day and asked for the source code,
> and they responded "sorry, I can't give you that.  I threw it all
> away, such that I wouldn't be able to give it to you.", they would
> still be disrespecting my freedoms, as well as the license, right?

Yes, they would. They are distributing a modification - in the words of the 
GPLv2 "a work based on the work" - without complying with the terms of the 
license.

> You see where I'm going?  Now, if they gave me the source code, but I
> still couldn't install modified versions, because they introduced
> technical measures with the purpose of denying me this possibility,
> then the inability to modify the program wouldn't be caused by
> something like a physical impossibility (something like ROM), but
> rather by an active measure to trample my ability to adapt the program
> and run it for any purpose.
>
> So, if I called them to ask how to install and run modified versions
> of the GPLed programs, and they responded "sorry, I can't give you
> that.  I threw it all away, such that I wouldn't be able to give it to
> you.", they would still be disrespecting my freedoms, as well as the
> license.

Not even the GPLv3dd4 - because they don't have the information anymore 
either. If, however, they still retained the information - in any form - they 
would be violating the GPLv3dd4.

The GPLv2 doesn't make the actions described above - "how to install and 
run" - a license violation.

> The reasons as to why they'd want to disrespect the freedoms don't
> matter.  It could be "making a profit", "complying with the law",
> "abiding by contractual restrictions", anything.  Imposing
> restrictions to the exercise of the freedoms is not in line with the
> spirit of the GPL, as such restrictions render the Software non-Free.

Then anyone using GPLv3'd software to drive WiFi devices, radio (HAM radio) 
networks, etc - in the US, at least - isn't allowed to do such. US Law makes 
some provisions of the GPLv3 illegal to comply with. Thanks to section 6 of 
the GPLv3 that invalidates the rights granted under the license.

>
> The conclusion?  Throwing keys away, or using split-key techniques, as
> I have suggested as potential alternatives to ROM for compliance with
> GPLv3 are not meant to be permitted by GPLv3.  There might be
> practical advantages to compromising and permitting these techniques,
> but that would amount to endorsing disrespect for users' freedoms, and
> more, betraying those who licensed their works under GPLv1+ or v2+
> with an intent to not permit these practices.  I don't think FSF is
> willing to be part of this, and this is how it should be.

Umm... making things more strict will just do more damage. As it is there are 
restrictions on companies that make them unable to comply with the GPL. What 
the GPLv3 has done is take away options they might otherwise have had. If one 
of the goals of the FSF is to force proprietary software into a minority then 
its just done damage to that goal.


DRH

> As for those who didn't mean the GPL this way, they can always grant
> additional permissions, or simply refrain from enforcing the license
> in these cases.
>
>
> I apologize for my terrible misunderstanding, and for spreading it.
>
> Hopefully this message will reach everyone I have misled.
>
> I've tried to Cc: everyone who'd received copies of my mistaken claims
> directly from me.  If I left you out by accident, please holler ;-)



-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: mea culpa on the meaning of Tivoization

[Alexandre Oliva]

On Jun 17, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> On Sunday 17 June 2007 19:11:13 Alexandre Oliva wrote:

>> Let me start with an example: I bought a wireless router some time
>> ago, and it had a GNU+Linux distribution installed in it.  No source
>> code or written offer for source code, though.

> Just want to point out that, when I read this, my reaction was
> "But... That is a direct violation of the GPLv2. No specific reading
> of the license needed."

Yes.  Anyone feels like enforcing the GPLv2 in Brazil?  I can even
recommend lawyers that speak English reasonably well and are somewhat
familiar with the GPL, and I've already tracked the distribution chain
back to the initial infringer.  Harald is aware of the issue, but
AFAIK he's decided not to pursue that yet.

>> Now, if I called the vendor next day and asked for the source code,
>> and they responded "sorry, I can't give you that.  I threw it all
>> away, such that I wouldn't be able to give it to you.", they would
>> still be disrespecting my freedoms, as well as the license, right?

> Yes, they would. They are distributing a modification

There's no reason to assume it's a modification.  They're distributing
a copy, and that's enough.

>> So, if I called them to ask how to install and run modified versions
>> of the GPLed programs, and they responded "sorry, I can't give you
>> that.  I threw it all away, such that I wouldn't be able to give it to
>> you.", they would still be disrespecting my freedoms, as well as the
>> license.

> Not even the GPLv3dd4 - because they don't have the information
> anymore either. If, however, they still retained the information -
> in any form - they would be violating the GPLv3dd4.

I'm told by the authors of GPLv3dd4 that this case is not meant to be
permitted.  I suppose they're going to change the wording, or at least
the rationale for it.

> The GPLv2 doesn't make the actions described above - "how to install and 
> run" - a license violation.

This is true.  They didn't have any such duty, under the GPLv2.

However, if I figured that out by myself, but found that I was unable
to run a modified version because something in there checks for a hash
computed over the program I'd like to modify, and refuses to run it
because of the hash, then the hash is effectively part of the program,
and they haven't provided me with the corresponding sources of that
portion of the program.

I know you don't want that to be true, and a court might actually
decide your way some day.  But until then, your claim that this is
permitted by the GPL is just as good as mine that it's not.  And I
really mean "just as good", since my claim is in line with the stated
purpose of the authors of the GPL, and yours is in line with their
opinion (according to others, I don't think I've got this straight
from them) as to whether the license effectively prohibits this
practice.

> Then anyone using GPLv3'd software to drive WiFi devices, radio (HAM radio) 
> networks, etc - in the US, at least - isn't allowed to do such. US Law makes 
> some provisions of the GPLv3 illegal to comply with. Thanks to section 6 of 
> the GPLv3 that invalidates the rights granted under the license.

Actually, this is false.  Not only because of the ROM provisions in
the GPLv3, but because the law requirements aren't anywhere as strict
as the WiFi vendors who want to disrespect your freedoms want you to
believe.

> What the GPLv3 has done is take away options they might otherwise
> have had.

It doesn't.  Authors can always grant these options separately if they
want to.  Authors can always choose GPLv2 if they want to.

GPLv3 is an option for those who want to defend freedoms, even if they
don't share the perception that this is a moral and ethical issue.  If
they're in it only for the self benefits, that's fine, GPLv3 can get
them that, even better than GPLv2 could, in spite of the short-sighted
claims to the contrary.

> If one of the goals of the FSF is to force proprietary software into
> a minority then its just done damage to that goal.

That's not the goal.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: mea culpa on the meaning of Tivoization

[Daniel Hazelton]

On Monday 18 June 2007 15:09:47 Alexandre Oliva wrote:
> On Jun 17, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> > On Sunday 17 June 2007 19:11:13 Alexandre Oliva wrote:
> >> Let me start with an example: I bought a wireless router some time
> >> ago, and it had a GNU+Linux distribution installed in it.  No source
> >> code or written offer for source code, though.
> >
> > Just want to point out that, when I read this, my reaction was
> > "But... That is a direct violation of the GPLv2. No specific reading
> > of the license needed."
>
> Yes.  Anyone feels like enforcing the GPLv2 in Brazil?  I can even
> recommend lawyers that speak English reasonably well and are somewhat
> familiar with the GPL, and I've already tracked the distribution chain
> back to the initial infringer.  Harald is aware of the issue, but
> AFAIK he's decided not to pursue that yet.

I don't know if I have the right. None of the code is mine - the fact that 
they are in violation of the license is not in question (I trust your word on 
this), but it is the licensor who has the right to press charges. (I will 
check with the lawyers and law professionals I know, because the GPL makes no 
statements about the legal jurisdiction under which violations will be tried. 
It might be that I actually can file suit under Brazillian law)

> >> Now, if I called the vendor next day and asked for the source code,
> >> and they responded "sorry, I can't give you that.  I threw it all
> >> away, such that I wouldn't be able to give it to you.", they would
> >> still be disrespecting my freedoms, as well as the license, right?
> >
> > Yes, they would. They are distributing a modification
>
> There's no reason to assume it's a modification.  They're distributing
> a copy, and that's enough.

Bad word choice on my part. Of course you are correct.

> >> So, if I called them to ask how to install and run modified versions
> >> of the GPLed programs, and they responded "sorry, I can't give you
> >> that.  I threw it all away, such that I wouldn't be able to give it to
> >> you.", they would still be disrespecting my freedoms, as well as the
> >> license.
> >
> > Not even the GPLv3dd4 - because they don't have the information
> > anymore either. If, however, they still retained the information -
> > in any form - they would be violating the GPLv3dd4.
>
> I'm told by the authors of GPLv3dd4 that this case is not meant to be
> permitted.  I suppose they're going to change the wording, or at least
> the rationale for it.

Okay. So its possible to change whats running on the hardware - but even 
though nobody has the information needed to do it, it's a violation. Hrm... I 
can see some valid reasoning behind this, but it'll take creative legalese to 
make sure that things like (EE)PROMS are properly covered.

> > The GPLv2 doesn't make the actions described above - "how to install and
> > run" - a license violation.
>
> This is true.  They didn't have any such duty, under the GPLv2.
>
> However, if I figured that out by myself, but found that I was unable
> to run a modified version because something in there checks for a hash
> computed over the program I'd like to modify, and refuses to run it
> because of the hash, then the hash is effectively part of the program,
> and they haven't provided me with the corresponding sources of that
> portion of the program.

"Effectively" - yes, that is the perfect way to describe it. And even though 
it isn't directly part, a situation like that should be covered. (In other 
words, if this was the way the "tivoization" section was written to make 
this "effectively part of the work" bit the focus a lot of my objections to 
it would be nullified. Give me a few hours to work on some solid and 
unambiguous language and I'll send something your way for review) 

> I know you don't want that to be true, and a court might actually
> decide your way some day.  But until then, your claim that this is
> permitted by the GPL is just as good as mine that it's not.  And I
> really mean "just as good", since my claim is in line with the stated
> purpose of the authors of the GPL, and yours is in line with their
> opinion (according to others, I don't think I've got this straight
> from them) as to whether the license effectively prohibits this
> practice.

Agreed. As Linus pointed out, we've been arguing over opinions and that's 
pointless. The only thing to do when someone states an opinion is to nod and 
accept it.

> > Then anyone using GPLv3'd software to drive WiFi devices, radio (HAM
> > radio) networks, etc - in the US, at least - isn't allowed to do such. US
> > Law makes some provisions of the GPLv3 illegal to comply with. Thanks to
> > section 6 of the GPLv3 that invalidates the rights granted under the
> > license.
>
> Actually, this is false.  Not only because of the ROM provisions in
> the GPLv3, but because the law requirements aren't anywhere as strict
> as the WiFi vendors who want to disrespect your freedoms want you to
> believe.

Perhaps. I haven't looked into the specific regulations in over a year, so my 
memory may be failing me entirely.

> > What the GPLv3 has done is take away options they might otherwise
> > have had.
>
> It doesn't.  Authors can always grant these options separately if they
> want to.  Authors can always choose GPLv2 if they want to.

Okay. I think that someone pointed out a problem with the "optional grant" 
idea, but I can't remember the specifics and don't feel like digging through 
the 500 or so posts that make up this discussion.

> GPLv3 is an option for those who want to defend freedoms, even if they
> don't share the perception that this is a moral and ethical issue.  If
> they're in it only for the self benefits, that's fine, GPLv3 can get
> them that, even better than GPLv2 could, in spite of the short-sighted
> claims to the contrary.

If this is your opinion, then run with it. My opinion on the matter is the 
opposite - that the GPLv2 does the job in a better way - but, well, that's my 
opinion. (and like my mother used to say - "Opinions are like assholes. 
Everyone has one, and the only one that doesn't stink is your own.")

> > If one of the goals of the FSF is to force proprietary software into
> > a minority then its just done damage to that goal.
>
> That's not the goal.

I didn't say it was "the goal", I said "one of the goals". I'm the first to 
admit when I'm wrong, but in this case I've read interviews with RMS where he 
has said that one of the reasons he founded the FSF was to marginalize 
proprietary software. (No, I don't know where this was - the interview was 
done several years ago)

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: mea culpa on the meaning of Tivoization

[Alexandre Oliva]

On Jun 18, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> On Monday 18 June 2007 15:09:47 Alexandre Oliva wrote:

>> Yes.  Anyone feels like enforcing the GPLv2 in Brazil?

> I don't know if I have the right. None of the code is mine

It would have to be some major copyright holder of core Linux code, or
the mips port (IIRC it's mips hardware), or some other driver they're
using.

> Okay. So its possible to change whats running on the hardware - but
> even though nobody has the information needed to do it, it's a
> violation.  Hrm... I can see some valid reasoning behind this,

Yup.  Same reasoning as "I threw the source code away", really.

> "Effectively" - yes, that is the perfect way to describe it. And
> even though it isn't directly part, a situation like that should be
> covered.

And if you look at GPLv3dd1 or dd2 IIRC, that's how it started.  For
some reason, the FSF turned it into the more lax (in some senses)
installation information for user products in dd3.  Maybe they decided
that the argument about the signature being effectively part of the
executable, and therefore the key being effectively part of the source
code, was less likely to be upheld in a court of law than this
alternate phrasing.  All in all, the effect is the same AFAICT, and
the spirit is being complied with.

>> > What the GPLv3 has done is take away options they might otherwise
>> > have had.

>> It doesn't.  Authors can always grant these options separately if they
>> want to.  Authors can always choose GPLv2 if they want to.

> Okay. I think that someone pointed out a problem with the "optional grant" 
> idea, but I can't remember the specifics and don't feel like digging through 
> the 500 or so posts that make up this discussion.

Linus claimed he would then have to refrain from accepting
contributions from anyone who removed this additional permission.

I don't see how this is different from refraining from accepting
contributions under any other license, except that you can't use
license incompatibility to reason it out as an impossibility you
established for yourself in just the very same way.

>> > If one of the goals of the FSF is to force proprietary software into
>> > a minority then its just done damage to that goal.

>> That's not the goal.

> I didn't say it was "the goal", I said "one of the goals".

I stand corrected.  Sorry.  It's been a long thread and a long week.

My objection was mainly about the "forcing".  FSF's stance is about
educating users as to the moral and ethical reasons, such that they
reject non-Free Software, while at the same time providing software
authors with means to stop others from hurting users, by depriving
them of the freedoms they're morally entitled to have.

Others often perceive FSF's tactics as forceful, and I don't deny that
this may be justified, based on past interactions with the FSF.  That
said, I think they've improved a lot, even if they're not perfect (who
is?).  But the perception and the consequent rejection unfortunately
remains as strong as ever.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: mea culpa on the meaning of Tivoization

[Daniel Hazelton]

On Monday 18 June 2007 17:31:47 Alexandre Oliva wrote:
> On Jun 18, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> > On Monday 18 June 2007 15:09:47 Alexandre Oliva wrote:
> >> Yes.  Anyone feels like enforcing the GPLv2 in Brazil?
> >
> > I don't know if I have the right. None of the code is mine
>
> It would have to be some major copyright holder of core Linux code, or
> the mips port (IIRC it's mips hardware), or some other driver they're
> using.

That's what I thought. Anyone here feel like pursuing that? It is, directly, a 
license violation. :)

> > Okay. So its possible to change whats running on the hardware - but
> > even though nobody has the information needed to do it, it's a
> > violation.  Hrm... I can see some valid reasoning behind this,
>
> Yup.  Same reasoning as "I threw the source code away", really.

Not really. "I threw the source away" violates the license in a very clear 
manner - they are distributing a "work based on the work" without complying 
with the terms of the license under which they gained access to the work. 
However, as I said, I can see some valid reasoning. But its more a part of 
the "Hey, I paid good money for this thing and I can't use it how I want!" 
type of reasoning.

> > "Effectively" - yes, that is the perfect way to describe it. And
> > even though it isn't directly part, a situation like that should be
> > covered.
>
> And if you look at GPLv3dd1 or dd2 IIRC, that's how it started.  For
> some reason, the FSF turned it into the more lax (in some senses)
> installation information for user products in dd3.  Maybe they decided
> that the argument about the signature being effectively part of the
> executable, and therefore the key being effectively part of the source
> code, was less likely to be upheld in a court of law than this
> alternate phrasing.  All in all, the effect is the same AFAICT, and
> the spirit is being complied with.

But the change has some massive problems. If dd1 or dd2 was clearly and 
concisely written such that the conditions were not open to a different 
interpretation without creative re-definition of words then changes would not 
be needed. (I'm still working on the version I mentioned - give me a bit, 
writing english in such a way that a lawyer can't twist it to mean whatever 
they are paid to make it mean is difficult.)

> >> > What the GPLv3 has done is take away options they might otherwise
> >> > have had.
> >>
> >> It doesn't.  Authors can always grant these options separately if they
> >> want to.  Authors can always choose GPLv2 if they want to.
> >
> > Okay. I think that someone pointed out a problem with the "optional
> > grant" idea, but I can't remember the specifics and don't feel like
> > digging through the 500 or so posts that make up this discussion.
>
> Linus claimed he would then have to refrain from accepting
> contributions from anyone who removed this additional permission.
>
> I don't see how this is different from refraining from accepting
> contributions under any other license, except that you can't use
> license incompatibility to reason it out as an impossibility you
> established for yourself in just the very same way.

I think there was more to it than that, but the point doesn't matter. If the 
license used on contributed code *isn't* completely compatible with the 
license on the project it can't be used anyway. (doesn't the GPLv3 cover 
situations like that?)

> >> > If one of the goals of the FSF is to force proprietary software into
> >> > a minority then its just done damage to that goal.
> >>
> >> That's not the goal.
> >
> > I didn't say it was "the goal", I said "one of the goals".
>
> I stand corrected.  Sorry.  It's been a long thread and a long week.

Understandable.

> My objection was mainly about the "forcing".  FSF's stance is about
> educating users as to the moral and ethical reasons, such that they
> reject non-Free Software, while at the same time providing software
> authors with means to stop others from hurting users, by depriving
> them of the freedoms they're morally entitled to have.

Hrm... When I first hit the end of this massive sentence I was really 
confused. Took about five minutes for me to remember that "morally entitled" 
is based on the morals promoted by the FSF.

> Others often perceive FSF's tactics as forceful, and I don't deny that
> this may be justified, based on past interactions with the FSF.  That
> said, I think they've improved a lot, even if they're not perfect (who
> is?).  But the perception and the consequent rejection unfortunately
> remains as strong as ever.

IMHO RMS and the FSF go about things in the wrong manner. Yes, the goal is 
admirable, but the means by which they are attempting to reach them are not. 
However, I have no connection to the FSF or RMS and thus cannot dictate 
anything. What I can do is try to educate people as to why I have the dislike 
of the FSF and RMS that I do. (and how I can *respect* RMS and the FSF while 
not liking them :)

In this discussion I've tried to point out things that occur that are at odds 
with the text of the GPLv3 (as it exists - the "31/5/2007 'Final Comment 
Draft'). I have also tried to point out that what the FSF and RMS claim is 
the "spirit" of the GPL cannot be conclusively proven to be the same "spirit" 
it had when it was written. When it comes to that somewhat slippery 
term "intent" I think I made my point very clear - ie: the intent of the 
license is what each licensor believes it to be. (meaning my view of the 
intent is no less valid than the FSF's view)

Everyone that has been part of this discussion - my personal code of morals 
will not let me get away without this: "Forgive me if, in the heat of the 
moment, I offended any of you." (it shouldn't need to be said that I've 
forgiven the few people that managed to offend me)

Now - this discussion has been a lot of fun and really kept me on my toes. It 
has been years since I've had a "debate" as lively as this one has been. So 
thank you all for it. Now please, remove me from the CC: list as, IMNSHO, 
this discussion has run its course and come to all the resolution it ever 
will.

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: mea culpa on the meaning of Tivoization

[Alexandre Oliva]

On Jun 18, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> On Monday 18 June 2007 17:31:47 Alexandre Oliva wrote:
>> And if you look at GPLv3dd1 or dd2 IIRC, that's how it started.  For
>> some reason, the FSF turned it into the more lax (in some senses)
>> installation information for user products in dd3.  Maybe they decided
>> that the argument about the signature being effectively part of the
>> executable, and therefore the key being effectively part of the source
>> code, was less likely to be upheld in a court of law than this
>> alternate phrasing.  All in all, the effect is the same AFAICT, and
>> the spirit is being complied with.

> But the change has some massive problems.

Such as?  Is the effect really any different?

> If dd1 or dd2 was clearly and concisely written such that the
> conditions were not open to a different interpretation without
> creative re-definition of words then changes would not be
> needed. (I'm still working on the version I mentioned - give me a
> bit, writing english in such a way that a lawyer can't twist it to
> mean whatever they are paid to make it mean is difficult.)

