From: Linus Torvalds <torvalds@osdl.org>
To: Alan Cox <alan@lxorguk.ukuu.org.uk>
Cc: Jan Engelhardt <jengelh@linux01.gwdg.de>,
Sergey Panov <sipan@sipan.org>,
James Bottomley <James.Bottomley@SteelEye.com>,
linux-kernel <linux-kernel@vger.kernel.org>
Subject: Re: GPLv3 Position Statement
Date: Wed, 27 Sep 2006 10:28:42 -0700 (PDT) [thread overview]
Message-ID: <Pine.LNX.4.64.0609271000510.3952@g5.osdl.org> (raw)
In-Reply-To: <1159359540.11049.347.camel@localhost.localdomain>
[ This is not so much really a reply to Alan, as a rant on some of the
issues that Alan takes up ]
On Wed, 27 Sep 2006, Alan Cox wrote:
> Ar Mer, 2006-09-27 am 10:58 +0200, ysgrifennodd Jan Engelhardt:
> > I think Linus once said that he does not consider the kernel to
> > become part of a combined work when an application uses the kernel.
>
> COPYING top of the kernel source tree.
Yes. But also somethign much more fundamental:
Copyright Law - regardless of country
You can claim anything you damn well want in a license, but in the end,
it's all about "derived works". If something is not a derived work, it
doesn't matter _what_ ownership rights you have, it's simply not an issue.
So even if the kernel had a big neon sign that said "You're bound by this
license in user space too!" (or, more likely, didn't have a sign at all,
like was the case originally), that has absolutely _zero_ legal meaning
for a copyright license. A license cannot cover something just because it
"says so".
For example, I could write a copyright license that said
"This license forbids you from ever copying the kernel, or any
work made by Leo Tolstoy, which is just a pseudonym for the
easter bunny"
and the fact that the license claims to control the works of Leo "the
easter bunny" Tolstoy, claiming so simply doesn't make it so.
And yes, the above is obviously ridiculous, but the point is, it's no more
ridiculous than a license that would claim that it extends to programs
just because you can run then on Linux.
In fact, it's also no more ridiculous than a license that claims it
extends copyright the other way - to the hardware or platform that you run
a program on. From a legal standpoint, such wording is just totally
idiotic.
[ So the wording at the top of the license is a clarification of fact, not
really any kind of change of the license itself. It actually does have
some legal meaning (it shows "intent"), but most importantly it allows
people to not have to even worry about frivolous lawsuits. Nobody can
sue people for not running GPLv2 user-space through normal system calls,
because the statement of intent makes it clear that a judge would throw
out such a suit immediately.
So I think the important thing here to take away is that "frivolous"
part of the lawsuit. The language doesn't actually _change_ the legal
meaning, but it protects against idiots. And a lot of lawyers worry
about idiots and money-grubbing douchebags *cough*SCO*cough*, and
as such obvious clarifications _can_ be useful. ]
Another real-world example of this mis-understanding is that a lot of
people seem to think that the GPLv2 disallows "linking" with non-GPLv2
code. Almost everybody I ever meet say that, and the FSF has written long
pieces on shared libraries etc.
People don't have a clue!
The GPLv2 never _ever_ mentions "linking" or any other technical measure
at all. Doing so would just be stupid (another problem with the GPLv3,
btw). So people who think that the GPLv2 disallows "linking" with
non-GPLv2 code had better go back and read the license again.
Grep for it, if you want to. The word "link" simply DOES NOT EXIST IN THE
LICENSE!
(It does exist at the end of the _file_ that contains the license, but
that's not actually the license at all, it's just the "btw, this is how
you might _use_ the license", and while legally that can be used to show
"intent", it does not actually extend the copyright in the work in any
way, shape, or form).
What the GPLv2 actually talks about is _only_ about "derived work". And
whether linking (dynamically, statically, or by using an army of worker
gnomes that re-arrange the bits to their liking) means something is
"derived or not" is a totally different issue, and is not something you
can actually say one way or the other, because it will depend on the
circumstances.
I'm always surprised by how many people talk abut the GPLv2 (and, quite
frankly, about the GPLv3 draft) without actually seemingly having ever
read the damn thing, or, more likely, ever really understood any legal
stuff what-so-ever, or the difference between the _use_ of a license, and
the license itself.