It's very difficult and, worse, it might turn out to be unenforceable.
You'd have to count on signing keys being copyrightable, and they are
unlikely to be, and on signatures being derived works of both, which
is a tough call.  The whole idea resonates very well with the spirit
of the license, but we need more than that, we need it to be very
likely to work.  I suspect this is why the FSF has decided to take
another route to achieve the same (AFAICT) effect.

>> I don't see how this is different from refraining from accepting
>> contributions under any other license, except that you can't use
>> license incompatibility to reason it out as an impossibility you
>> established for yourself in just the very same way.

> I think there was more to it than that, but the point doesn't
> matter. If the license used on contributed code *isn't* completely
> compatible with the license on the project it can't be used
> anyway. (doesn't the GPLv3 cover situations like that?)

I'm not sure what you're asking.  GPLv3 covers additional permissions,
that are really no different from dual-licensing, so anyone can choose
to drop them when combining with works (including their own) that
don't offer such additional permissions.

>> My objection was mainly about the "forcing".  FSF's stance is about
>> educating users as to the moral and ethical reasons, such that they
>> reject non-Free Software, while at the same time providing software
>> authors with means to stop others from hurting users, by depriving
>> them of the freedoms they're morally entitled to have.

> Hrm... When I first hit the end of this massive sentence I was really 
> confused. Took about five minutes for me to remember that "morally entitled" 
> is based on the morals promoted by the FSF.

Yes.  And the 'them' after the last comma refers to the users, not the
authors (although they can be users too), in case it's not clear ;-)
:-D

> Everyone that has been part of this discussion - my personal code of morals 
> will not let me get away without this: "Forgive me if, in the heat of the 
> moment, I offended any of you."

FWIW, I never felt offended by you, but I second your request and
extend it to all participants in the thread too, particularly to Ingo,
to whom I remember having directed some harsh words.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: mea culpa on the meaning of Tivoization

[Daniel Hazelton]

On Monday 18 June 2007 22:57:20 Alexandre Oliva wrote:
> On Jun 18, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> > On Monday 18 June 2007 17:31:47 Alexandre Oliva wrote:
> >> And if you look at GPLv3dd1 or dd2 IIRC, that's how it started.  For
> >> some reason, the FSF turned it into the more lax (in some senses)
> >> installation information for user products in dd3.  Maybe they decided
> >> that the argument about the signature being effectively part of the
> >> executable, and therefore the key being effectively part of the source
> >> code, was less likely to be upheld in a court of law than this
> >> alternate phrasing.  All in all, the effect is the same AFAICT, and
> >> the spirit is being complied with.
> >
> > But the change has some massive problems.
>
> Such as?  Is the effect really any different?

I haven't looked at it, in depth, today but one of the problems I saw was the 
apparent loopholes in the text. No specifics, but I remember thinking "a 
lawyer would have a field day with this - dozens of ways they could sidestep 
these issues"

> > If dd1 or dd2 was clearly and concisely written such that the
> > conditions were not open to a different interpretation without
> > creative re-definition of words then changes would not be
> > needed. (I'm still working on the version I mentioned - give me a
> > bit, writing english in such a way that a lawyer can't twist it to
> > mean whatever they are paid to make it mean is difficult.)
>
> It's very difficult and, worse, it might turn out to be unenforceable.
> You'd have to count on signing keys being copyrightable, and they are
> unlikely to be, and on signatures being derived works of both, which
> is a tough call.  The whole idea resonates very well with the spirit
> of the license, but we need more than that, we need it to be very
> likely to work.  I suspect this is why the FSF has decided to take
> another route to achieve the same (AFAICT) effect.

Agreed. I'm still stuck trying to keep the language concise and understandable 
without delving into the descriptive flights of fancy I enjoy. (I write a lot 
more fiction than I do code, even though I started writing code a long time 
before I started on fiction)

> >> I don't see how this is different from refraining from accepting
> >> contributions under any other license, except that you can't use
> >> license incompatibility to reason it out as an impossibility you
> >> established for yourself in just the very same way.
> >
> > I think there was more to it than that, but the point doesn't
> > matter. If the license used on contributed code *isn't* completely
> > compatible with the license on the project it can't be used
> > anyway. (doesn't the GPLv3 cover situations like that?)
>
> I'm not sure what you're asking.  GPLv3 covers additional permissions,
> that are really no different from dual-licensing, so anyone can choose
> to drop them when combining with works (including their own) that
> don't offer such additional permissions.

What I was getting at, here, is that the GPLv3 isn't backwards compatible with 
GPLv2, because you aren't allowed to remove rights from the GPLv3. Remember, 
there are rights encoded in the GPLv3 that don't appear in v2. In fact, if 
you want to use GPLv3 code in a GPLv2 project you have to use GPLv3. For some 
projects, like the Linux Kernel, the upgrade is impossible to accomplish.

> >> My objection was mainly about the "forcing".  FSF's stance is about
> >> educating users as to the moral and ethical reasons, such that they
> >> reject non-Free Software, while at the same time providing software
> >> authors with means to stop others from hurting users, by depriving
> >> them of the freedoms they're morally entitled to have.
> >
> > Hrm... When I first hit the end of this massive sentence I was really
> > confused. Took about five minutes for me to remember that "morally
> > entitled" is based on the morals promoted by the FSF.
>
> Yes.  And the 'them' after the last comma refers to the users, not the
> authors (although they can be users too), in case it's not clear ;-)
>
> :-D

Yes. I almost replied "-ENOPARSE" because, when I first read it, I parsed it 
as "by depriving [the authors] of the freedoms they're morally entitled to 
have". When my brain finally rebooted after that bought of idiocy I was able 
to parse it properly.

DRH

> > Everyone that has been part of this discussion - my personal code of
> > morals will not let me get away without this: "Forgive me if, in the heat
> > of the moment, I offended any of you."
>
> FWIW, I never felt offended by you, but I second your request and
> extend it to all participants in the thread too, particularly to Ingo,
> to whom I remember having directed some harsh words.



-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: mea culpa on the meaning of Tivoization

[Alexandre Oliva]

On Jun 19, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> I haven't looked at it, in depth, today but one of the problems I
> saw was the apparent loopholes in the text. No specifics, but I
> remember thinking "a lawyer would have a field day with this -
> dozens of ways they could sidestep these issues"

*Pretty* *please* file comments about the apparent loopholes at
gplv3.fsf.org/comments

> What I was getting at, here, is that the GPLv3 isn't backwards
> compatible with GPLv2,

It couldn't possibly be.  The whole point of upgrading the GPL is such
that it complies better with its spirit of defending the freedoms, so
as to keep free software free.  This can only be accomplished with
additional restrictions that stop practices that deny users'
freedoms.

Relaxing the provisions, a necessary condition for compatibility,
wouldn't make for better defenses.

> because you aren't allowed to remove rights from the GPLv3. Remember, 
> there are rights encoded in the GPLv3 that don't appear in v2.

I'm not sure what you mean by "rights" in the two sentences above.
You know you can grant additional permissions, so I assume that's not
what you mean, even more so because you *can* indeed take them out.
Is it "conditions", "restrictions" or some such, that in turn
translate into freedoms for downstream users, or is it about the
granted rights per se?

> In fact, if you want to use GPLv3 code in a GPLv2 project you have
> to use GPLv3. For some projects, like the Linux Kernel, the upgrade
> is impossible to accomplish.

Impossible is a bit too strong.  I understand it would take a huge
amount of work though, so I sympathize with "it wouldn't be worth it",
even if, in my scale of moral values, I'd disagree.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: mea culpa on the meaning of Tivoization

[Daniel Hazelton]

On Tuesday 19 June 2007 02:10:02 Alexandre Oliva wrote:
> On Jun 19, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> > I haven't looked at it, in depth, today but one of the problems I
> > saw was the apparent loopholes in the text. No specifics, but I
> > remember thinking "a lawyer would have a field day with this -
> > dozens of ways they could sidestep these issues"
>
> *Pretty* *please* file comments about the apparent loopholes at
> gplv3.fsf.org/comments

To do that I'd have to go back, take the time to re-read the GPLv3 *in* 
*depth*, think about each paragraph of each section individually...

Like I said, I just got a general impression that a lawyer would have a 
field-day with it.

> > What I was getting at, here, is that the GPLv3 isn't backwards
> > compatible with GPLv2,
>
> It couldn't possibly be.  The whole point of upgrading the GPL is such
> that it complies better with its spirit of defending the freedoms, so
> as to keep free software free.  This can only be accomplished with
> additional restrictions that stop practices that deny users'
> freedoms.
>
> Relaxing the provisions, a necessary condition for compatibility,
> wouldn't make for better defenses.
>
> > because you aren't allowed to remove rights from the GPLv3. Remember,
> > there are rights encoded in the GPLv3 that don't appear in v2.
>
> I'm not sure what you mean by "rights" in the two sentences above.
> You know you can grant additional permissions, so I assume that's not
> what you mean, even more so because you *can* indeed take them out.
> Is it "conditions", "restrictions" or some such, that in turn
> translate into freedoms for downstream users, or is it about the
> granted rights per se?

Sorry, bad choice of words. There are "guarantees" encoded into every license. 
There are some encoded into the GPLv3 that aren't encoded into the GPLv2. You 
can't remove or restrict those guarantees without violating the license. And 
removing those guarantees would be the only way to make the GPLv3 fully 
compatible with the GPLv2.

> > In fact, if you want to use GPLv3 code in a GPLv2 project you have
> > to use GPLv3. For some projects, like the Linux Kernel, the upgrade
> > is impossible to accomplish.
>
> Impossible is a bit too strong.  I understand it would take a huge
> amount of work though, so I sympathize with "it wouldn't be worth it",
> even if, in my scale of moral values, I'd disagree.

In this case I wasn't speaking literally. I should have been a lot more 
specific there - it should say "practically impossible". 

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

RE: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Schwartz]

> On Jun 17, 2007, Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:
>
> >> I don't know any law that requires tivoization.
>
> > In the USSA it is arguable that wireless might need it (if done in
> > software) for certain properties. (The argument being it must be
> > tamperproof to random end consumers).
>
> But this is not tivoization.

> Tivoization is a manufacturer using technical measures to prevent the
> user from tampering (*) with the device, *while* keeping the ability
> to tamper with it changes itself.

You're splitting those hairs might finely. So when you ask whether there's
any law that "requirse tivoization", you won't accept a law that creates a
situation where the only practical solution is tivoization?

> (*) tampering brings in negative connotations that I'd rather avoid,
> but since that was the term you used, and the term "modifying" might
> bring in legal-based technicalities such as that replacing isn't
> modification, I just went with it.

> So, given a proper definition, do you know any law that requires
> tivoization?

> Taking it further, do you know whether any such law requires
> *worldwide* tivoization, as in, applying the restrictions in the law
> even outside its own jurisdiction?

A law that requires certaint things be tamper-proof, where engineering
realities requires that they be controlled by software and the software be
upgradable (for security reasons and for support of future protocol
revisions) isn't good enough for you?

DS

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 18, 2007, "David Schwartz" <davids@webmaster.com> wrote:

>> On Jun 17, 2007, Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:
>> 
>> >> I don't know any law that requires tivoization.
>> 
>> > In the USSA it is arguable that wireless might need it (if done in
>> > software) for certain properties. (The argument being it must be
>> > tamperproof to random end consumers).
>> 
>> But this is not tivoization.

>> Tivoization is a manufacturer using technical measures to prevent the
>> user from tampering (*) with the device, *while* keeping the ability
>> to tamper with it changes itself.

> You're splitting those hairs might finely.

And I was wrong.  Please see the "mea culpa on the meaning of
tivoization" thread.

> So when you ask whether there's any law that "requirse tivoization",
> you won't accept a law that creates a situation where the only
> practical solution is tivoization?

I guess it amounts to what you mean by "*only* practical solution".

"I can't fit the corresponding sources in this CD, so you won't get
them." is no excuse to disrespect users' freedoms, why should this be
different?

>> Taking it further, do you know whether any such law requires
>> *worldwide* tivoization, as in, applying the restrictions in the law
>> even outside its own jurisdiction?

> A law that requires certaint things be tamper-proof, where engineering
> realities requires that they be controlled by software and the software be
> upgradable (for security reasons and for support of future protocol
> revisions) isn't good enough for you?

"engineering realities" is the weak point of your argument, see above.
Is ROM still software?  Is replaceable ROM still software?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Drake]

Let's take a certain class of medical devices into account: ones that 
are absolutely definitely for medical treatment, but are not life 
threatening if they fail.

Say, a dental treatment device -- if the device produces a crown or 
bridge that doesn't fit properly, the dentist says "nope" and throws it 
away. No harm done.

Alexandre Oliva wrote:
> There may be business models that require the ability to make changes.

I'd say that its sensible for the manufacturer to attempt to retain this 
ability in every case. You never know what's going to go wrong, so it's 
a plus to have this option so that you can roll out some types of fixes 
without going bankrupt.

Now, for medical devices, this is tricky stuff: medical devices require 
all sorts of certifications, so modifying your product after you have 
certified it has it's complications. However, despite all the 
regulations it's realistic to be able to do this, and it does happen. 
Hell, windows-based devices in this field download new antivirus 
definitions and run windows update every few days.

> Then it's fair to enable the user to make changes as well, such that
> they don't become dependent on the vendor

Now this is where the regulations get really heavy. If the user is 
offered the ability to modify the device, theres *no way* it would get 
certified. Your business is dead - you do not have a product you can 
sell. In such case, the license has completely excluded free software 
from the market and everyone is forced to use completely closed systems.


I realise that the latest GPLv3 draft would not pose restrictions here, 
as such devices would not be classified as consumer products. That said, 
talking purely in terms of business models and fairness: there ARE 
decent reasons for manufacturer lockdown in some industries.

Daniel

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Tim Post]

On Mon, 2007-06-18 at 23:21 -0400, Daniel Drake wrote:
> Let's take a certain class of medical devices into account: ones that 
> are absolutely definitely for medical treatment, but are not life 
> threatening if they fail.
> 
> Say, a dental treatment device -- if the device produces a crown or 
> bridge that doesn't fit properly, the dentist says "nope" and throws it 
> away. No harm done.

I've done quite a bit of research, I'm not nearly done.

These regulations (from what I can tell) seemed to follow suit with the
National Electric Code (NEC) [latest] when dealing with mandatory
isolated ground devices and special cabling methods when it comes into a
device touching a patient. If that remains consistent, this won't be so
bad.

If the patient never comes in contact with it, its not regulated as much
and (from what I've seen) has no requirement for tamper proofing. I
point out again, I am not _nearly_ done with my research.

I think of nothing else, anyone with an interest should closely monitor
how these devices are being regulated by the FDA as more of them begin
to look like penguins.

I won't argue one way or another as to the presence of benevolent intent
in those laws-to-come, I'm simply pointing out the questionable
technical competency of those who will be writing them and their need
for guidance when doing so.


Best,
--Tim

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 19, 2007, Daniel Drake <dsd@gentoo.org> wrote:

> I realise that the latest GPLv3 draft would not pose restrictions
> here, as such devices would not be classified as consumer
> products.

And even if they were, there's always ROM.

I don't know whether hardware seals that state "once you break this
seal, law prohibits the use of this device with human patients".

Then the restriction is not being imposed by the manufacturer, only by
law, and this does make lot of a difference as far as software freedom
is concerned.

But then, law might not find this to be enough.  Software patents are
not the only stupid law that harms Free Software :-(

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[david]

On Tue, 19 Jun 2007, Alexandre Oliva wrote:

> On Jun 19, 2007, Daniel Drake <dsd@gentoo.org> wrote:
>
>> I realise that the latest GPLv3 draft would not pose restrictions
>> here, as such devices would not be classified as consumer
>> products.
>
> And even if they were, there's always ROM.
>
> I don't know whether hardware seals that state "once you break this
> seal, law prohibits the use of this device with human patients".

once you break the seal the device is no longer certified. an uncertified 
device cannot be used.

this is very common (in some areas it's widely ignored, in others it 
isn't)

this is just like the 'you void the warranty if you disrupt this sticker' 
stickers that you see on just about any hardware you buy today. some 
vendors are stickers for this, others don't really care.

David Lang

> Then the restriction is not being imposed by the manufacturer, only by
> law, and this does make lot of a difference as far as software freedom
> is concerned.
>
> But then, law might not find this to be enough.  Software patents are
> not the only stupid law that harms Free Software :-(
>
>

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 19, 2007, david@lang.hm wrote:

> On Tue, 19 Jun 2007, Alexandre Oliva wrote:
>> On Jun 19, 2007, Daniel Drake <dsd@gentoo.org> wrote:
>> 
>>> I realise that the latest GPLv3 draft would not pose restrictions
>>> here, as such devices would not be classified as consumer
>>> products.
>> 
>> And even if they were, there's always ROM.
>> 
>> I don't know whether hardware seals that state "once you break this
>> seal, law prohibits the use of this device with human patients".

> once you break the seal the device is no longer certified. an
> uncertified device cannot be used.

Yup.  That's the law.

At which point it's not the hardware vendor imposing the restriction,
so this use is perfectly acceptable.


One could presumably implement similar seals in software.  Nothing
wrong with a signature used to indicate that the device has been
tampered with.  Even a led somewhere that reflects this status.

None of this prevents the user from enjoying the freedoms he's
entitled to, according to the laws of the place where he lives.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Lennart Sorensen]

On Fri, Jun 15, 2007 at 04:26:34PM -0300, Alexandre Oliva wrote:
> If the bug is in the non-GPLed BIOS, not in the GPLed code, too bad.
> One more reason to dislike non-Free Software.

Maybe the Tivo only loading signed kernels is a bug in their bios. :)

> The freedom the GPL defends is not the freedom to modify and debug the
> system, but rather the covered software.
> 
> Now, if you find evidence that the "bug" is actually intentionally put
> there to stop you from doing what you wanted with the software, then
> there's clearly a violation of the spirit of the license, and you
> might even have a case of copyright infringement, but IANAL.

There are many interesting bugs out there.  Who is to say what was
intensional?

--
Len Sorensen

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 20, 2007, lsorense@csclub.uwaterloo.ca (Lennart Sorensen) wrote:

> On Fri, Jun 15, 2007 at 04:26:34PM -0300, Alexandre Oliva wrote:
>> If the bug is in the non-GPLed BIOS, not in the GPLed code, too bad.
>> One more reason to dislike non-Free Software.

> Maybe the Tivo only loading signed kernels is a bug in their bios. :)

That might be so.  And in the US, a court might end up finding that
piece of code was taken at random, for no particular reason, from a
sample of garbage produced by code monkeys (in a very literal sense)
typing at random on computer keyboards over a large period of time ;-)

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Rob Landley]

On Wednesday 13 June 2007 21:04:42 Alexandre Oliva wrote:
> On Jun 13, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> > Now stop parroting the FSF's worn and tired tripe.
>
> Are you playing Linus' sheeple and parroting his lines just to make a
> point, or are you like that all the time? ;-)

Read the hover text on http://xkcd.org/c202.html

I'm wondering if it's time for a "Munroe's law" concerning that word...

Rob

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alan Cox]

> > Only courts of law can do that.
> 
> Wrong! Anyone with half a brain can make the distinction. What TiVO did is 

Maybe half a brain can, but anyone with a whole brain can assure you its
a bit more complex and you are wrong..

> version of it that we provide on our hardware". Why is that legal? Because 
> TiVO produces the hardware and sells it to you with a certain *LICENSE* - 

The keys required to make the code run with the hardware are part of the
software. The license requires the software and relevant scripts etc are
included. Thus there is a very good argument that the keys are part of
the software.

And since there is no court ruling to high enough level in the USA, UK or
any other jurisdiction on that it remains a matter of opinion.

Tivo may control the hardware but the authors control the software (via
the GPL), and subject to the limits of what may be specified by a
copyright license (as opposed to contract) can make such demands as they
see fit about their software and anything derivative of it.

> because it does contain hardware covered under any number of patents. That 
> license grants you the right to use the patents - in this case algorithms - 

You can't patent algorithms either

Alan

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Wednesday 13 June 2007 21:16:19 Alan Cox wrote:
> > > Only courts of law can do that.
> >
> > Wrong! Anyone with half a brain can make the distinction. What TiVO did
> > is
>
> Maybe half a brain can, but anyone with a whole brain can assure you its
> a bit more complex and you are wrong..
>
> > version of it that we provide on our hardware". Why is that legal?
> > Because TiVO produces the hardware and sells it to you with a certain
> > *LICENSE* -
>
> The keys required to make the code run with the hardware are part of the
> software. The license requires the software and relevant scripts etc are
> included. Thus there is a very good argument that the keys are part of
> the software.

Good argument, but I'll stand by my interpretation of the law, the GPL and the 
situation until there is solid proof that a signing-key is part of the source 
code. Doubly so because the language of the GPLv2 makes it clear that "all 
relevant scripts, etc" are only needed to build and run the "covered work" - 
not for proper installation of it. (and, in the case of a TiVO, the signing 
keys are part of the installation, not the running or building. Besides 
needing the proper signing key, the kernel in a TiVO is run the same as any 
other Linux kernel) 

> And since there is no court ruling to high enough level in the USA, UK or
> any other jurisdiction on that it remains a matter of opinion.
>
> Tivo may control the hardware but the authors control the software (via
> the GPL), and subject to the limits of what may be specified by a
> copyright license (as opposed to contract) can make such demands as they
> see fit about their software and anything derivative of it.

Agreed. However, AFAICT, TiVO meets the provisions of the GPLv2 - they make 
the source of the GPL'd part of their system available. (And I'm not going to 
get into arguments over whether kernel modules are "derivative works" or not, 
since those invariably end up with "They aren't, even though we think they 
should be")

> > because it does contain hardware covered under any number of patents.
> > That license grants you the right to use the patents - in this case
> > algorithms -
>
> You can't patent algorithms either

Then explain the patents on the MP3 algorithm, the LZW algorithm, etc... Those 
patents are real and while the LZW one may have lapsed, still relevant.

DRH 

> Alan



-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 13, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:

> (and, in the case of a TiVO, the signing 
> keys are part of the installation, not the running or building.

Is installation not a precondition for running?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Valdis.Kletnieks]

[-- Attachment #1: Type: text/plain, Size: 533 bytes --]

On Wed, 13 Jun 2007 23:40:47 -0300, Alexandre Oliva said:
> On Jun 13, 2007, Daniel Hazelton <dhazelton@enter.net> wrote:
> 
> > (and, in the case of a TiVO, the signing 
> > keys are part of the installation, not the running or building.
> 
> Is installation not a precondition for running?

If a company sells you hardware that includes a ROM that contains GPL'ed
software, are they in violation of the GPL if they don't include a ROM burner
in the hardware?  Or are ROM burners like compilers, where you have to supply
your own?


[-- Attachment #2: Type: application/pgp-signature, Size: 226 bytes --]

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Valdis.Kletnieks@vt.edu wrote:

> If a company sells you hardware that includes a ROM that contains GPL'ed
> software, are they in violation of the GPL if they don't include a ROM burner
> in the hardware?  Or are ROM burners like compilers, where you have to supply
> your own?

  this requirement does not apply if neither you nor any third party
  retains the ability to install modified object code on the User
  Product (for example, the work has been installed in ROM).

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Valdis.Kletnieks]

[-- Attachment #1: Type: text/plain, Size: 665 bytes --]

On Thu, 14 Jun 2007 03:03:40 -0300, Alexandre Oliva said:
> On Jun 14, 2007, Valdis.Kletnieks@vt.edu wrote:
> 
> > If a company sells you hardware that includes a ROM that contains GPL'ed
> > software, are they in violation of the GPL if they don't include a ROM burner
> > in the hardware?  Or are ROM burners like compilers, where you have to supply
> > your own?
> 
>   this requirement does not apply if neither you nor any third party
>   retains the ability to install modified object code on the User
>   Product (for example, the work has been installed in ROM).

Do they have to provide a ROM burner if the ROM is socketed rather than
soldered into place?

[-- Attachment #2: Type: application/pgp-signature, Size: 226 bytes --]

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 14, 2007, Valdis.Kletnieks@vt.edu wrote:

> On Thu, 14 Jun 2007 03:03:40 -0300, Alexandre Oliva said:
>> On Jun 14, 2007, Valdis.Kletnieks@vt.edu wrote:
>> 
>> > If a company sells you hardware that includes a ROM that contains GPL'ed
>> > software, are they in violation of the GPL if they don't include a ROM burner
>> > in the hardware?  Or are ROM burners like compilers, where you have to supply
>> > your own?
>> 
>> this requirement does not apply if neither you nor any third party
>> retains the ability to install modified object code on the User
>> Product (for example, the work has been installed in ROM).

> Do they have to provide a ROM burner if the ROM is socketed rather than
> soldered into place?

Of course not.  They just can't impose restrictions on your obtaining
a ROM burner and doing the work yourself.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Ingo Molnar]

* Alexandre Oliva <aoliva@redhat.com> wrote:

> > Do they have to provide a ROM burner if the ROM is socketed rather 
> > than soldered into place?
> 
> Of course not.  They just can't impose restrictions on your obtaining 
> a ROM burner and doing the work yourself.

do you realize that you have just admitted that the Tivo is perfectly 
fine and legal?

because you can solder off the ROM from the Tivo and can put in a new 
ROM with another bootloader that does not check the SHA1 key. Tivo puts 
no restrictions on you to obtain a ROM and a ROM burner and do this work 
yourself. (they dont help you either, but you just conceded in another 
thread that the hardware maker does not have to go out on his way to 
help you in your software modification efforts.)

	Ingo

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alan Cox]

> because you can solder off the ROM from the Tivo and can put in a new 
> ROM with another bootloader that does not check the SHA1 key. Tivo puts 

Then you've committed an offence because of the SHA1 key removal. Tivo
deliberately create a system where removal of the ROM is an offence
(sometimes in criminal law) so that argument doesn't hold water.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Woodhouse]

On Wed, 2007-06-13 at 21:29 -0400, Daniel Hazelton wrote:
> Agreed. However, AFAICT, TiVO meets the provisions of the GPLv2 - they make 
> the source of the GPL'd part of their system available. (And I'm not going to 
> get into arguments over whether kernel modules are "derivative works" or not, 
> since those invariably end up with "They aren't, even though we think they 
> should be")

Who cares about whether the module is a derivative work? That's only
relevant when you distribute the module as a separate work. When you
ship a combined work including both the kernel and the module in
question, it's a _whole_ lot easier to interpret the GPL.

-- 
dwmw2

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Thursday 14 June 2007 18:24:55 David Woodhouse wrote:
> On Wed, 2007-06-13 at 21:29 -0400, Daniel Hazelton wrote:
> > Agreed. However, AFAICT, TiVO meets the provisions of the GPLv2 - they
> > make the source of the GPL'd part of their system available. (And I'm not
> > going to get into arguments over whether kernel modules are "derivative
> > works" or not, since those invariably end up with "They aren't, even
> > though we think they should be")
>
> Who cares about whether the module is a derivative work? That's only
> relevant when you distribute the module as a separate work. When you
> ship a combined work including both the kernel and the module in
> question, it's a _whole_ lot easier to interpret the GPL.

Agreed. I said I wasn't going to argue about it because there *ARE* 
distinctions that the law makes and the GPL ignores. You can't have it both 
ways. If the module is distributed *with* the kernel *SOURCE* then it doesn't 
matter if it's a derivative work or not, because it becomes covered by the 
kernels license. If it's distributed with the kernel *binaries* then it is 
covered by its own license. In that case the only reason you'd have a right 
to the source is if the module is considered a "derivative work".

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Woodhouse]

On Thu, 2007-06-14 at 21:44 -0400, Daniel Hazelton wrote:
> Agreed. I said I wasn't going to argue about it because there *ARE* 
> distinctions that the law makes and the GPL ignores. You can't have it both 
> ways. If the module is distributed *with* the kernel *SOURCE* then it doesn't 
> matter if it's a derivative work or not, because it becomes covered by the 
> kernels license. 

Yes.

> If it's distributed with the kernel *binaries* then it is 
> covered by its own license. In that case the only reason you'd have a right 
> to the source is if the module is considered a "derivative work".

Not necessarily. I'm not entirely sure where you got that idea from.

If the module is distributed 'as a separate work', _then_ what you say
is true: the only reason you'd have a right to the source is if the
module is considered a 'derivative work'.

But when you distribute the same module as part of a whole which is a
work based on the kernel, the distribution of the whole must be on the
terms of GPL, whose permissions for other licensees extend to the entire
whole, and thus to each and every part regardless of who wrote it.

The words you used were 'with the kernel', which could actually mean
either of the above. In the case of embedded Linux-based firmware
though, it's definitely the latter. It's a coherent whole, and it
contains both the kernel and the module. Thus the GPL extends to each
and every part, regardless of who wrote it. Including the module.

-- 
dwmw2

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Friday 15 June 2007 04:25:24 David Woodhouse wrote:
> On Thu, 2007-06-14 at 21:44 -0400, Daniel Hazelton wrote:
> > Agreed. I said I wasn't going to argue about it because there *ARE*
> > distinctions that the law makes and the GPL ignores. You can't have it
> > both ways. If the module is distributed *with* the kernel *SOURCE* then
> > it doesn't matter if it's a derivative work or not, because it becomes
> > covered by the kernels license.
>
> Yes.
>
> > If it's distributed with the kernel *binaries* then it is
> > covered by its own license. In that case the only reason you'd have a
> > right to the source is if the module is considered a "derivative work".
>
> Not necessarily. I'm not entirely sure where you got that idea from.
>
> If the module is distributed 'as a separate work', _then_ what you say
> is true: the only reason you'd have a right to the source is if the
> module is considered a 'derivative work'.
>
> But when you distribute the same module as part of a whole which is a
> work based on the kernel, the distribution of the whole must be on the
> terms of GPL, whose permissions for other licensees extend to the entire
> whole, and thus to each and every part regardless of who wrote it.

-ELOGIC

> The words you used were 'with the kernel', which could actually mean
> either of the above. In the case of embedded Linux-based firmware
> though, it's definitely the latter. It's a coherent whole, and it
> contains both the kernel and the module. Thus the GPL extends to each
> and every part, regardless of who wrote it. Including the module.

Just because two things are bundled together doesn't put them under the same 
license or copyright. Take a look at the GPL, which specifically mentions 
that "mere aggregation" does not cause something to fall under the GPL. Not 
that the GPL can even change the law - in the US copyright law specifically 
states that "mechanical translation" and "mechanical processes" *CANNOT* 
create a "new" work. Since the process of compiling source into a binary is, 
by definition, a *mechanical* process then the binary can't suddenly become 
covered by a different copyright license than the source code merely because 
of the medium on which its distributed or the manner in which it is 
distributed.

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Woodhouse]

On Fri, 2007-06-15 at 04:58 -0400, Daniel Hazelton wrote:
> > If the module is distributed 'as a separate work', _then_ what you say
> > is true: the only reason you'd have a right to the source is if the
> > module is considered a 'derivative work'.
> >
> > But when you distribute the same module as part of a whole which is a
> > work based on the kernel, the distribution of the whole must be on the
> > terms of GPL, whose permissions for other licensees extend to the entire
> > whole, and thus to each and every part regardless of who wrote it.
> 
> -ELOGIC

What's logic got to do with it? It was fairly much a direct quote from
the licence. You have _read_ the licence, haven't you?

> > The words you used were 'with the kernel', which could actually mean
> > either of the above. In the case of embedded Linux-based firmware
> > though, it's definitely the latter. It's a coherent whole, and it
> > contains both the kernel and the module. Thus the GPL extends to each
> > and every part, regardless of who wrote it. Including the module.
> 
> Just because two things are bundled together doesn't put them under the same 
> license or copyright. Take a look at the GPL, which specifically mentions 
> that "mere aggregation" does not cause something to fall under the GPL. Not 
> that the GPL can even change the law - in the US copyright law specifically 
> states that "mechanical translation" and "mechanical processes" *CANNOT* 
> create a "new" work. Since the process of compiling source into a binary is, 
> by definition, a *mechanical* process then the binary can't suddenly become 
> covered by a different copyright license than the source code merely because 
> of the medium on which its distributed or the manner in which it is 
> distributed.

You're confused.

If I grant you a licence on the condition that you give me money, would
you object on the basis that the money is not a 'derived work' of my
code? No. It's just a condition of the licence, and you're not allowed
to use my code unless you give me money.

If I grant you a licence on the condition that you sacrifice your
first-born son to Satan, would you object on the basis that your son is
not a 'derived work' of my code? No. It's just a condition of the
licence. If you don't do it, you don't have the right to use my code. 
(You may be able to get me locked up, but you still don't get to use my
code without a licence).

If I grant you a licence on the condition that you release _everything_
you write this year under the GPLv2, would you object on the basis that
your code is not a 'derived work' of my own? No. It's just a condition
of the licence, which you choose to accept or not.

If I grant you a licence on the condition that anything you release in
_combination_ with my code must also be released under the GPL, would
you object on the basis that you code is not a 'derived work' of my own?
No. Again, it's just a condition of the licence. If you don't want to
obey the licence, you don't get to use the kernel in the first place.

Talking about how your code can't possibly be a derived work is just a
red herring. The GPL explicitly talks about works which are 'independent
and separate works in themselves', to which the GPL does not apply 'when
you distribute them as separate works'.

But when you distribute the same sections as part of a whole which is a
work based on the Program, the distribution of the whole must be on the
terms of this License, whose permissions for other licensees extend to
the entire whole, and thus to each and every part regardless of who
wrote it.

It's your choice -- you're not _forced_ to use the kernel, and you're
not _forced_ to distribute a product which combines it with other code
of your own. But if you do, you're bound by the licence. And whether
your code is a 'derived work' has nothing to do with it.

Yes, there are exceptions for mere aggregation onto a storage medium --
if the kernel and your own work are next to each other on a backup tape
or your laptop's hard drive, or even both burned to a 'Gratis Software'
CD as _separate_ works, then that doesn't count. But we're talking about
a product which has a Linux kernel, a module built specifically for that
kernel, and cannot function unless both of them are present. That ain't
"mere aggregation on a storage medium".

-- 
dwmw2

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Friday 15 June 2007 05:17:44 David Woodhouse wrote:
> On Fri, 2007-06-15 at 04:58 -0400, Daniel Hazelton wrote:
> > > If the module is distributed 'as a separate work', _then_ what you say
> > > is true: the only reason you'd have a right to the source is if the
> > > module is considered a 'derivative work'.
> > >
> > > But when you distribute the same module as part of a whole which is a
> > > work based on the kernel, the distribution of the whole must be on the
> > > terms of GPL, whose permissions for other licensees extend to the
> > > entire whole, and thus to each and every part regardless of who wrote
> > > it.
> >
> > -ELOGIC
>
> What's logic got to do with it? It was fairly much a direct quote from
> the licence. You have _read_ the licence, haven't you?

Hrm... Perhaps I misread your post originally. Let me read it again and see if 
I didn't encounter a parsing error somewhere... Nope. Error of omission. The 
text you cut changes the meaning of the passage in its entirety.

Here, I'll quote it, in it's entirety:

These requirements apply to the modified work as a whole.  If
identifiable sections of that work are not derived from the Program,
and can be reasonably considered independent and separate works in
themselves, then this License, and its terms, do not apply to those
sections when you distribute them as separate works.  But when you
distribute the same sections as part of a whole which is a work based
on the Program, the distribution of the whole must be on the terms of
this License, whose permissions for other licensees extend to the
entire whole, and thus to each and every part regardless of who wrote it.

In other words, it applies to *SECTIONS* of the code, not to individual object 
code files. This is why kernel modules can have their own, separate license 
from the kernel. It isn't until the code is shipped as a *standard* part of 
the kernel that it has to be GPLv2. (Dynamic Linking, being a totally 
mechanical process, cannot create a derivative work under US copyright law, 
so please, don't try that old saw)

What this means is that it doesn't matter that a non-GPL module is shipped, 
in "object code" form with the "object code" form of the linux kernel it is 
designed to interface with - it *still* doesn't become automatically covered 
by the GPL.

> > > The words you used were 'with the kernel', which could actually mean
> > > either of the above. In the case of embedded Linux-based firmware
> > > though, it's definitely the latter. It's a coherent whole, and it
> > > contains both the kernel and the module. Thus the GPL extends to each
> > > and every part, regardless of who wrote it. Including the module.
> >
> > Just because two things are bundled together doesn't put them under the
> > same license or copyright. Take a look at the GPL, which specifically
> > mentions that "mere aggregation" does not cause something to fall under
> > the GPL. Not that the GPL can even change the law - in the US copyright
> > law specifically states that "mechanical translation" and "mechanical
> > processes" *CANNOT* create a "new" work. Since the process of compiling
> > source into a binary is, by definition, a *mechanical* process then the
> > binary can't suddenly become covered by a different copyright license
> > than the source code merely because of the medium on which its
> > distributed or the manner in which it is distributed.
>
> You're confused.

Nope. Not confused at all.

> If I grant you a licence on the condition that you give me money, would
> you object on the basis that the money is not a 'derived work' of my
> code? No. It's just a condition of the licence, and you're not allowed
> to use my code unless you give me money.

But you obviously are. After all, what does this have to do with whether the 
GPLv2 can "magically" change the law? 

> If I grant you a licence on the condition that you sacrifice your
> first-born son to Satan, would you object on the basis that your son is
> not a 'derived work' of my code? No. It's just a condition of the
> licence. If you don't do it, you don't have the right to use my code.
> (You may be able to get me locked up, but you still don't get to use my
> code without a licence).
>
> If I grant you a licence on the condition that you release _everything_
> you write this year under the GPLv2, would you object on the basis that
> your code is not a 'derived work' of my own? No. It's just a condition
> of the licence, which you choose to accept or not.

Again, what does this have to do with your apparent belief that me putting a 
binary of a kernel module that isn't GPL'd on a disc with the Linux kernel 
causes that module to become covered by the GPL?

> If I grant you a licence on the condition that anything you release in
> _combination_ with my code must also be released under the GPL, would
> you object on the basis that you code is not a 'derived work' of my own?
> No. Again, it's just a condition of the licence. If you don't want to
> obey the licence, you don't get to use the kernel in the first place.
>
> Talking about how your code can't possibly be a derived work is just a
> red herring. The GPL explicitly talks about works which are 'independent
> and separate works in themselves', to which the GPL does not apply 'when
> you distribute them as separate works'.

And it also says:
In addition, mere aggregation of another work not based on the Program
with the Program (or with a work based on the Program) on a volume of
a storage or distribution medium does not bring the other work under
the scope of this License.

In other words, even though I've built a program that isn't GPL'd, I can still 
put it on the same "volume of storage or distribution medium" as a GPL'd work 
and not have to put it under the GPL. Imagine that.

And again, I have to point out that, no matter what RMS and the rest of the 
people that wrote the GPLv2 may believe, it can't change the law on which it 
is based. If the law says "mechanical translation or processes cannot produce 
a new work", then no matter what the GPLv2 says, I can write a parser 
skeleton in the "YACC" language, run it through Bison and *NOT* need 
the "exception" that the FSF "thoughtfully" provides. Why? Because the 
*skeleton* I wrote carries the copyright - not the output of Bison, despite 
what the GPL *might* say on this in Section 0. A license *cannot* change the 
law, because it gets all its power from the law.

> But when you distribute the same sections as part of a whole which is a
> work based on the Program, the distribution of the whole must be on the
> terms of this License, whose permissions for other licensees extend to
> the entire whole, and thus to each and every part regardless of who
> wrote it.

Hrm... I see, you've included the section unmolested here - but you still seem 
unable to read it correctly. Perhaps I'm wrong though.

> It's your choice -- you're not _forced_ to use the kernel, and you're
> not _forced_ to distribute a product which combines it with other code
> of your own. But if you do, you're bound by the licence. And whether
> your code is a 'derived work' has nothing to do with it.

And is this what the process of making a module does? Because that *IS* what I 
had mentioned. If the mail I'm responding to was a response to that mail then 
you are sadly confused as to what I was talking about. 

> Yes, there are exceptions for mere aggregation onto a storage medium --
> if the kernel and your own work are next to each other on a backup tape
> or your laptop's hard drive, or even both burned to a 'Gratis Software'
> CD as _separate_ works, then that doesn't count. But we're talking about
> a product which has a Linux kernel, a module built specifically for that
> kernel, and cannot function unless both of them are present. That ain't
> "mere aggregation on a storage medium".

Yet it still doesn't make them a "combined work". If it did then the simple 
act of me installing a copy of the ATI or NVidia modules makes them GPL'd. 
But it doesn't - if it did I'm sure that somebody would have filed a lawsuit 
over this already.

DRH


-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Woodhouse]

On Fri, 2007-06-15 at 06:03 -0400, Daniel Hazelton wrote:
> In other words, it applies to *SECTIONS* of the code, not to individual object 
> code files. This is why kernel modules can have their own, separate license 
> from the kernel. It isn't until the code is shipped as a *standard* part of 
> the kernel that it has to be GPLv2. (Dynamic Linking, being a totally 
> mechanical process, cannot create a derivative work under US copyright law, 
> so please, don't try that old saw)
> 
> What this means is that it doesn't matter that a non-GPL module is shipped, 
> in "object code" form with the "object code" form of the linux kernel it is 
> designed to interface with - it *still* doesn't become automatically covered 
> by the GPL.

You're interpreting 'sections' to mean individual linker sections? And
you think it's talking about distributing those 'sections' "as separate
works"? Despite the fact that all the rest of the language in the
document is high-level and doesn't even mention _linking_? 

That's... an interesting interpretation.

> > If I grant you a licence on the condition that you give me money, would
> > you object on the basis that the money is not a 'derived work' of my
> > code? No. It's just a condition of the licence, and you're not allowed
> > to use my code unless you give me money.
> 
> But you obviously are. After all, what does this have to do with whether the 
> GPLv2 can "magically" change the law? 

The GPLv2 doesn't "magically" change the law, and has no need to. The
above is just a demonstration that a licence can have conditions which
involve things _other_ than derived works.

I can release code under a 'viral' licence which requires you to release
_everything_ you write for a whole year under that same licence. You
don't _have_ to obey, and there's no suggestion that your own code would
be 'derived' from mine -- but if you don't follow the conditions of the
licence, then you don't get permission to use my code. There's no
"magical" change to the law required.

> > If I grant you a licence on the condition that you sacrifice your
> > first-born son to Satan, would you object on the basis that your son is
> > not a 'derived work' of my code? No. It's just a condition of the
> > licence. If you don't do it, you don't have the right to use my code.
> > (You may be able to get me locked up, but you still don't get to use my
> > code without a licence).
> >
> > If I grant you a licence on the condition that you release _everything_
> > you write this year under the GPLv2, would you object on the basis that
> > your code is not a 'derived work' of my own? No. It's just a condition
> > of the licence, which you choose to accept or not.
> 
> Again, what does this have to do with your apparent belief that me putting a 
> binary of a kernel module that isn't GPL'd on a disc with the Linux kernel 
> causes that module to become covered by the GPL?

It's just a demonstration that a licence _can_ make requirements about
non-derived code. You seemed to be making two bogus claims -- first that
it _can_ not, and then that the GPL _does_ not.

I've dealt with the first; let's look at the second, leaving aside your
weird digression about bison... (hint: it's about the code stubs in the
output which weren't _produced_ mechanically in the first place; they
were only _put_ there mechanically by the software).

> > Talking about how your code can't possibly be a derived work is just a
> > red herring. The GPL explicitly talks about works which are 'independent
> > and separate works in themselves', to which the GPL does not apply 'when
> > you distribute them as separate works'.
> 
> And it also says:
> In addition, mere aggregation of another work not based on the Program
> with the Program (or with a work based on the Program) on a volume of
> a storage or distribution medium does not bring the other work under
> the scope of this License.
> 
> In other words, even though I've built a program that isn't GPL'd, I can still 
> put it on the same "volume of storage or distribution medium" as a GPL'd work 
> and not have to put it under the GPL. Imagine that.

Yes. That's why I said 'not necessarily' rather than 'no'. If it just
happens to be on the same hard drive / tape / CD-ROM that's not
important. The important question is whether it's distributed 'as a
separate work' or whether it's part of a larger, coherent whole.

> > But when you distribute the same sections as part of a whole which is a
> > work based on the Program, the distribution of the whole must be on the
> > terms of this License, whose permissions for other licensees extend to
> > the entire whole, and thus to each and every part regardless of who
> > wrote it.
> 
> Hrm... I see, you've included the section unmolested here - but you still seem 
> unable to read it correctly. Perhaps I'm wrong though.

If by 'correctly' you mean I should interpret 'sections' to mean linker
sections, then you're right -- I really can't bring myself to read it
that way. Of course, you're not actually _wrong_ until/unless it's
interpreted in court. But I'm willing to place bets :)

> > It's your choice -- you're not _forced_ to use the kernel, and you're
> > not _forced_ to distribute a product which combines it with other code
> > of your own. But if you do, you're bound by the licence. And whether
> > your code is a 'derived work' has nothing to do with it.
> 
> And is this what the process of making a module does? Because that *IS* what I 
> had mentioned. If the mail I'm responding to was a response to that mail then 
> you are sadly confused as to what I was talking about. 

I'm certainly confused as to what you're talking about _now_; it seems
to make little sense. What do you mean by 'this' in the context of 'is
this what the process of making a module does?'? And how is it relevant?

You said that modules aren't derivative works. I said 'Who cares?'
because there are far more obvious reasons why the module would be under
GPL, because of the GPL's requirements about collective works. What part
of that confuses you? You seem to _keep_ going back to talking about
derived works.

> > Yes, there are exceptions for mere aggregation onto a storage medium --
> > if the kernel and your own work are next to each other on a backup tape
> > or your laptop's hard drive, or even both burned to a 'Gratis Software'
> > CD as _separate_ works, then that doesn't count. But we're talking about
> > a product which has a Linux kernel, a module built specifically for that
> > kernel, and cannot function unless both of them are present. That ain't
> > "mere aggregation on a storage medium".
> 
> Yet it still doesn't make them a "combined work". If it did then the simple 
> act of me installing a copy of the ATI or NVidia modules makes them GPL'd. 
> But it doesn't - if it did I'm sure that somebody would have filed a lawsuit 
> over this already.

The GPL applies if you _distribute_ a work which combines such modules
and the kernel into a larger whole. You can do what you like if you
don't distribute it.

The 'mere aggregation' thing means you can even distribute them together
if they just happen to be side-by-side on a storage medium as if by
coincidence. But putting them together into a product which actually
uses and requires both of them is not 'mere aggregation'. Some people
_were_ bundling the ATI and nVidia modules together with the kernel in a
'product', and they stopped under threat of legal action.

The module on its own is questionable -- it is a matter of opinion
whether that's a derived work or not. But when you ship a product which
combines both kernel and module into a coherent whole, it doesn't
_matter_ if the module is a derived work or not. You are not permitted
to distribute the kernel unless your module is also licensed under the
GPL. Not because it's a derived work, but because the licence of the
kernel says that's the price you pay if you want to distribute the
kernel.

-- 
dwmw2

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Friday 15 June 2007 06:49:05 David Woodhouse wrote:
> On Fri, 2007-06-15 at 06:03 -0400, Daniel Hazelton wrote:
> > In other words, it applies to *SECTIONS* of the code, not to individual
> > object code files. This is why kernel modules can have their own,
> > separate license from the kernel. It isn't until the code is shipped as a
> > *standard* part of the kernel that it has to be GPLv2. (Dynamic Linking,
> > being a totally mechanical process, cannot create a derivative work under
> > US copyright law, so please, don't try that old saw)
> >
> > What this means is that it doesn't matter that a non-GPL module is
> > shipped, in "object code" form with the "object code" form of the linux
> > kernel it is designed to interface with - it *still* doesn't become
> > automatically covered by the GPL.
>
> You're interpreting 'sections' to mean individual linker sections? And
> you think it's talking about distributing those 'sections' "as separate
> works"? Despite the fact that all the rest of the language in the
> document is high-level and doesn't even mention _linking_?
>
> That's... an interesting interpretation.

And it's your own, twisted interpretation. When I say "Sections" I 
mean "portions of the source code", ie: files, functions, etc... In that 
manner the person who wrote that function, file, etc... (provided it meets 
the "minimum of artistic expression" required by US copyright law) has 
copyright to it. If it doesn't become compiled into - ie: linked into as a 
*standard* part of the build - the final executable the license on the final 
executable *cannot* have any effect on the stated code. QED: a kernel module, 
like "i8042.ko" *can* and *does* have a separate copyright *and* license from 
the linux kernel itself.

> > > If I grant you a licence on the condition that you give me money, would
> > > you object on the basis that the money is not a 'derived work' of my
> > > code? No. It's just a condition of the licence, and you're not allowed
> > > to use my code unless you give me money.
> >
> > But you obviously are. After all, what does this have to do with whether
> > the GPLv2 can "magically" change the law?
>
> The GPLv2 doesn't "magically" change the law, and has no need to. The
> above is just a demonstration that a licence can have conditions which
> involve things _other_ than derived works.
>
> I can release code under a 'viral' licence which requires you to release
> _everything_ you write for a whole year under that same licence. You
> don't _have_ to obey, and there's no suggestion that your own code would
> be 'derived' from mine -- but if you don't follow the conditions of the
> licence, then you don't get permission to use my code. There's no
> "magical" change to the law required.
>
> > > If I grant you a licence on the condition that you sacrifice your
> > > first-born son to Satan, would you object on the basis that your son is
> > > not a 'derived work' of my code? No. It's just a condition of the
> > > licence. If you don't do it, you don't have the right to use my code.
> > > (You may be able to get me locked up, but you still don't get to use my
> > > code without a licence).
> > >
> > > If I grant you a licence on the condition that you release _everything_
> > > you write this year under the GPLv2, would you object on the basis that
> > > your code is not a 'derived work' of my own? No. It's just a condition
> > > of the licence, which you choose to accept or not.
> >
> > Again, what does this have to do with your apparent belief that me
> > putting a binary of a kernel module that isn't GPL'd on a disc with the
> > Linux kernel causes that module to become covered by the GPL?
>
> It's just a demonstration that a licence _can_ make requirements about
> non-derived code. You seemed to be making two bogus claims -- first that
> it _can_ not, and then that the GPL _does_ not.

I never said it couldn't - a license can do whatever the hell it wants. What I 
said was that the license on one copyright work *cannot* just magically 
change the license on another work. The change of license *requires* the 
person holding copyright to *agree* to the change of license.

And no, the GPL *DOES* *NOT* have the requirement that a non-GPL'd work 
included on the same medium - for distribution or otherwise - must change its 
license to the GPL. If you think otherwise then you are sadly mistaken.

> I've dealt with the first; let's look at the second, leaving aside your
> weird digression about bison... (hint: it's about the code stubs in the
> output which weren't _produced_ mechanically in the first place; they
> were only _put_ there mechanically by the software).

About bison: doesn't matter. The code that they are included in *is* 
mechanically generated, as guided by the input file. QED: The output of Bison 
is a mechanical translation process *exactly* like the compiling of a C 
source file is a mechanical translation.

> > > Talking about how your code can't possibly be a derived work is just a
> > > red herring. The GPL explicitly talks about works which are
> > > 'independent and separate works in themselves', to which the GPL does
> > > not apply 'when you distribute them as separate works'.
> >
> > And it also says:
> > In addition, mere aggregation of another work not based on the Program
> > with the Program (or with a work based on the Program) on a volume of
> > a storage or distribution medium does not bring the other work under
> > the scope of this License.
> >
> > In other words, even though I've built a program that isn't GPL'd, I can
> > still put it on the same "volume of storage or distribution medium" as a
> > GPL'd work and not have to put it under the GPL. Imagine that.
>
> Yes. That's why I said 'not necessarily' rather than 'no'. If it just
> happens to be on the same hard drive / tape / CD-ROM that's not
> important. The important question is whether it's distributed 'as a
> separate work' or whether it's part of a larger, coherent whole.

And the GPL cannot define, on its own, what a "Separate Work" or a "Coherent 
Whole" is. That is defined by the relevant parts of copyright law. QED: The 
passage is largely irrelevant - if not, then the FSF claim that linking to a 
GPL'd library means your program is "magically" now GPL'd. It would also mean 
that every Linux live-cd that includes a non-GPL program is violating the 
GPL.

> > > But when you distribute the same sections as part of a whole which is a
> > > work based on the Program, the distribution of the whole must be on the
> > > terms of this License, whose permissions for other licensees extend to
> > > the entire whole, and thus to each and every part regardless of who
> > > wrote it.
> >
> > Hrm... I see, you've included the section unmolested here - but you still
> > seem unable to read it correctly. Perhaps I'm wrong though.
>
> If by 'correctly' you mean I should interpret 'sections' to mean linker
> sections, then you're right -- I really can't bring myself to read it
> that way. Of course, you're not actually _wrong_ until/unless it's
> interpreted in court. But I'm willing to place bets :)

Nope. To mean "source code files" or "functions in the source code". "Linker 
Sections" is your idea.

> > > It's your choice -- you're not _forced_ to use the kernel, and you're
> > > not _forced_ to distribute a product which combines it with other code
> > > of your own. But if you do, you're bound by the licence. And whether
> > > your code is a 'derived work' has nothing to do with it.
> >
> > And is this what the process of making a module does? Because that *IS*
> > what I had mentioned. If the mail I'm responding to was a response to
> > that mail then you are sadly confused as to what I was talking about.
>
> I'm certainly confused as to what you're talking about _now_; it seems
> to make little sense. What do you mean by 'this' in the context of 'is
> this what the process of making a module does?'? And how is it relevant?

The confusion is probably because you assumed I was an idiot who didn't have a 
clue what he was talking about. "This" means, if you really don't 
understand, "combining a GPL'd work with code of your own". So, does creating 
a kernel module that needs to be loaded into memory *separately* from the 
kernel "combine GPL'd code with your own". If that's what you think then you 
are wrong - see my comments about "mechanical processes" and such.

> You said that modules aren't derivative works. I said 'Who cares?'
> because there are far more obvious reasons why the module would be under
> GPL, because of the GPL's requirements about collective works. What part
> of that confuses you? You seem to _keep_ going back to talking about
> derived works.

Because the *only* relevant passages in the GPL, as applied to not-in-tree 
kernel modules is that about "derived works". *IF* distributing GPL'd 
binaries as part of an "Operating System" created a "collective work" then 
the FreeBSD people, who - last I looked - shipped with GCC as part of the OS, 
are in direct violation of the GPL. That means that either your argument is 
wrong or that every non-GPL'd OS that ships with GCC as an integral part is 
in violation of the GPL. 

> > > Yes, there are exceptions for mere aggregation onto a storage medium --
> > > if the kernel and your own work are next to each other on a backup tape
> > > or your laptop's hard drive, or even both burned to a 'Gratis Software'
> > > CD as _separate_ works, then that doesn't count. But we're talking
> > > about a product which has a Linux kernel, a module built specifically
> > > for that kernel, and cannot function unless both of them are present.
> > > That ain't "mere aggregation on a storage medium".
> >
> > Yet it still doesn't make them a "combined work". If it did then the
> > simple act of me installing a copy of the ATI or NVidia modules makes
> > them GPL'd. But it doesn't - if it did I'm sure that somebody would have
> > filed a lawsuit over this already.
>
> The GPL applies if you _distribute_ a work which combines such modules
> and the kernel into a larger whole. You can do what you like if you
> don't distribute it.
>
> The 'mere aggregation' thing means you can even distribute them together
> if they just happen to be side-by-side on a storage medium as if by
> coincidence. But putting them together into a product which actually
> uses and requires both of them is not 'mere aggregation'. Some people
> _were_ bundling the ATI and nVidia modules together with the kernel in a
> 'product', and they stopped under threat of legal action.
>
> The module on its own is questionable -- it is a matter of opinion
> whether that's a derived work or not. But when you ship a product which
> combines both kernel and module into a coherent whole, it doesn't
> _matter_ if the module is a derived work or not. You are not permitted
> to distribute the kernel unless your module is also licensed under the
> GPL. Not because it's a derived work, but because the licence of the
> kernel says that's the price you pay if you want to distribute the
> kernel.

Hrm... I see your point. When applied *solely* to the kernel and associated 
modules. But I can *still* see problems with that interpretation - very big 
ones that a good lawyer could exploit.

DRH

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Fri, 15 Jun 2007, David Woodhouse wrote:

> On Fri, 2007-06-15 at 04:58 -0400, Daniel Hazelton wrote:
> > >
> > > But when you distribute the same module as part of a whole which is a
> > > work based on the kernel, the distribution of the whole must be on the
> > > terms of GPL, whose permissions for other licensees extend to the entire
> > > whole, and thus to each and every part regardless of who wrote it.
> > 
> > -ELOGIC
> 
> What's logic got to do with it? It was fairly much a direct quote from
> the licence. You have _read_ the licence, haven't you?

Actually, I suspect Daniel has read it, and is probably referring to 
another facet of the license: distribution of two things together does 
*not* imply that those two things have to both be GPLv2's.

The GPLv2 explicitly mentions "mere aggregation". Strictly speaking, it 
doesn't even *have* to mention it, since it does mention in other places 
that it only covers "derived work", and "derivation" has nothing to do 
with "distributing two things together". But it's a good clarification.

So you guys are *both* right, for different cases!

The issue is simply what you mean by "part of the whole"? If you mean 
"part of the whole kernel distribution", then yes, the kernel is one work, 
and it is, in its entirety, under the GPLv2. But if the "part of the 
whole" is about something like a DVD with the whole being a collection of 
"mere aggregation", the licenses do not necessarily meld together.

Let's say that you're a Linux vendor, and you distribute a DVD with both 
the Linux kernel binary (and all the normal modules that go with it, that 
obviously are "part of the whole kernel") *and* say the NVidia proprietary 
kernel module.

Is that the *only* way to read things? No. It's a matter of 
interpretation, and which "whole" you are talking about. The whole 
aggregation, or the whole program?

And is the NVidia module a "derived work" or not? That's a gray area, and 
that's really what it hinges on. I personally think it's not, but I know 
others think it is.

Which is why I think you're both *potentially* right. Which one of you is 
*actually* right will depend on the exact circumstances ;)

			Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Woodhouse]

On Fri, 2007-06-15 at 08:58 -0700, Linus Torvalds wrote:
> The GPLv2 explicitly mentions "mere aggregation". Strictly speaking, it 
> doesn't even *have* to mention it, since it does mention in other places 
> that it only covers "derived work", and "derivation" has nothing to do 
> with "distributing two things together". But it's a good clarification.

Actually, I don't see where it explicitly states that it only covers
derived work.

On the other hand, I _do_ see where it explicitly states that in some
cases it _does_ 'infect' non-derived works. That's §2, which goes along
the lines of ... 'sections...not derived from Program' ...
'License...does...not apply...when...distribute...as separate works'.
'But..same sections as part of whole... License...extend...to each and
every part regardless of who wrote it'.

As you say, it goes on to _clarify_ that 'mere aggregation on a volume
of a storage or distribution medium' isn't what it's talking about here.

But there is _some_ class of collective work which combines entirely
non-derived work with the original GPL'd Program, and for which the GPL
requires that you release your non-derived code under the terms of the
GPL. You could argue till the cows come home about precisely what falls
into that category and what doesn't.

> So you guys are *both* right, for different cases!
> 
> The issue is simply what you mean by "part of the whole"? If you mean 
> "part of the whole kernel distribution", then yes, the kernel is one work, 
> and it is, in its entirety, under the GPLv2. But if the "part of the 
> whole" is about something like a DVD with the whole being a collection of 
> "mere aggregation", the licenses do not necessarily meld together.

Yep, those are two extremes which fall either side of the grey area and
are relatively easy to agree on... perhaps.

> Let's say that you're a Linux vendor, and you distribute a DVD with both 
> the Linux kernel binary (and all the normal modules that go with it, that 
> obviously are "part of the whole kernel") *and* say the NVidia proprietary 
> kernel module. 

A DVD with a Linux distribution isn't _quite_ the same as an arbitrary
bunch of gratis software which happens to be thrown together on a disc.
A _lot_ of work goes into making that a coherent product where
everything interoperates sanely, rather than just such an arbitrary sack
of bits.

There is at least one prominent North American Linux Vendor who has been
observed to claim that the distribution _is_ a collective work and
copyrightable in its own right -- which would mean that it _is_ a work
based on the Program, and thus that including binary-only modules in it
is not permitted. And there are other distributors who've stopped
including binary-only modules under threat of legal action (not that
_that_ necessarily means anything -- anyone can make baseless threats).

> And is the NVidia module a "derived work" or not? That's a gray area, and 
> that's really what it hinges on. I personally think it's not, but I know 
> others think it is.

If the DVD of the distribution is considered to be a work in itself; a
'work based on the Program', then it doesn't actually matter whether the
nVidia module is a derived work or not. Unless you're willing to
disregard those two paragraphs of §2 entirely?

The case which interests me most is when someone makes an embedded
device, for example a router -- and they  distribute a 'blob' of
firmware for it, containing both the kernel a binary-only network driver
module. Again we have to ask ourselves "is this a work based on the
kernel?". Obviously there isn't a 'right' answer outside a court of law,
but personally I reckon it's a fairly safe bet that it _is_ going to be
considered to be a work based on Linux. Especially as the thing would be
totally useless if you took away either part. They're both fundamental
to its operation.

> Which is why I think you're both *potentially* right. Which one of you
> is *actually* right will depend on the exact circumstances ;)

True.

-- 
dwmw2

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Fri, 15 Jun 2007, David Woodhouse wrote:
> 
> Actually, I don't see where it explicitly states that it only covers
> derived work.

See "Section 0":

	The "Program", below, refers to any such program or work, and a 
	"work based on the Program" means either the Program or any 
	derivative work under copyright law:

so yes, if you grepped for "derived work", you wouldn't have found it. The 
exact wording used in the license is "derivative work under copyright 
law".

So the very *definition* of the word "Program" is indeed limited by the 
notion of "derived work" - as defined by copyright law, and NOT the GPLv2.

> The case which interests me most is when someone makes an embedded
> device, for example a router -- and they  distribute a 'blob' of
> firmware for it, containing both the kernel a binary-only network driver
> module. Again we have to ask ourselves "is this a work based on the
> kernel?". Obviously there isn't a 'right' answer outside a court of law,
> but personally I reckon it's a fairly safe bet that it _is_ going to be
> considered to be a work based on Linux.

Hey, I kind of disagree.

What is a DVD? It's just a "blob" of a UDF image, potentially containing 
the Linux kernel.

How is that different from a "blob" of some other kind of image (say, a 
cramfs or similar image) on a rom?

What makes UDF so different from cramfs? What makes a DVD so different 
from a ROM chip? Why would copyright law care about one and not the other?

So I really do _not_ think it's at all obvious. Personally, I think it's 
exactly the same case. Others disagree, but I've never really seen a good 
*reason* for them disagreeing.

		Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Woodhouse]

On Fri, 2007-06-15 at 11:23 -0700, Linus Torvalds wrote:
> 
> On Fri, 15 Jun 2007, David Woodhouse wrote:
> > 
> > Actually, I don't see where it explicitly states that it only covers
> > derived work.
> 
> See "Section 0":
> 
> 	The "Program", below, refers to any such program or work, and a 
> 	"work based on the Program" means either the Program or any 
> 	derivative work under copyright law:
> 
> so yes, if you grepped for "derived work", you wouldn't have found it. The 
> exact wording used in the license is "derivative work under copyright 
> law".
> 
> So the very *definition* of the word "Program" is indeed limited by the 
> notion of "derived work" - as defined by copyright law, and NOT the GPLv2.

Yep. And §2 talks explicitly about independent and separate works when
they are distributed _with_ the Program, as part of a larger work based
on the Program.

> > The case which interests me most is when someone makes an embedded
> > device, for example a router -- and they  distribute a 'blob' of
> > firmware for it, containing both the kernel a binary-only network driver
> > module. Again we have to ask ourselves "is this a work based on the
> > kernel?". Obviously there isn't a 'right' answer outside a court of law,
> > but personally I reckon it's a fairly safe bet that it _is_ going to be
> > considered to be a work based on Linux.
> 
> Hey, I kind of disagree.
> 
> What is a DVD? It's just a "blob" of a UDF image, potentially containing 
> the Linux kernel.
> 
> How is that different from a "blob" of some other kind of image (say, a 
> cramfs or similar image) on a rom?
> 
> What makes UDF so different from cramfs? What makes a DVD so different 
> from a ROM chip? Why would copyright law care about one and not the other?

The differences are subtle, but they do exist. They're not really about
whether it's iso9660 or cramfs; it's about whether what you put on them
is a coherent work in its own right or just a bunch of bits which happen
to be thrown together onto the same medium.

And in the router case, there's little point to its existence without
the binary-only module. At least with the DVD it _can_ work without the
binary-only module. Although as I said, some distributors definitely
claim that the distribution is a 'coherent whole' too.

> So I really do _not_ think it's at all obvious. Personally, I think it's 
> exactly the same case. Others disagree, but I've never really seen a good 
> *reason* for them disagreeing.

It's a grey area, and nobody's 'right' until/unless a court decides. And
then only until/unless a higher court contradicts it. The reason I
jumped in was to point out that it isn't _just_ about whether the module
is a derived work or not. The GPL goes further than that.

-- 
dwmw2

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Ingo Molnar]

* David Woodhouse <dwmw2@infradead.org> wrote:

> > So I really do _not_ think it's at all obvious. Personally, I think 
> > it's exactly the same case. Others disagree, but I've never really 
> > seen a good *reason* for them disagreeing.
> 
> It's a grey area, and nobody's 'right' until/unless a court decides. 
> And then only until/unless a higher court contradicts it. The reason I 
> jumped in was to point out that it isn't _just_ about whether the 
> module is a derived work or not. The GPL goes further than that.

i think you are putting too much weight behind the distinction between 
derivatives/modifications and collective works / collections. The two 
are very closely related: a derivative/modification is a work that 
arises out of an existing work, while a collective work arises out of a 
set of pre-existing works by adding some 'glue' to it (where the glue 
itself is not a work in itself - the whole thing is the new work). In 
fact one could argue that a derivative is a collective work based on a 
_single_ preexisting work!

and thus the whole issue of "what is a whole", how strong the "glue" 
needs to be so that the copyright of the collective work is meaningful 
on its own and starts to affect every component and requires each of 
them to be GPL licensed. This largely depends on how deeply a 
distributor integrates said binary blobs. For example some Linux 
distributors certainly found it safe enough to ship restricted software 
on separate medium - even if they happen to be in the same physical 
package. (which one could attempt to argue to be 'one work'.)

That boundary is indeed fuzzy, because life is fuzzy too and the 
possibilities are virtually unlimited. But one thing is pretty sure: as 
long as some component is merely put alongside of a larger body of work, 
even if that component has no life of its own without _some_ larger body 
of work, that component is not necessarily part of a collective work and 
does not necessarily fall under the GPL. (unless it falls under the GPL 
for entirely different reasons: for example it was continuously 
developed out of internals of the 'larger body of work' and thus became 
a derivative of the larger body of work.)

For driver blobs that are shared between Windows and Linux it would be 
hard to argue that they are derived from the Linux kernel. Merely 
linking to some larger body of work does not necessarily mean that the 
two become a collective work. No matter how much the FSF is trying to 
muddy the waters with the LGPL/GPL.

	Ingo

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Woodhouse]

On Fri, 2007-06-15 at 22:20 +0200, Ingo Molnar wrote:
> That boundary is indeed fuzzy, because life is fuzzy too and the 
> possibilities are virtually unlimited. But one thing is pretty sure: as 
> long as some component is merely put alongside of a larger body of work, 
> even if that component has no life of its own without _some_ larger body 
> of work, that component is not necessarily part of a collective work and 
> does not necessarily fall under the GPL.

Not _necessarily_ a collective work. But not necessarily _not_ a
collective work either.

> For driver blobs that are shared between Windows and Linux it would be 
> hard to argue that they are derived from the Linux kernel. 

You're back to the 'derived work' thing again, which wasn't relevant.

> Merely linking to some larger body of work does not necessarily mean
> that the two become a collective work. No matter how much the FSF is
> trying to muddy the waters with the LGPL/GPL.

I think it's quite clear that the intent of the GPL _is_ to 'muddy the
waters', as you put it, and to indicate that bundling stuff together
_should_ put the non-derived parts under the GPL too; at least in some
circumstances. But still, nothing's true until it's ruled by a court.

-- 
dwmw2

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Ingo Molnar]

* David Woodhouse <dwmw2@infradead.org> wrote:

> On Fri, 2007-06-15 at 22:20 +0200, Ingo Molnar wrote:
> > That boundary is indeed fuzzy, because life is fuzzy too and the 
> > possibilities are virtually unlimited. But one thing is pretty sure: as 
> > long as some component is merely put alongside of a larger body of work, 
> > even if that component has no life of its own without _some_ larger body 
> > of work, that component is not necessarily part of a collective work and 
> > does not necessarily fall under the GPL.
> 
> Not _necessarily_ a collective work. But not necessarily _not_ a
> collective work either.
> 
> > For driver blobs that are shared between Windows and Linux it would be 
> > hard to argue that they are derived from the Linux kernel. 
> 
> You're back to the 'derived work' thing again, which wasn't relevant.

yeah - i keep interchanging the two because they are so closely related. 
(the same goes for modification and derivation - for software the two 
are quite similar.) Depending on how deeply a distributor integrates a 
binary blob, it might or might not fall under the umbrella of a 
collective work, and the GPL (covering other components of the 
collective work) might or might not apply.

> > Merely linking to some larger body of work does not necessarily mean 
> > that the two become a collective work. No matter how much the FSF is 
> > trying to muddy the waters with the LGPL/GPL.
> 
> I think it's quite clear that the intent of the GPL _is_ to 'muddy the 
> waters', as you put it, and to indicate that bundling stuff together 
> _should_ put the non-derived parts under the GPL too; at least in some 
> circumstances. But still, nothing's true until it's ruled by a court.

but it's not up to the GPL to define that! Whether something is a 
collective work is a matter of law (which operates on the specific facts 
of the case), not a matter of licensing.

and that's where the GPLv3 errs: it arbitrarily attempts to "define" 
some work that can _easily_ be completely separate from the GPL-ed work 
to be under the scope of "source code". Yes, it can do that legally 
because its framers knew what they were doing and they did not attempt 
to implement it as a 'this work belongs to us' thing (which would be 
misuse of copyright) but as a 'you got to pay with your work for our 
permission' - but the external communications about this is all false: 
the pretense that the key in the Tivo case somehow belongs to the GPL-ed 
work is just bogus. A key _can_ belong to a GPL-ed work, but it does not 
automatically so. The GPLv3 automatically and unconditionally moves it 
under the scope of the license and that aspect of the GPLv3 is just 
wrong and moves the license closer to a Microsoft EULA contract than 
towards a pure and just copyright license.

	Ingo

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Fri, 15 Jun 2007, David Woodhouse wrote:
> > 
> > What is a DVD? It's just a "blob" of a UDF image, potentially containing 
> > the Linux kernel.
> > 
> > How is that different from a "blob" of some other kind of image (say, a 
> > cramfs or similar image) on a rom?
> > 
> > What makes UDF so different from cramfs? What makes a DVD so different 
> > from a ROM chip? Why would copyright law care about one and not the other?
> 
> The differences are subtle, but they do exist. They're not really about
> whether it's iso9660 or cramfs; it's about whether what you put on them
> is a coherent work in its own right or just a bunch of bits which happen
> to be thrown together onto the same medium.

I think that's a somewhat valid argument, although I'm not really sure 
whether there is any difference between, say, a Fedora 7 "livecd", and a 
router with a cramfs filesystem in rom.

Both really work the same way, and both really are very much targeted 
towards a specific hardware platform.

Yes, it's true that a small router migth be a more *coherent* hardware 
platform than the Fedora 7 livecd is, but that's more a factor of the wild 
and crazy PC hardware culture than of the small router.

For example, what about a livecd for PPC-based hardware? Those tend to be 
much more uniform (read: I think the livecd's generally work on mac 
clones).

So it's a question of degrees of separation. Does it make a difference 
that some of these embedded images work across a whole range of (rather 
similar, but still.. not identical) routers? 

I don't really have any point, except that there is no real *technical* 
difference, and in many respects the only difference in the end really 
seems to be about "intended target device or audience" rather than 
anything else.

How can you make hard licensing decisions in situations like that? I don't 
think you really can. In fact, the OSI rules even forbid making licensing 
decisions based on things that get rather close to the differences you are 
describing (both the "not specific to a product" and "license must not 
discriminate against fields of endeavor").

It's also really really *hard* to make a choice based on a gradual scale. 
Where do you put the limit? Wherever you put it, it's going to be 
arbitrary. Is that really a good thing?

So I would at least *personally* suggest that people not look into the 
license for these kinds of things, and also that you really need to have a 
very specific case, and just basically put it in front of a judge.

At some point, *somebody* has to decide in a gray area, and I'm not saying 
that a judge is really _technically_ any better really to decide the 
issue, but at least he is hopefully _independent_ of both parties, so when 
a judge makes an arbitrary decision, the "arbitrariness" is hopefully at 
least somewhat "fair".

Hmm? 

			Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Woodhouse]

On Fri, 2007-06-15 at 13:34 -0700, Linus Torvalds wrote:
> I think that's a somewhat valid argument, although I'm not really sure 
> whether there is any difference between, say, a Fedora 7 "livecd", and a 
> router with a cramfs filesystem in rom.
> 
> Both really work the same way, and both really are very much targeted 
> towards a specific hardware platform.

I'm inclined to agree. And I'd probably suggest that the Fedora 7
'livecd' would be in violation of the GPL if it were to include the
binary-only modules, too. Enough people agree with me that we _don't_ in
fact include those modules. And other people have been convinced to
_stop_ shipping those modules, when once they did.

> So I would at least *personally* suggest that people not look into the 
> license for these kinds of things, and also that you really need to have a 
> very specific case, and just basically put it in front of a judge.

> At some point, *somebody* has to decide in a gray area, and I'm not saying 
> that a judge is really _technically_ any better really to decide the 
> issue, but at least he is hopefully _independent_ of both parties, so when 
> a judge makes an arbitrary decision, the "arbitrariness" is hopefully at 
> least somewhat "fair".

Indeed.

-- 
dwmw2

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Friday 15 June 2007 15:49:15 David Woodhouse wrote:
> On Fri, 2007-06-15 at 11:23 -0700, Linus Torvalds wrote:
> > On Fri, 15 Jun 2007, David Woodhouse wrote:
> > > Actually, I don't see where it explicitly states that it only covers
> > > derived work.
> >
> > See "Section 0":
> >
> > 	The "Program", below, refers to any such program or work, and a
> > 	"work based on the Program" means either the Program or any
> > 	derivative work under copyright law:
> >
> > so yes, if you grepped for "derived work", you wouldn't have found it.
> > The exact wording used in the license is "derivative work under copyright
> > law".
> >
> > So the very *definition* of the word "Program" is indeed limited by the
> > notion of "derived work" - as defined by copyright law, and NOT the
> > GPLv2.
>
> Yep. And §2 talks explicitly about independent and separate works when
> they are distributed _with_ the Program, as part of a larger work based
> on the Program.

>From section 0 of the GPLv2:
The "Program", below, refers to any such program or work, and a "work based on 
the Program" means either the Program or any derivative work under copyright 
law: that is to say, a work containing the Program or a portion of it, either 
verbatim or with modifications and/or translated into another language.  
(Hereinafter, translation is included without limitation in the 
term "modification".)

In other words, the "that is to say, a work containing the Program or a 
portion of it, either verbatim or with modifications and/or translated into 
another language." is a clarification of the terms so that you do *NOT* have 
to know copyright law. However, the license, being based in copyright law, 
*CANNOT* change that law without making itself invalid. QED: What copyright 
law says is a "derivative work" is what matters, not the definition provided 
in the license.

DRH

> > > The case which interests me most is when someone makes an embedded
> > > device, for example a router -- and they  distribute a 'blob' of
> > > firmware for it, containing both the kernel a binary-only network
> > > driver module. Again we have to ask ourselves "is this a work based on
> > > the kernel?". Obviously there isn't a 'right' answer outside a court of
> > > law, but personally I reckon it's a fairly safe bet that it _is_ going
> > > to be considered to be a work based on Linux.
> >
> > Hey, I kind of disagree.
> >
> > What is a DVD? It's just a "blob" of a UDF image, potentially containing
> > the Linux kernel.
> >
> > How is that different from a "blob" of some other kind of image (say, a
> > cramfs or similar image) on a rom?
> >
> > What makes UDF so different from cramfs? What makes a DVD so different
> > from a ROM chip? Why would copyright law care about one and not the
> > other?
>
> The differences are subtle, but they do exist. They're not really about
> whether it's iso9660 or cramfs; it's about whether what you put on them
> is a coherent work in its own right or just a bunch of bits which happen
> to be thrown together onto the same medium.
>
> And in the router case, there's little point to its existence without
> the binary-only module. At least with the DVD it _can_ work without the
> binary-only module. Although as I said, some distributors definitely
> claim that the distribution is a 'coherent whole' too.
>
> > So I really do _not_ think it's at all obvious. Personally, I think it's
> > exactly the same case. Others disagree, but I've never really seen a good
> > *reason* for them disagreeing.
>
> It's a grey area, and nobody's 'right' until/unless a court decides. And
> then only until/unless a higher court contradicts it. The reason I
> jumped in was to point out that it isn't _just_ about whether the module
> is a derived work or not. The GPL goes further than that.



-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 15, 2007, David Woodhouse <dwmw2@infradead.org> wrote:

> But when you distribute the same module as part of a whole which is a
> work based on the kernel, the distribution of the whole must be on the
> terms of GPL,

... unless the other parts of the whole fall under the mere
aggregation exception, methinks, but IANAL.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Ingo Molnar]

* David Woodhouse <dwmw2@infradead.org> wrote:

> Who cares about whether the module is a derivative work? That's only 
> relevant when you distribute the module as a separate work. When you 
> ship a combined work including both the kernel and the module in 
> question, it's a _whole_ lot easier to interpret the GPL.

i fully support the notion you articulate, that whether bin-only modules 
are part of a derivative work of the kernel or whether they are 
independent works is not an automatic thing at all. The answer is: "it 
depends, talk to your lawyer". For example i'd say VMWare's ESX bin-only 
module is likely derived from the Linux kernel and should be distributed 
under the GPL, but that for example the ATI and nvidia drivers, although 
being a large PITA for all of us, are possibly independent works.

but lets note that this is irrelevant to the Tivo argument. Tivo is not 
using bin-only modules AFAIK, all their source code is available for 
download. (their kernel source is totally uninteresting by the way - 
they have some weird crap IDE controller with hacks that will never go 
upstream.)

	Ingo

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Woodhouse]

On Fri, 2007-06-15 at 13:49 +0200, Ingo Molnar wrote:
> i fully support the notion you articulate, that whether bin-only modules 
> are part of a derivative work of the kernel or whether they are 
> independent works is not an automatic thing at all. The answer is: "it 
> depends, talk to your lawyer".

I was actually trying to avoid the question altogether. It's not that
interesting, largely because the answer is indeed 'talk to your lawyer'.

> For example i'd say VMWare's ESX bin-only module is likely derived
> from the Linux kernel and should be distributed under the GPL, but
> that for example the ATI and nvidia drivers, although being a large
> PITA for all of us, are possibly independent works.

And thus not affected by the GPL _if_ they are distributed as separate
works in their own right. But if you bundle them with the kernel into a
product, the GPL has something to say about that.

> but lets note that this is irrelevant to the Tivo argument. Tivo is not 
> using bin-only modules AFAIK,

Right. It was a digression, which I picked up on because people were
talking about derived works in the context of modules again, and missing
the point that the most _obvious_ GPL violation with modules doesn't
actually involve those modules being a derived work at all.

-- 
dwmw2

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Ingo Molnar]

* David Woodhouse <dwmw2@infradead.org> wrote:

> > For example i'd say VMWare's ESX bin-only module is likely derived 
> > from the Linux kernel and should be distributed under the GPL, but 
> > that for example the ATI and nvidia drivers, although being a large 
> > PITA for all of us, are possibly independent works.
> 
> And thus not affected by the GPL _if_ they are distributed as separate 
> works in their own right. But if you bundle them with the kernel into 
> a product, the GPL has something to say about that.

yeah. Section 2 of the GPLv2 takes a permissive (and IMO correct) 
approach here:

  Thus, it is not the intent of this section to claim rights or contest
  your rights to work written entirely by you; rather, the intent is to
  exercise the right to control the distribution of derivative or
  collective works based on the Program.

  In addition, mere aggregation of another work not based on the Program
  with the Program (or with a work based on the Program) on a volume of
  a storage or distribution medium does not bring the other work under
  the scope of this License.

the argument is quite strong that the linking of two independent works 
is "mere aggregation" as well. (as long as they are truly separate 
works)

	Ingo

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Woodhouse]

On Fri, 2007-06-15 at 14:29 +0200, Ingo Molnar wrote:
> the argument is quite strong that the linking of two independent works 
> is "mere aggregation" as well. (as long as they are truly separate 
> works) 

You think so?

If even linking was considered 'mere aggregation on a volume of a
storage or distribution medium', then when would the 'But when you
distribute those same sections as part of a whole...' bit _ever_ apply?
It _explicitly_ talks of sections which are independent and separate
works in their own right, but which must be licensed under the GPL when
they're distributed as part of a larger whole.

I don't see how we could hold the view that _even_ linking is 'mere
aggregation on a volume of a storage or distribution medium', without
conveniently either ignoring entire paragraphs of the GPL or declaring
them to be entirely meaningless.

Of course, that doesn't mean that a court _wouldn't_ do that. Given
enough money, I'm sure you could get US court to declare that the world
is flat. But it doesn't seem to be a reasonable viewpoint, to me. Or a
likely outcome.


-- 
dwmw2

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Ingo Molnar]

* David Woodhouse <dwmw2@infradead.org> wrote:

> If even linking was considered 'mere aggregation on a volume of a 
> storage or distribution medium', then when would the 'But when you 
> distribute those same sections as part of a whole...' bit _ever_ 
> apply? It _explicitly_ talks of sections which are independent and 
> separate works in their own right, but which must be licensed under 
> the GPL when they're distributed as part of a larger whole.
>
> I don't see how we could hold the view that _even_ linking is 'mere 
> aggregation on a volume of a storage or distribution medium', without 
> conveniently either ignoring entire paragraphs of the GPL or declaring 
> them to be entirely meaningless.

as long as it's not distributed in one collective work, where is the 
problem? A driver could be argued to be part of a mere compilation of 
works (not part of a collective work), or just two separate works. But 
... this is a much greyer area than the key stuff.

> Of course, that doesn't mean that a court _wouldn't_ do that. Given 
> enough money, I'm sure you could get US court to declare that the 
> world is flat. But it doesn't seem to be a reasonable viewpoint, to 
> me. Or a likely outcome.

i'm not that cynical about US courts.

	Ingo

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Woodhouse]

On Fri, 2007-06-15 at 14:58 +0200, Ingo Molnar wrote:
> * David Woodhouse <dwmw2@infradead.org> wrote:
> 
> > If even linking was considered 'mere aggregation on a volume of a 
> > storage or distribution medium', then when would the 'But when you 
> > distribute those same sections as part of a whole...' bit _ever_ 
> > apply? It _explicitly_ talks of sections which are independent and 
> > separate works in their own right, but which must be licensed under 
> > the GPL when they're distributed as part of a larger whole.
> >
> > I don't see how we could hold the view that _even_ linking is 'mere 
> > aggregation on a volume of a storage or distribution medium', without 
> > conveniently either ignoring entire paragraphs of the GPL or declaring 
> > them to be entirely meaningless.
> 
> as long as it's not distributed in one collective work, where is the 
> problem?

As long as it's not distributed "as part of a whole which is a work
based on the Program", there's no problem.

You seem to be suggesting that even linking the Program together with
other stuff doesn't create a 'work based on the Program'. You seem claim
it's "mere aggregation on a volume of a storage or distribution medium".
Am I understanding you correctly?

Is there _anything_ which you admit would actually constitute a 'work
based on the Program', when that work wouldn't have been be a derived
work anyway? Or do you claim that those whole paragraphs of the GPL are
just meaningless drivel, when they explicitly make reference to applying
the GPL to works which would _normally_ be 'considered independent and
separate works in themselves'?

If your interpretation of the GPL means that those paragraphs don't make
any sense at all, then I feel your interpretation may be suspect.

-- 
dwmw2

RE: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Schwartz]

> As long as it's not distributed "as part of a whole which is a work
> based on the Program", there's no problem.
>
> You seem to be suggesting that even linking the Program together with
> other stuff doesn't create a 'work based on the Program'. You seem claim
> it's "mere aggregation on a volume of a storage or distribution medium".
> Am I understanding you correctly?

Correct. Linking does not create a "work based on the program" because
linking does not create a work. Only a creative process can create a work.
Anything else is mere aggregation.

> Is there _anything_ which you admit would actually constitute a 'work
> based on the Program', when that work wouldn't have been be a derived
> work anyway?

These terms are synonymous. And neither of them can apply to something that
is not a work.

> Or do you claim that those whole paragraphs of the GPL are
> just meaningless drivel, when they explicitly make reference to applying
> the GPL to works which would _normally_ be 'considered independent and
> separate works in themselves'?

The license is just clarifying copyright law. Even if it intended to do
something else, it can't. The license cannot set its own scope.

> If your interpretation of the GPL means that those paragraphs don't make
> any sense at all, then I feel your interpretation may be suspect.

They make perfect sense. They're clarifications of copyright law to help
people who might not be familiar with the law understand what their
obligations under the license are. All of those sections are reasonable
explanations of what a "derivative work" is. It would be extremely strange
to try to parse them for subtle differences.

In any event, even if the GPL said "if you ever look at any source code to a
GPL'd work, the FSF owns everything you code after that", it wouldn't
matter. The GPL can't set its own scope. Copyright law, and the definition
of a derivative work, set the GPL's scope anyway.

The GPL only makes sense if you understand "mere aggregation" to mean 'as
opposed to creative combination'.

DS

RE: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Schwartz]

By the way, the unfortunate answer to the question of what the default
position is when contributions to a collective work are received without
explicit license, at least in the United States, is:

"In the absence of an express transfer of the copyright or of any rights
under it, the owner of copyright in the collective work is presumed to have
acquired only the privilege of reproducing and distributing the contribution
as part of that particular collective work, any revision of that collective
work, and any later collective work in the same series." -- USC 201(c)

That is, as I understand the law, if you receive a contribution to a project
without any specific license in that contribution, it works like this:

1) You automatically receive a license to use that piece as part of that
project by virtue of the fact that it was contributed by the author to that
project. (Because 210(c) says so.)

2) If the contribution is itself a derivative work of a GPL'd work, then you
receive a GPL license. (Because the GPL says so).

So it would be very unwise to add a contribution that wasn't itself a
derivative work without clear indication from the author that the
contribution is offered under the license you need. I had assumed no law set
a default, and therefore the default would be the project's license. THIS IS
INCORRECT. 201(c) sets the default, and it's the wrong one.

This means that contributions of non-derivative works to GPL projects should
not be added to the project unless the author specifically licenses that
piece under the GPL. I would not consider it safe to assume that the fact
that the author knowingly contributed the work to a GPL'd project is
sufficient to change the 201(c) default.

What's worse, section 203 appears to grant the author various ways, by law,
to *terminate* a license grant. This termination removes the ability to
create subsequent derivative works. Ouch.

http://www.copyright.gov/title17/92chap2.html

I sure hope I'm misunderstanding something.

DS

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Friday 15 June 2007 09:12:43 David Woodhouse wrote:
> On Fri, 2007-06-15 at 14:58 +0200, Ingo Molnar wrote:
> > * David Woodhouse <dwmw2@infradead.org> wrote:
> > > If even linking was considered 'mere aggregation on a volume of a
> > > storage or distribution medium', then when would the 'But when you
> > > distribute those same sections as part of a whole...' bit _ever_
> > > apply? It _explicitly_ talks of sections which are independent and
> > > separate works in their own right, but which must be licensed under
> > > the GPL when they're distributed as part of a larger whole.
> > >
> > > I don't see how we could hold the view that _even_ linking is 'mere
> > > aggregation on a volume of a storage or distribution medium', without
> > > conveniently either ignoring entire paragraphs of the GPL or declaring
> > > them to be entirely meaningless.
> >
> > as long as it's not distributed in one collective work, where is the
> > problem?
>
> As long as it's not distributed "as part of a whole which is a work
> based on the Program", there's no problem.

Agreed.

> You seem to be suggesting that even linking the Program together with
> other stuff doesn't create a 'work based on the Program'. You seem claim
> it's "mere aggregation on a volume of a storage or distribution medium".
> Am I understanding you correctly?

Yes, you are.

> Is there _anything_ which you admit would actually constitute a 'work
> based on the Program', when that work wouldn't have been be a derived
> work anyway? Or do you claim that those whole paragraphs of the GPL are
> just meaningless drivel, when they explicitly make reference to applying
> the GPL to works which would _normally_ be 'considered independent and
> separate works in themselves'?

Nope. In fact, "work based on the program" is so unclear that it means that if 
I wrote a book about the creation of the Linux Kernel that is entirely 
original - containing nothing that is copyright someone else - I would have 
to release it under the GPL simply because it is a "work based on the 
program".

Is it okay to make that demand? I don't think so. But that is *exactly* what 
it means. And it is "the GPL applying itself to works which would normally be 
considered independent and separate works in themselves".

DRH

> If your interpretation of the GPL means that those paragraphs don't make
> any sense at all, then I feel your interpretation may be suspect.



-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Willy Tarreau]

On Wed, Jun 13, 2007 at 03:06:51PM -0700, Linus Torvalds wrote:
> And I actually am of the very firm opinion that a world with gray areas 
> (and purple, and pink, and green) is a hell of a lot better than one where 
> everything is black-and-white.

agreed, because you cannot imagine at the beginning all fair uses of your
project. It's a good thing that people can use it, thinking "hey, it's not
explicitly allowed but I think I can defend my case".

> Only lawyers want a black-and-white world.

not really. They would lose their job. They need a gray world to get
customers, but they want to decide what half is black and what half
is white in front of the judge depending on their customer's needs.

> Indeed. And it's _fine_ to even be in it "just to make a quick buck". We 
> do want all kinds of input. I think the community is much healthier having 
> lots of different reasons for people wanting to be involved, rather than 
> concentrating on just some specific reason.
> 
> For some it's the technology. For some it's the license. For some it's 
> just a thing to pass boredom. Others like to learn. Whatever. It's all 
> good!

And I think that for many people (including myself), it's all of these in
this order :
  - something to learn (when you're at school)
  - something to pass boredom (when you're at school too)
  - the technology (when you're working on designing new products)
  - the license (when you finally try to put your products on the market)

Regards,
Willy

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 13, 2007, Linus Torvalds <torvalds@linux-foundation.org> wrote:

> On Wed, 13 Jun 2007, Alexandre Oliva wrote:
>> 
>> Look, there was room for misunderstandings in earlier drafts of the
>> license.  Based on the public comments, the wording was improved.  I'd
>> like to think the issues that arose from misunderstandings of the
>> earlier drafts are no longer an issue.  Is it not so?

> No. The anti-DRM language is still there, and no, it was never a 
> misunderstanding.

It was claimed that GPLv3 would forbid implementations of DRM.  That's
just plain false.  If you don't think so, please show what terms in
the latest draft prohibit DRM (as opposed to merely making it
ineffective, a necessary consequence of abiding by the spirit of all
GNU GPLs)

> It's offensive because Tivo never did anything wrong, and the FSF
> even acknowledged that.

Another misunderstanding.  The FSF never said TiVo didn't do anything
wrong.  It only said it didn't think there was a license violation.  I
personally disagree with that assessment, but IANAL.

Anyhow, deciding whether it's right or wrong is not the same as
deciding whether it's legal or illegal.  Law doesn't define what's
right or wrong.  That's what morals and ethics do.

> want to protect the integrity of that hardware.

> The kernel license covers the *kernel*.

When they choose to include a copy of the kernel in their hardware
that they can modify but others can't, they're failing to comply with
the spirit of the license.  For brevity, I won't repeat the quotes
from the GPLv2 preamble, that I just included in the message I sent to
Lennart Sorensen in this same thread.  Can you justify how you came to
the conclusion (if you did) that TiVo is abiding by the spirit of the
license?

>> Keeping on making false claims about the license drafts can be one of
>> two things: misunderstandings, out of ambiguity in the text or
>> preconceptions, or ill intentions.  I'd rather believe it's the
>> former.

> No, it was not the former.

Wow, I didn't see that coming.  Public admission of ill intentions?
;-) :-D :-P

> And I think the whole "the kernel developers misunderstand the
> license" crap that the FSF was saying (several times) was very
> trying to confuse the issue: the FSF knew damn well which part of
> the license was obnoxious, they just tried to confuse the issue by
> pointing to *another* part of the license.

Let me see if I got this right.  There was a section entitled
"3. Digital Restrictions Management" in GPLv3dd1.  Are you saying
that, when people complained about the DRM clause, they actually meant
the provisions in "1. Source Code", that established a requirement to
include the source code corresponding to functional signatures, namely
the signing keys, as part of the corresponding source code?

> And you're just parrotting their idiotic line.

Please watch your tone.  If you find offense at the allegedly
condescending tone in which the FSF says "misunderstanding", how do
you expect me and the FSF to take this?

It is also odd that you claim the right to be entitled to your own
opinion and reading about stuff, while denying myself the same right.
Please don't do that.  I have a mind of my own, and the fact that I
reach similar conclusions doesn't make me a parrot.  Even more so when
I actually have some influence on those conclusions.

>> Now, of course you can look at the licenses and decide that you never
>> agreed with the spirit of the GPL in the first place, and that GPLv2
>> models better your intentions than GPLv3.

> And this is again the same *disease*. You claim that I "misunderstood" the 
> "spirit of the GPL".

> Dammit, the GPL is a license. I understand it quite well. Probably better 
> than most. The fact that the FSF then noticed that there were *other* 
> things that they wanted to do, and that were *not* covered by the GPLv2, 
> does *not* mean that they can claim that others "misunderstood" the 
> license.

> I understood it perfectly fine, and it fit my needs. So tell me: who is 
> the more confused one: the one who chose the license fifteen years ago, 
> and realized what it means legally, and still stands behind it? I don't 
> think so.

You are definitely confused.  You're talking about the legal terms,
while I'm talking about the spirit.  The legal terms tried to reflect
the spirit as best as they could, but they left some holes.  Some
people found them and started exploiting them.

Sure, if you want to leave those holes unplugged in your code, that's
your decision.  I don't doubt that the GPLv2 legal terms fit the bill
for you.  I think GPLv3 would do even better in this regard.  But none
of this is about the spirit of the GPL.  Claiming GPLv3 changes the
spirit is totally missing the point of what the spirit amounts to.
The spirit is described in the preamble, it's not the legal terms.

> The beauty of the GPLv2 is exactly that it's a "tit-for-tat"
> license,

Ok, let's explore this argument.  In what sense is it tit-for-tat?
What is tit-for-tat about it?  What is the payback an author who
releases software under the GPL can legitimately expect to get?

>> In fact, the spirit has always been described in its preamble, and it 
>> didn't change at all: it's all about respecting others' freedoms.

> That's a lot of bullshit. You are apparently the grand poobah, and can 
> decide _which_ freedoms and for _what_ others' that matter.

The freedoms I'm talking about are very clearly described in the
spirit (preamble) of the license you chose for your project.  Go look
at the preamble one more time, "grand poobah" ;-)

  [...] the GNU General Public License is intended to guarantee your
  freedom to share and change free software--to make sure the software
  is free for all its users [...]

  [...] Our General Public Licenses are designed to make sure that you
  have the freedom to distribute copies of free software (and charge
  for this service if you wish), that you receive source code or can
  get it if you want it, that you can change the software or use
  pieces of it in new free programs; and that you know you can do
  these things.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Linus Torvalds]

On Wed, 13 Jun 2007, Alexandre Oliva wrote:
> 
> Let me see if I got this right.  There was a section entitled
> "3. Digital Restrictions Management" in GPLv3dd1.  Are you saying
> that, when people complained about the DRM clause, they actually meant
> the provisions in "1. Source Code"

Yes. I said that multiple times. It was obvious. But people didn't listen.

It's now in Section 7, or whatever.

The section 3 never mattered.

		Linus

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Jörn Engel]

On Wed, 13 June 2007 14:33:07 -0700, Linus Torvalds wrote:
> 
> The beauty of the GPLv2 is exactly that it's a "tit-for-tat" license, and 
> you can use it without having to drink the kool-aid.

One could even add that "tit-for-tat" appears to be the best strategy
in game theory for continuous runs of the prisoners dilemma.  At times I
wonder why game theory isn't taught in schools yet - it might shorten
discussions like these.

Jörn

-- 
All art is but imitation of nature.
-- Lucius Annaeus Seneca

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Wednesday 13 June 2007 19:35:41 Jörn Engel wrote:
> On Wed, 13 June 2007 14:33:07 -0700, Linus Torvalds wrote:
> > The beauty of the GPLv2 is exactly that it's a "tit-for-tat" license, and
> > you can use it without having to drink the kool-aid.
>
> One could even add that "tit-for-tat" appears to be the best strategy
> in game theory for continuous runs of the prisoners dilemma.  At times I
> wonder why game theory isn't taught in schools yet - it might shorten
> discussions like these.
>
> Jörn

With the sheer amount of sheeple[1] in the world (and on this list), I doubt 
anything could make these discussions any shorter.

(While I hate thinking that sheeple are on this list, it is an unavoidable 
fact. (I had hoped I wouldn't find any sheeple here, as my favorite theory is 
that they are all "Fundamentalist Christians" like the "Creationist" fools))

DRH
1: Sheeple (n): People that act like sheep - ie: they cannot think or form 
opinions for themselves and always look to someone else for their thoughts 
and parrot the opinions of some "trusted" figure.

-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Neil Brown]

On Wednesday June 13, dhazelton@enter.net wrote:
> 1: Sheeple (n): People that act like sheep - ie: they cannot think or form 
> opinions for themselves and always look to someone else for their thoughts 
> and parrot the opinions of some "trusted" figure.

I recommend that you avoid definitions like this.  Using them simply
makes you appear to have a very poor understanding of your fellow
humans.

"cannot" and "always" are absolutes that never apply (well, hardly
ever).

In my experience, most people do think for themselves and do form
opinions about areas where they have an interest/ability, and tend to
follow trusted others in areas where they have less interest or
ability.

The problem that I think you see is particularly the "parrot the
opinions" bit.  When a person tries to argue a case based largely on
the opinion of someone else with little personal understanding, they
are in very risky territory.   This is akin to 'fundamentalism' that
you also mentioned in your post.

Accusing people of arguing opinions that are not their own may well be
appropriate. Accusing them of not be able to think is not.

NeilBrown

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Wednesday 13 June 2007 20:14:47 Neil Brown wrote:
> On Wednesday June 13, dhazelton@enter.net wrote:
> > 1: Sheeple (n): People that act like sheep - ie: they cannot think or
> > form opinions for themselves and always look to someone else for their
> > thoughts and parrot the opinions of some "trusted" figure.
>
> I recommend that you avoid definitions like this.  Using them simply
> makes you appear to have a very poor understanding of your fellow
> humans.
>
> "cannot" and "always" are absolutes that never apply (well, hardly
> ever).

In this case I gave a very hasty definition of the term, though it does fit 
very well. 

> In my experience, most people do think for themselves and do form
> opinions about areas where they have an interest/ability, and tend to
> follow trusted others in areas where they have less interest or
> ability.

Agreed. But those *aren't* "Sheeple".

> The problem that I think you see is particularly the "parrot the
> opinions" bit.  When a person tries to argue a case based largely on
> the opinion of someone else with little personal understanding, they
> are in very risky territory.   This is akin to 'fundamentalism' that
> you also mentioned in your post.
>
> Accusing people of arguing opinions that are not their own may well be
> appropriate. Accusing them of not be able to think is not.

I never accused any specific person of either. I was making an observation 
about the general nature of humanity as I've observed it. A better definition 
of "Sheeple" would be "Someone who continually parrots some "trusted" figures 
opinion even when presented with proof that that opinion is in error. They 
also, generally, do not "think deeply" on any topic."

And while I won't apologize to anyone that may have felt the "Sheeple" remark 
was directed at them, I will apologize for not taking more time to choose a 
better definition for the term originally.

DRH

> NeilBrown
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/



-- 
Dialup is like pissing through a pipette. Slow and excruciatingly painful.

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 13, 2007, Neil Brown <neilb@suse.de> wrote:

> Accusing people of arguing opinions that are not their own may well be
> appropriate. Accusing them of not be able to think is not.

I agree with the latter, but the former is seldom appropriate.  Any
accusation ought to be provable, and it's nearly impossible to prove
that an opinion held by someone is not his own.  People quite often
arrive at similar opinions independently.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 13, 2007, Jörn Engel <joern@logfs.org> wrote:

> On Wed, 13 June 2007 14:33:07 -0700, Linus Torvalds wrote:
>> 
>> The beauty of the GPLv2 is exactly that it's a "tit-for-tat" license, and 
>> you can use it without having to drink the kool-aid.

> One could even add that "tit-for-tat" appears to be the best strategy
> in game theory for continuous runs of the prisoners dilemma.

It is, indeed.

Now the remaining piece of the proof is to show that the GPLv2 is
tit-for-tat.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Satyam Sharma]

[ Gmail did horrible things to the original post by giving it a base64
content transfer encoding, so majordomo@vger dropped it. It's just
an off-topic digression, but I cared enough to resend anyway, fwiw. ]

On 6/14/07, Jörn Engel <joern@logfs.org> wrote:
> On Wed, 13 June 2007 14:33:07 -0700, Linus Torvalds wrote:
> >
> > The beauty of the GPLv2 is exactly that it's a "tit-for-tat" license, and
> > you can use it without having to drink the kool-aid.
>
> One could even add that "tit-for-tat" appears to be the best strategy
> in game theory for continuous runs of the prisoners dilemma.

Tit-for-tat is the best *deterministic* strategy when playing
iterated prisoner's dilemma. But note that "deterministic" and
"rational" are not adjectives that go well with "humans", and
most real-world (social) situations are noisy environments --
miscommunication and misunderstandings are the usual noise.
A double-D noise perceived by any player would throw a
tit-for-tat-playing couple into a perennial spiral of D's, for example,
which is clearly not a Pareto-efficient solution for either.

> At times I
> wonder why game theory isn't taught in schools yet - it might shorten
> discussions like these.

Yes, and no.

Yes - for teaching game theory (and its social relevance) in schools;
and add behavioral economics to this list :-)

No - it doesn't shorten discussions, however. And it shouldn't either.
Human / social situations are complex, Jörn; tit-for-tat can win
computer contests, for example, but it's not a behaviour one person
would find as entirely agreeable in another.

Satyam

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[H. Peter Anvin]

Tarkan Erimer wrote:
> And maybe another questions should be : How long a copyright owner can
> hold the copyright, if died or lost for sometime ? if died, the
> copyright still should be valid or not ? If lost, what the law orders at
> this point for copyright holding ?

In most countries, copyright lasts for 70 years after the death of the
author.  Either way, it is most definitely valid.

	-hpa

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Tarkan Erimer]

Hi Neil,

Neil Brown wrote:
> On Saturday June 9, tarkan@netone.net.tr wrote:
>   
>> Hi,
>>
>> As we know the forthcoming GPL V3 will be not compatible with the GPL V2 
>> and Linux Kernel is GPL V2 only.
>> So, another point is, which is previously mentioned by Linus and others, 
>> that if it is decided to upgrade the Linux Kernel's License to GPL V3, 
>> it is needed the permission of all the maintainers permission who 
>> contributed to the Linux Kernel and there are a lot of lost or dead 
>> maintainers. Which makes it impossible to get all the maintainers' 
>> permission.
>>     
>
> You don't need the permission of maintainers.  You need the permission
> of copyright owners.  The two groups overlap, but are not the same.
> Dead people cannot own anything, even copyright.  Their estate
> probably can.  I don't think it is theoretically impossible to get
> everyone's permission, though it may be quite close to practically
> impossible. 
>
>   
So, does it mean we can change the license of the dead people's code ?

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Neil Brown]

On Sunday June 10, tarkan@netone.net.tr wrote:
> Hi Neil,
> 
> Neil Brown wrote:
> > On Saturday June 9, tarkan@netone.net.tr wrote:
> >   
> >> Hi,
> >>
> >> As we know the forthcoming GPL V3 will be not compatible with the GPL V2 
> >> and Linux Kernel is GPL V2 only.
> >> So, another point is, which is previously mentioned by Linus and others, 
> >> that if it is decided to upgrade the Linux Kernel's License to GPL V3, 
> >> it is needed the permission of all the maintainers permission who 
> >> contributed to the Linux Kernel and there are a lot of lost or dead 
> >> maintainers. Which makes it impossible to get all the maintainers' 
> >> permission.
> >>     
> >
> > You don't need the permission of maintainers.  You need the permission
> > of copyright owners.  The two groups overlap, but are not the same.
> > Dead people cannot own anything, even copyright.  Their estate
> > probably can.  I don't think it is theoretically impossible to get
> > everyone's permission, though it may be quite close to practically
> > impossible. 
> >
> >   
> So, does it mean we can change the license of the dead people's code ?
> 

I presume the heirs of the dead people could change the license.  And
if they have no heir, then there is no-one to sue for breach of
copyright, so I assume the copyright lapses.

And I wouldn't be surprised if there were some legal precedent that
allowed for some process whereby we could make a "best effort" to
contact copyright holders (including registered paper letters and
entries in the "Public Notices" section of major newspapers) and if
no-one stepped forward to claim copyright in a reasonable period of
time we could assume that the copyright had lapsed.  But you would
need to ask a lawyer, and it would be different in different
countries.

But I think this is largely academic.  You only need a fairly small
number of fairly significant contributors to say "no" and the rest of
the process would be pointless.  And at last count, the number of
kernel people who were not keen on GPLv3 was fairly high.  Of course
no-one knows for certain yet what the final GPLv3 will be, and maybe
lots of people would change their mind when it comes out.

There would certainly be value in a straw-pole once GPLv3 was out and
had been discussed for a while - to see if a license change to GPLv3
would be accepted by a majority of current developers.  Doing that
would at least provide a clear statistic to point people at.

NeilBrown

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Jiri Kosina]

On Sun, 10 Jun 2007, Neil Brown wrote:

> I presume the heirs of the dead people could change the license.  And if 
> they have no heir, then there is no-one to sue for breach of copyright, 
> so I assume the copyright lapses.

In most of the law systems out there the copyright stays valid for 70 
years (or so) after the holder's death.

-- 
Jiri Kosina

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Sunday 10 June 2007 08:45:41 Jiri Kosina wrote:
> On Sun, 10 Jun 2007, Neil Brown wrote:
> > I presume the heirs of the dead people could change the license.  And if
> > they have no heir, then there is no-one to sue for breach of copyright,
> > so I assume the copyright lapses.
>
> In most of the law systems out there the copyright stays valid for 70
> years (or so) after the holder's death.

I'm almost certain that it is the same in the US, not the death+90 previously 
stated. (I've read the copyright laws a number of times to deal with some 
involved conversations) In some of the writings tied the change that made it 
death+70 its stated that said change was made to "bring US laws in line with 
Europe and most of the rest of the world" (paraphrase - I didn't bother going 
and digging out the page again). It's a relatively common belief - and, IIRC, 
was even brought before the US Supreme Court - that the copyrights length was 
changed to give Disney longer protection, if just because a lot of Disney's 
copyrights were going to expire and the change was applied retroactively. 
(And, unsurprisingly, the suit was shot down - 70 years is still a "limited" 
period. However, IIRC, there was some noted concern by the Supreme Court that 
the US Congress would exploit the legal loophole and just keep extending the 
copyright period retroactively to make it, effectively, never-ending. That 
scenario, while not *technically* in violation of the language of the US 
Constitution (which grants the US Congress the power to set the length of 
copyrights) would violate the spirit of it)

DRH

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Michael Gerdau]

[-- Attachment #1: Type: text/plain, Size: 1369 bytes --]

> In most of the law systems out there the copyright stays valid for 70 
> years (or so) after the holder's death.

[I don't want to appear picky but IMO it is important to be as precise
as one could possibly be, therefor...]

That is not quite correct. At least for the german law system and
presumably for most if not all similar systems the corrrect stmts is:

In most of the law systems out there the copyright stays valid for 70 
years (or so) after initial publication of the copyrighted work. For
work whose publication date can't be determined (*) (e.g. unpublished work)
the creators death is assumed to be the publication date for the sake
of determining the  aforementioned 70 years protection period.

(*) "can't be determined" does not apply for published work with an
insecure initial publication date. In those cases the earliest documented
publication (i.e. prooveable in court) is considered the initial publication
.date

Note that the original phrase could last indefinitely by simply moving
the copyright from person to person.

Best,
Michael
-- 
 Technosis GmbH, Geschäftsführer: Michael Gerdau, Tobias Dittmar
 Sitz Hamburg; HRB 89145 Amtsgericht Hamburg
 Vote against SPAM - see http://www.politik-digital.de/spam/
 Michael Gerdau       email: mgd@technosis.de
 GPG-keys available on request or at public keyserver

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Adrian Bunk]

On Sun, Jun 10, 2007 at 04:23:20PM +0200, Michael Gerdau wrote:
> > In most of the law systems out there the copyright stays valid for 70 
> > years (or so) after the holder's death.
> 
> [I don't want to appear picky but IMO it is important to be as precise
> as one could possibly be, therefor...]
> 
> That is not quite correct. At least for the german law system and
> presumably for most if not all similar systems the corrrect stmts is:
> 
> In most of the law systems out there the copyright stays valid for 70 
> years (or so) after initial publication of the copyrighted work.
>...

Please read the German "Gesetz über Urheberrecht und verwandte 
Schutzrechte" before making completely false claims - in Germany it's
70 years after the death of the author.

> Best,
> Michael

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Michael Gerdau]

[-- Attachment #1: Type: text/plain, Size: 830 bytes --]

> Please read the German "Gesetz über Urheberrecht und verwandte 
> Schutzrechte" before making completely false claims - in Germany it's
> 70 years after the death of the author.

Oups, I had been mixing up named published works and anonymously
published works.

My previous claim is valid for anonymously published work only and
thus holds no relevance for the situation discussed. I apologize for
having raised a false claim.

The copyright for works with a known is valid for 70 years after
the authors death.

Sorry,
Michael
-- 
 Technosis GmbH, Geschäftsführer: Michael Gerdau, Tobias Dittmar
 Sitz Hamburg; HRB 89145 Amtsgericht Hamburg
 Vote against SPAM - see http://www.politik-digital.de/spam/
 Michael Gerdau       email: mgd@technosis.de
 GPG-keys available on request or at public keyserver

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Adrian Bunk]

On Sun, Jun 10, 2007 at 07:36:39PM +1000, Neil Brown wrote:
>...
> And I wouldn't be surprised if there were some legal precedent that
> allowed for some process whereby we could make a "best effort" to
> contact copyright holders (including registered paper letters and
> entries in the "Public Notices" section of major newspapers) and if
> no-one stepped forward to claim copyright in a reasonable period of
> time we could assume that the copyright had lapsed.  But you would
> need to ask a lawyer, and it would be different in different
> countries.
>...

A legal precedent valid in all jurisdictions?

Harald suceessfully takes legal actions against people violating his 
copyright on the Linux kernel under the terms of the GPL in Germany at 
German courts based on German laws.

If someone finds any legal precedent in Finland or the USA or Russia 
that some copyright would have lapsed for some reason, would this have 
any legal effect in Germany?

> NeilBrown

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Michael Gerdau]

[-- Attachment #1: Type: text/plain, Size: 1334 bytes --]

[legal precedence to force waiving copyright> 

> A legal precedent valid in all jurisdictions?
> 
> Harald suceessfully takes legal actions against people violating his 
> copyright on the Linux kernel under the terms of the GPL in Germany at 
> German courts based on German laws.
> 
> If someone finds any legal precedent in Finland or the USA or Russia 
> that some copyright would have lapsed for some reason, would this have 
> any legal effect in Germany?

That is difficult but AFAIK it work in the "increase protection"
direction but not the other way around, at least in germany.

For example even though the copyright on Micky Mouse is no longer
valid by german law it is still enforcible under german law because
the german jurisdiction does acknowledge the (recently extended)
longer such period in the US.

Apart from that I'm sure you'd find some state "legally" waiving a
copyright on whatever. If that would start an avalanche of similar
terminations everywhere else I'd be utterly surprised.

Best,
Michael
-- 
 Technosis GmbH, Geschäftsführer: Michael Gerdau, Tobias Dittmar
 Sitz Hamburg; HRB 89145 Amtsgericht Hamburg
 Vote against SPAM - see http://www.politik-digital.de/spam/
 Michael Gerdau       email: mgd@technosis.de
 GPG-keys available on request or at public keyserver

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alan Cox]

> But I think this is largely academic.  You only need a fairly small
> number of fairly significant contributors to say "no" and the rest of
> the process would be pointless.  And at last count, the number of
> kernel people who were not keen on GPLv3 was fairly high.  Of course
> no-one knows for certain yet what the final GPLv3 will be, and maybe
> lots of people would change their mind when it comes out.

You can take a fair bet someone will say no, or much more likely they or
whoever inherited their copyright will say $50,000

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Daniel Hazelton]

On Sunday 10 June 2007 09:40:23 Alan Cox wrote:
> > But I think this is largely academic.  You only need a fairly small
> > number of fairly significant contributors to say "no" and the rest of
> > the process would be pointless.  And at last count, the number of
> > kernel people who were not keen on GPLv3 was fairly high.  Of course
> > no-one knows for certain yet what the final GPLv3 will be, and maybe
> > lots of people would change their mind when it comes out.
>
> You can take a fair bet someone will say no, or much more likely they or
> whoever inherited their copyright will say $50,000

I seeds shades of Merkey there :P

Seriously, though, this was all settled a long time ago. Linus said "While 
individual parts of the kernel *could* be licensed [with another license] the 
kernel as a whole is strictly GPLv2" (I've tried to get it right, but my 
memory isn't as good as it used to be when it comes to useful quotes like 
that)

DRH

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Tim Post]

On Sun, 2007-06-10 at 19:36 +1000, Neil Brown wrote:

> I presume the heirs of the dead people could change the license.  And
> if they have no heir, then there is no-one to sue for breach of
> copyright, so I assume the copyright lapses.
> 
> And I wouldn't be surprised if there were some legal precedent that
> allowed for some process whereby we could make a "best effort" to
> contact copyright holders (including registered paper letters and
> entries in the "Public Notices" section of major newspapers) and if
> no-one stepped forward to claim copyright in a reasonable period of
> time we could assume that the copyright had lapsed.  But you would
> need to ask a lawyer, and it would be different in different
> countries.

I've done some research on this and from what I can tell you are
correct. There is some sort of "Due Diligence" law that you have to
satisfy that is slightly different from country to country. 

>From what I can tell the process is similar to changing your name, an ad
in the paper would be o.k. in most places. Since the intent of the
copyright holder to make their work free is clear, its pretty clear cut.

It was hard to find specifics, because not many people worry about it if
they know the copyright holder to be existentially challenged. I didn't
find many cases of people even bothering with the formality.

I say existentially challenged because you don't need to be dead to
vanish. I also read it as ".. a person of questionable existence." as
well as "A null (or moot) party".

> But I think this is largely academic.  You only need a fairly small
> number of fairly significant contributors to say "no" and the rest of
> the process would be pointless.  And at last count, the number of
> kernel people who were not keen on GPLv3 was fairly high.  Of course
> no-one knows for certain yet what the final GPLv3 will be, and maybe
> lots of people would change their mind when it comes out.

I think its good that people have their own viewpoints instead of just
watching the Linus - RMS tennis match as the points of the license get
debated and assuming the views of whoever 'wins' in their eyes. Neither
person is going to be around forever. 

> There would certainly be value in a straw-pole once GPLv3 was out and
> had been discussed for a while - to see if a license change to GPLv3
> would be accepted by a majority of current developers.  Doing that
> would at least provide a clear statistic to point people at.

Unless you are distributing the Linux kernel in a whole or parts of it
in some way you don't even HAVE to accept the terms of the GPL2, or GPL3
for that matter. This doesn't really concern me as an end user, but it
should interest me (and does).

I'm more worried about someone finding a way to leverage existing or
future patents against users of the Linux kernel, or people who provide
the use of a computer using the Linux kernel as a service.

I'm not *overly* concerned about that. Anyone trying to get people to
pay for the right to use the Linux kernel would put themselves in clear
and certain danger of becoming ' a person of questionable existence '
themselves.

Best,
--Tim

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Tim Post]

On Sun, 2007-06-10 at 11:49 +0300, Tarkan Erimer wrote:
>  
> So, does it mean we can change the license of the dead people's code ?
> 

Please realize that one doesn't need to be dead to become
uncommunicative incapacitated or vanish. The only need to be somewhere
other than where they were without updating anyone. 

Here is a very humorous, but sort of scary theoretical :

"Linus was so disturbed by the code in a submitted patch that he had a
nervous breakdown and spent the next 30 years in a padded room.
Unfortunately, no provisions were left to determine what happens to his
copyrights should he become incapacitated."

That is a LOT different than 

"Linus passed away at age 397 today, and left an appropriate will so his
work remains useful for those enjoying it."

If you are a substantial contributor, address this in a will should you
die and limited power of attorney to someone that you trust should you
fall off the face of the planet. Just like you would see to it that your
assets were properly dispersed if something happened to you.

If your unfortunate enough to not have anyone in your life that you can
trust, you can grant power of attorney to any reputable charitable
organization that you feel mirrors your beliefs and ideals. 

Prevention is so much better than a cure and quite cheap. 

Best,
--Tim

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Pavel Machek]

Hi!

> > So, does it mean we can change the license of the dead people's code ?
> > 
> 
> Please realize that one doesn't need to be dead to become
> uncommunicative incapacitated or vanish. The only need to be somewhere
> other than where they were without updating anyone. 
> 
> Here is a very humorous, but sort of scary theoretical :
> 
> "Linus was so disturbed by the code in a submitted patch that he had a
> nervous breakdown and spent the next 30 years in a padded room.
> Unfortunately, no provisions were left to determine what happens to his
> copyrights should he become incapacitated."

License is only promise not to sue. If Linus is safely in a padded
room, I'll happily relicense Linux under GPLv17, knowing he is not
going to sue me :-).

(And actually, if he's in a padded room, there's probably someone who
can act in his name. At least czech law works like that.)
							Pavel
-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

RE: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Schwartz]

> But; if the Linux kernel should Dual-Licensed (GPL V2 and GPL V3), it
> will allow us the both worlds' fruits like code exchanging from other
> Open Source Projects (OpenSolaris etc.) that is compatible with GPL V3
> and not with GPL V2 and of course the opposite is applicable,too.

That is a misleading claim. While being dual-licensed would make it either
for other projects to adopt Linux code, it would have three downsides:

1) If Linux code were adopted into other projects that were not
dual-licensed, changes could not be imported back into Linux unless the
changes were dual-licensed which is not likely when the contributions are
made to a project that's not dual-licensed.

2) Linux could no longer take code from other projects that are GPL v2
licensed unless it could obtain them under a dual license.

And, last and probably most serious:

3) Linux derivatives could be available with just a GPL v3 license and no
GPL v2. license if the derivers wanted things that way.

DS

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Tarkan Erimer]

Hi David,

David Schwartz wrote:
>> But; if the Linux kernel should Dual-Licensed (GPL V2 and GPL V3), it
>> will allow us the both worlds' fruits like code exchanging from other
>> Open Source Projects (OpenSolaris etc.) that is compatible with GPL V3
>> and not with GPL V2 and of course the opposite is applicable,too.
>>     
>
> That is a misleading claim. While being dual-licensed would make it either
> for other projects to adopt Linux code, it would have three downsides:
>
> 1) If Linux code were adopted into other projects that were not
> dual-licensed, changes could not be imported back into Linux unless the
> changes were dual-licensed which is not likely when the contributions are
> made to a project that's not dual-licensed.
>
> 2) Linux could no longer take code from other projects that are GPL v2
> licensed unless it could obtain them under a dual license.
>
> And, last and probably most serious:
>
> 3) Linux derivatives could be available with just a GPL v3 license and no
> GPL v2. license if the derivers wanted things that way.
>
>   

Thanks for the corrections ;-) The whole picture is more clear now for 
me :-)
BTW,I found a really interesting blog entry about which code in Linux 
Kernel is using which version of GPL :

http://6thsenseless.blogspot.com/2007/02/how-much-linux-kernel-code-is-gpl-2.html

The work done on a Linux 2.6.20. The result is quite interesting. 
Because almost half  (Around %60 of the code licensed under "GPLv2 Only" 
and the rest is "GPLv2 or above","GPL-Version not specified,others that 
have not stated which and what version of License has been used) of the 
code is "GPLv2 or above" licensed. And also stated in the article that 
some of the codes should be "Dual Licensed" not the whole Linux kernel 
needed to be "Dual Licensed". So,if it is really like this, maybe we can 
make,for example: "File system related Codes", "Dual Licensed" and it 
will allow us to port ZFS from OpenSolaris requested by a lot of people 
or other things maybe ?

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Jan Engelhardt]

On Jun 10 2007 11:37, Tarkan Erimer wrote:
>
> Thanks for the corrections ;-) The whole picture is more clear now for me :-)
> BTW,I found a really interesting blog entry about which code in Linux Kernel is
> using which version of GPL :
>
> http://6thsenseless.blogspot.com/2007/02/how-much-linux-kernel-code-is-gpl-2.html

You've got to take MODULE_LICENSE() into account. There is

	MODULE_LICENSE("GPL");
	MODULE_LICENSE("GPL v2");
	MODULE_LICENSE("GPL and additional rights");
	MODULE_LICENSE("Dual BSD/GPL");
	MODULE_LICENSE("Dual MIT/GPL");
	MODULE_LICENSE("Dual MPL/GPL");

I think it's time to set things right, making

	* MODULE_LICENSE the authoritative place for the license
	  (also makes it easier to grep for)

	* sync up license into MODULE_LICENSE



	Jan
-- 

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Simon Arlott]

On 10/06/07 09:37, Tarkan Erimer wrote:
> BTW,I found a really interesting blog entry about which code in Linux
> Kernel is using which version of GPL :
> 
> http://6thsenseless.blogspot.com/2007/02/how-much-linux-kernel-code-is-gpl-2.html
> 
> 
> The work done on a Linux 2.6.20. The result is quite interesting.
> Because almost half  (Around %60 of the code licensed under "GPLv2 Only"
> and the rest is "GPLv2 or above","GPL-Version not specified,others that
> have not stated which and what version of License has been used) of the
> code is "GPLv2 or above" licensed. And also stated in the article that
> some of the codes should be "Dual Licensed" not the whole Linux kernel
> needed to be "Dual Licensed". So,if it is really like this, maybe we can
> make,for example: "File system related Codes", "Dual Licensed" and it
> will allow us to port ZFS from OpenSolaris requested by a lot of people
> or other things maybe ?

Once code obtained under the GPLv2 only licence that the kernel is released 
under is modified and submitted back to Linus for inclusion, that code 
would become GPLv2 only - only the original would be BSD/LGPL/GPLv2+ and 
only separate changes to the original could continue to be available under 
dual licence. Since most files will have been modified at various stages 
in Linux's development when major internal changes occur, surely practically 
everything is now GPLv2 only?


> So, does it mean we can change the license of the dead people's code ? 

If you can contact whoever currently owns the copyright, they can release 
it under another licence... however this is no good because the derivative 
work would be GPLv2 only. Perhaps if you got *everyone* at all stages of 
development (including code that has been removed if existing code is a 
derivative work of it) to agree - then it could work.

It only takes one person's code, uncooperative or not contactable, to 
prevent a change to the licence, so there's not much point in trying 
unless you intend to start replacing such code.


On 10/06/07 10:03, Jan Engelhardt wrote:
> You've got to take MODULE_LICENSE() into account. There is
> 
> 	MODULE_LICENSE("GPL");
> 	MODULE_LICENSE("GPL v2");
> 	MODULE_LICENSE("GPL and additional rights");
> 	MODULE_LICENSE("Dual BSD/GPL");
> 	MODULE_LICENSE("Dual MIT/GPL");
> 	MODULE_LICENSE("Dual MPL/GPL");

Surely that doesn't work since the entire Linux kernel is (and can only be) 
released as GPLv2? Wouldn't anyone making changes to those files need to 
obtain a copy under the other licence and explicitly release it under both 
licenses in order to maintain that?

-- 
Simon Arlott

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Jan Engelhardt]

On Jun 10 2007 10:17, Simon Arlott wrote:
>On 10/06/07 09:37, Tarkan Erimer wrote:
>> BTW,I found a really interesting blog entry about which code in Linux
>> Kernel is using which version of GPL :
>> 
>> http://6thsenseless.blogspot.com/2007/02/how-much-linux-kernel-code-is-gpl-2.html
>> 
>> You've got to take MODULE_LICENSE() into account. There is
>> 
>> 	MODULE_LICENSE("GPL");
>> 	MODULE_LICENSE("GPL v2");
>> 	MODULE_LICENSE("GPL and additional rights");
>> 	MODULE_LICENSE("Dual BSD/GPL");
>> 	MODULE_LICENSE("Dual MIT/GPL");
>> 	MODULE_LICENSE("Dual MPL/GPL");
>
>Surely that doesn't work since the entire Linux kernel is (and can only be) 
>released as GPLv2? Wouldn't anyone making changes to those files need to 
>obtain a copy under the other licence and explicitly release it under both 
>licenses in order to maintain that?

http://www.eweek.com/article2/0,1759,1915720,00.asp
has the answer. Quoting Linus:

"If you want to license a program under any later version of the GPL, you have
to state so explicitly. Linux never did."

Hence, unless there's a "GPL 2 or later", all the "unspecified GPL" files
are GPL2 only.


	Jan
-- 

RE: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Schwartz]

> http://www.eweek.com/article2/0,1759,1915720,00.asp
> has the answer. Quoting Linus:
>
> "If you want to license a program under any later version of the
> GPL, you have
> to state so explicitly. Linux never did."
>
> Hence, unless there's a "GPL 2 or later", all the "unspecified GPL" files
> are GPL2 only.

The GPL states the default position:

"If the Program does not specify a version number of this License, you may
choose any version ever published by the Free Software Foundation."

Leaving the question of whether Linus's comment at the top of the license
changes the default:

"Also note that the only valid version of the GPL as far as the kernel is
concerned is _this_ particular version of the license (ie v2, not v2.2 or
v3.x or whatever), unless explicitly otherwise stated."

So we have dueling defaults. The GPL says the default is any version. Linus'
statement at the top of the GPL says the default is v2 only. It's not clear,
at least to me, that there is any clear reason why one should win out over
the other.

DS

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Rob Landley]

On Sunday 10 June 2007 16:42:35 David Schwartz wrote:
> > http://www.eweek.com/article2/0,1759,1915720,00.asp
> > has the answer. Quoting Linus:
> >
> > "If you want to license a program under any later version of the
> > GPL, you have
> > to state so explicitly. Linux never did."
> >
> > Hence, unless there's a "GPL 2 or later", all the "unspecified GPL" files
> > are GPL2 only.
>
> The GPL states the default position:
>
> "If the Program does not specify a version number of this License, you may
> choose any version ever published by the Free Software Foundation."
>
> Leaving the question of whether Linus's comment at the top of the license
> changes the default:
>
> "Also note that the only valid version of the GPL as far as the kernel is
> concerned is _this_ particular version of the license (ie v2, not v2.2 or
> v3.x or whatever), unless explicitly otherwise stated."
>
> So we have dueling defaults. The GPL says the default is any version.
> Linus' statement at the top of the GPL says the default is v2 only. It's
> not clear, at least to me, that there is any clear reason why one should
> win out over the other.

Except that Linux included a copy of the license it was distributed under in 
the tarball, which was GPLv2.  There's a case to be made that this DOES count 
as "selection of the license version".  It's hard to get more explicit than 
including the complete license text.

On top of that, Linus clarified his position back in 2000:
http://www.uwsg.iu.edu/hypermail/linux/kernel/0009.1/0096.html

This means that not only are his on contributions so licensed, but there's a 
strong case to be made that everything he merged since then is explicitly 
GPLv2 only when sourced from the Linux kernel.  As maintainer there's a good 
argument that he has a compliation copyright on the kernels he releases.  He 
was certainly accepting patches under GPLv2 only, and merging them into a 
GPLv2 only work.  Sourcing ANY code from that work and declaring it non-GPLv2 
is really fishy; much better to find a clean upstream source, such as the 
original authors.

Of course the FSF would very, very much like this not to have been the case, 
and they've been trying to wish it away ever since...

Rob

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Al Boldi]

Alexandre Oliva wrote:
> On Jun 17, 2007, "Gabor Czigola" <czigola@gmail.com> wrote:
> > I wonder why the linux kernel development community couldn't propose
> > an own GPL draft (say v2.2) that is "as free as v2" and that includes
> > some ideas (from v3) that are considered as good (free, innovative, in
> > the spirit of whatever etc.) by the majority of the kernel developers.
>
> For one, because the text of the GPL is copyrighted by the FSF, and
> licensed without permission for modification.  And that's as it should
> be, you don't want others to modify the terms of the license you chose
> for your code, do you?

Wow!

Under what circumstances would it be possible to receive permission for 
modification?


Thanks for being GPL!

--
Al

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Al Boldi]

Alexandre Oliva wrote:
> Consider this scenario: vendor tivoizes Linux in the device, and
> includes the corresponding sources only in a partition that is
> theoretically accessible using the shipped kernel, but that nothing in
> the software available in the machine will let you get to.  Further,
> sources (like everything else on disk) are encrypted, and you can only
> decrypt it with hardware crypto that is disabled if the boot loader
> doesn't find a correct signature for the boot partition, or maybe the
> signature is irrelevant, given that everything on disk is encrypted in
> such a way that, if you don't have the keys, you can't update the
> kernel properly anyway.  The vendor refuses to give customers other
> copies of the sources.  To add insult to the injury, the vendor
> configures the computer to set up the encrypted disk partition
> containing the sources as a swap device, such that the shared-secret
> key used to access that entire filesystem is overwritten upon the
> first boot, rendering the entire previous contents of the partition
> holding the source code into an incomprehensible stream of bits.
>
> Does anyone think this is permitted by the letter of GPLv2?

Yes.

> Is it in the spirit of GPLv2?

No, but that's besides the point.

You can only hold people responsible for the letter, lest there be chaos.

If there is a specific usage spirit you want to protect, then you must 
formulate it in letter.

> How are the sources passed on in this way going to benefit the user or the
> community?

They still have to provide the source by other GPL means of their choosing.

> Is this still desirable by the Linux developers?

Looks undesirable to me, but still valid.


Thanks!

--
Al

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 26, 2007, Al Boldi <a1426z@gawab.com> wrote:

>> Is it in the spirit of GPLv2?

> No, but that's besides the point.

Thanks for informing me about the point *I*'m trying to make ;-)

> You can only hold people responsible for the letter, lest there be chaos.

That's not *quite* how it works, but that's a general idea, yes.

>> How are the sources passed on in this way going to benefit the user or the
>> community?

> They still have to provide the source by other GPL means of their choosing.

This is contradictory.  You said the scenario I described was
permitted, and the scenario included the vendor's refusal to give
customers other copies of the sources.

Which is it?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Al Boldi]

Alexandre Oliva wrote:
> On Jun 26, 2007, Al Boldi <a1426z@gawab.com> wrote:
> >> Is it in the spirit of GPLv2?
> >
> > No, but that's besides the point.
>
> Thanks for informing me about the point *I*'m trying to make ;-)
>
> > You can only hold people responsible for the letter, lest there be
> > chaos.
>
> That's not *quite* how it works, but that's a general idea, yes.
>
> >> How are the sources passed on in this way going to benefit the user or
> >> the community?
> >
> > They still have to provide the source by other GPL means of their
> > choosing.
>
> This is contradictory.  You said the scenario I described was
> permitted, and the scenario included the vendor's refusal to give
> customers other copies of the sources.
>
> Which is it?

I read your scenario of the vendor not giving you the source to mean: not 
directly; i.e.  they could give you a third-party download link.


Thanks!

--
Al

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 26, 2007, Al Boldi <a1426z@gawab.com> wrote:

> I read your scenario of the vendor not giving you the source to mean: not 
> directly; i.e.  they could give you a third-party download link.

This has never been enough to comply with GPLv2.

FWIW, it is one of the improvements in GPLv3.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[david]

On Tue, 26 Jun 2007, Alexandre Oliva wrote:

> On Jun 26, 2007, Al Boldi <a1426z@gawab.com> wrote:
>
>> I read your scenario of the vendor not giving you the source to mean: not
>> directly; i.e.  they could give you a third-party download link.
>
> This has never been enough to comply with GPLv2.
>
> FWIW, it is one of the improvements in GPLv3.

either it's an improvement in the GPLv3 or it's a violation of GPLv2.

you can't say that the GPLv2 prohibits it _and_ it's an improvement in the 
GPLv3.

unless you are saying that the GPLv3 is saying that a third party link now 
_is_ sufficiant. this seems to be counter to what the FSF is claiming 
(with good reasoning. after all, you don't control the third party site, 
so it could change or go away and now the people who got the binaries from 
you can't get the sources)

David Lang

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 26, 2007, david@lang.hm wrote:

> unless you are saying that the GPLv3 is saying that a third party link
> now _is_ sufficiant.

Yup.  The improvement in GPLv3 is to relax the requirement of
providing source code in physical medium if you choose to not
distribute it along with the binaries.  It's recognizing that internet
access is no longer a barrier that could stop someone from obtaining
the sources they're entitled to.  Even someone who doesn't have
regular or fast internet access can hire a third party who does to
perform the download and record it.

I.e., with GPLv3, you *can* point at the sources you used, even in a
site that you don't control.

However, if the site takes the sources out, you're still responsible
for providing sources to those who received the sources from you from
that point on.  Or something like that, IANAL ;-)

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[david]

On Tue, 26 Jun 2007, Alexandre Oliva wrote:

> On Jun 26, 2007, david@lang.hm wrote:
>
>> unless you are saying that the GPLv3 is saying that a third party link
>> now _is_ sufficiant.
>
> Yup.  The improvement in GPLv3 is to relax the requirement of
> providing source code in physical medium if you choose to not
> distribute it along with the binaries.  It's recognizing that internet
> access is no longer a barrier that could stop someone from obtaining
> the sources they're entitled to.  Even someone who doesn't have
> regular or fast internet access can hire a third party who does to
> perform the download and record it.
>
> I.e., with GPLv3, you *can* point at the sources you used, even in a
> site that you don't control.
>
> However, if the site takes the sources out, you're still responsible
> for providing sources to those who received the sources from you from
> that point on.  Or something like that, IANAL ;-)

this sounds like a step backwards, you may not have the sources at that 
point if you were relying on the other site to host them.

and by the way, internet access never was a barrier that could stop 
someone from obtaining them, the only issue was you hosting the source vs 
someone else hosting the source.

David Lang

RE: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Schwartz]

david@lang.hm wrote:

> this sounds like a step backwards, you may not have the sources at that
> point if you were relying on the other site to host them.

You would then be violating the GPL, under any version. The GPL is quite
clear that being unable to comply with it means you do not get the benefits
it offers rather than excusing you from meeting its requirements.

You *MUST* have the source code in order to distribute it on request. You
cannot ship GPL'd works without offering source code just by arranging it
(deliberately or accidentally) so that you don't have the source code.

> and by the way, internet access never was a barrier that could stop
> someone from obtaining them, the only issue was you hosting the source vs
> someone else hosting the source.

The GPLv2 never specified one way or the other.

If you do allow someone else to host them, you are responsible for making
sure they remain available for at least three years from the last time you
used them as an offer. Should they stop distributing, you would be violating
the GPL. Nothing in the GPL says you can't rely on third parties for your
GPL compliance. Of course, this could be a very risky thing to do. However,
there is no GPL violation so long as they do in fact remain operational for
three years from the last time you distributed.

In fact, a third party is no more risky than any other setup. Any company
can go out of business within the three-year period after distribution.
There are many real-world cases where a third party having the source is
actually more likely to result in actual GPL compliance than the
distributor. (Consider a fly-by-night company selling Fedora binary
distributions burnt to CDROM on the stop for $1 on a street corner.)

One way to avoid this problem is to maintain your own web page that links to
the third party's download. You would have to host the sources yourself if
you couldn't make other arrangements at any point during the three year
period.

This is no different from any other case where the offer is not honored. If
the offer is not honored in a case where the GPL requires that it be, the
GPL is being violated.

DS

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 26, 2007, david@lang.hm wrote:

>> However, if the site takes the sources out, you're still responsible
>> for providing sources to those who received the sources from you from
>> that point on.  Or something like that, IANAL ;-)

> this sounds like a step backwards, you may not have the sources at
> that point if you were relying on the other site to host them.

You should have them.  This provision is not an excuse from your
obligations, it's just a pragmatic concession.

> and by the way, internet access never was a barrier that could stop
> someone from obtaining them

Back when GPLv2 was written, it really was.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

RE: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Schwartz]

Alexandre Oliva:

> On Jun 26, 2007, Al Boldi <a1426z@gawab.com> wrote:
>
> > I read your scenario of the vendor not giving you the source to
> > mean: not
> > directly; i.e.  they could give you a third-party download link.
>
> This has never been enough to comply with GPLv2.

A lot of people seem to say this, but I don't think it's true. Section 3b
says:

    Accompany it with a written offer, valid for at least three
    years, to give any third party, for a charge no more than your
    cost of physically performing source distribution, a complete
    machine-readable copy of the corresponding source code, to be
    distributed under the terms of Sections 1 and 2 above on a medium
    customarily used for software interchange; or,

A web page with a download URL is just such an offer. The Internet is a
medium customarily used for software interchange. I do not see why the
following statement doesn't meet the requirements above:

"The source code for this product is available under the terms of the GPL
from the following web page http://www.mycompanyname.com/gpl"

This assumes that no special steps are needed to obtain the software from
that web page.

DS

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Tomas Neme]

> "The source code for this product is available under the terms of the GPL
> from the following web page http://www.mycompanyname.com/gpl"
>
> This assumes that no special steps are needed to obtain the software from
> that web page.

But thats YOURcompanyname.com. Not a third party. If you gave as a
link somebodyelsescompany.com/gpl then somebodyelse could get rid of
the link, and your offer wouldn't be valid for "at least three years"

T

-- 
|_|0|_|
|_|_|0|
|0|0|0|

RE: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Schwartz]

> But thats YOURcompanyname.com. Not a third party. If you gave as a
> link somebodyelsescompany.com/gpl then somebodyelse could get rid of
> the link, and your offer wouldn't be valid for "at least three years"
>
> T

You mean it might not be valid for at least three years. It also might be.
You also might go out of business in less than three years.

You're not violating the GPL simply because you might not be able to honor
your offer in less than three years. You're violating the GPL when you in
fact fail to honor it.

I do agree that providing a link to a third party URL is risky, but I do not
agree that it doesn't comply with the GPL.

DS

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Alexandre Oliva]

On Jun 26, 2007, "David Schwartz" <davids@webmaster.com> wrote:

> Alexandre Oliva:

>> On Jun 26, 2007, Al Boldi <a1426z@gawab.com> wrote:

>> > I read your scenario of the vendor not giving you the source to
>> > mean: not directly; i.e.  they could give you a third-party
>> > download link.

>> This has never been enough to comply with GPLv2.

> A lot of people seem to say this, but I don't think it's true.

http://www.gnu.org/licenses/gpl-faq.html#TOCUnchangedJustBinary and
the 3 questions after that should be enlightening as to why people say
this ;-)

    cost of physically performing source distribution, a complete
            ^^^^^^^^^^

Why would 'physically' be there if it didn't mean anything?  When
interpreting legal texts, that's one sort of question you should ask
yourself.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

RE: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[David Schwartz]

> On Jun 26, 2007, "David Schwartz" <davids@webmaster.com> wrote:
>
> > Alexandre Oliva:
>
> >> On Jun 26, 2007, Al Boldi <a1426z@gawab.com> wrote:
>
> >> > I read your scenario of the vendor not giving you the source to
> >> > mean: not directly; i.e.  they could give you a third-party
> >> > download link.
>
> >> This has never been enough to comply with GPLv2.
>
> > A lot of people seem to say this, but I don't think it's true.
>
> http://www.gnu.org/licenses/gpl-faq.html#TOCUnchangedJustBinary

This consists of entirely unsupported statements.

> and
> the 3 questions after that should be enlightening as to why people say
> this ;-)
>
>     cost of physically performing source distribution, a complete
>             ^^^^^^^^^^
>
> Why would 'physically' be there if it didn't mean anything?

It does. It means you can't charge for adminsitrative or other costs
associated with performing the source distribution. It's necessary because
otherwise someone could claim that it costs them $10,000 to give you the
source code because they need to purchase a license from someone else. This
limits what you can charge for but does not specify what you have to do.

> When
> interpreting legal texts, that's one sort of question you should ask
> yourself.

It's obvious why it's there. If you're going to charge for the distribution,
the charge must be nominal and justified by actual distribution cost.

Note that even a distribution over the Internet must be physically performed
in this sense (actual physical activity by a human being is required to
perform this type of distribution, both in setup and in maintenance). I
would argue that the GPL allows you to charge these costs if you really
wanted to, though it's hard to imagine why anyone would bother.

DS

Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

[Al Boldi]

Alexandre Oliva wrote:
> On Jun 26, 2007, Al Boldi <a1426z@gawab.com> wrote:
> > I read your scenario of the vendor not giving you the source to mean:
> > not directly; i.e.  they could give you a third-party download link.
>
> This has never been enough to comply with GPLv2.

Section 3a of the GPLv2 mentions "a medium customarily used for software 
interchange".  I would think the Internet is a medium customarily used for 
software interchange, is it not?


Thanks!

--
Al