For example, in the GPLv3 discussions, I've seen more than one person
claim that I've used a special magic version of the GPLv2 that doesn't
have the "v2 or any later" clause. Again, those people don't have a _clue_
about what they are talking about. They feel very free in arguing about
other peoples copyrigted works, and the fact that I'm not a lawyer, but
then they ignore the fact that I actually _do_ know what I'm talking
about, and that they don't have a stinking clue.
> No. The definition of a derivative work is a legal one and not a
> technical one.
Exactly. A lot of people don't understand this, and a lot of people think
that "derivative" means "being close". Linking doesn't make something
derivative per se - the same way _not_ linking doesn't make it not be
derivative.
Now, it is also indisputable that if you _need_ to "link", it's a damn
good sign that something is _very_likely_ to be derivative, but as Alan
points out, you could do the same thing with RPC over a socket, and the
fact that you did it technically differently really makes no real
difference. So linking per se isn't the issue, and never has been.
Linus
next prev parent reply other threads:[~2006-09-27 17:28 UTC|newest]
Thread overview: 235+ messages / expand[flat|nested] mbox.gz Atom feed top
2006-09-22 16:15 GPLv3 Position Statement James Bottomley
2006-09-22 16:16 ` James Bottomley
2006-09-22 17:49 ` The GPL: No shelter for the Linux kernel? Adrian Bunk
2006-09-22 18:00 ` Greg KH
2006-09-22 18:01 ` Manu Abraham
2006-09-22 20:04 ` David Schwartz
2006-09-22 21:25 ` Linus Torvalds
2006-09-22 21:44 ` Linus Torvalds
2006-09-23 0:11 ` David Schwartz
2006-09-23 1:36 ` Linus Torvalds
2006-09-23 7:28 ` Paul Jackson
2006-09-23 8:05 ` Manu Abraham
2006-09-23 15:32 ` Oleg Verych
[not found] ` <Pine.LNX.4.64.0609230941530.4388@g5.osdl.org>
2006-09-23 21:04 ` Forwarded message from Linus Torvalds <torvalds@osdl.org> Oleg Verych
2006-09-27 1:19 ` The GPL: No shelter for the Linux kernel? Oleg Verych
2006-09-23 8:10 ` Jan Engelhardt
2006-09-23 17:38 ` David Schwartz
2006-09-23 18:00 ` Linus Torvalds
2006-09-23 18:14 ` Petr Baudis
2006-09-24 7:53 ` Jan Engelhardt
2006-09-24 16:34 ` Linus Torvalds
2006-09-25 5:59 ` Jan Engelhardt
2006-09-25 15:14 ` Linus Torvalds
[not found] ` <200609221359.39519.gene.heskett@verizon.net>
2006-09-22 18:08 ` GPLv3 Position Statement James Bottomley
2006-09-22 18:30 ` Gene Heskett
2006-09-22 18:34 ` Jeff Garzik
2006-09-22 18:52 ` Gene Heskett
2006-09-22 19:05 ` Alan Cox
2006-09-22 18:54 ` Gene Heskett
2006-09-25 6:33 ` Marc Perkel
2006-09-22 20:42 ` Jeff Garzik
2006-09-23 11:38 ` Florian Weimer
2006-09-25 2:44 ` An Ode to GPLv2 (was Re: GPLv3 Position Statement) Linus Torvalds
2006-09-25 4:40 ` Willy Tarreau
2006-09-25 12:00 ` Arjan van de Ven
2006-09-25 13:07 ` Willy Tarreau
2006-09-28 0:12 ` H. Peter Anvin
2006-09-25 8:53 ` GPLv3 Position Statement Michiel de Boer
2006-09-25 9:06 ` Russell King
2006-09-25 10:51 ` Neil Brown
2006-09-25 11:31 ` Alan Cox
2006-09-25 16:10 ` Thomas Gleixner
2006-09-29 10:15 ` Helge Hafting
2006-09-29 11:07 ` Alan Cox
2006-09-29 14:02 ` Stephen Clark
2006-09-29 16:51 ` Linus Torvalds
2006-09-29 17:47 ` Alan Cox
2006-09-29 17:49 ` Linus Torvalds
2006-09-29 18:17 ` Linus Torvalds
2006-09-29 18:26 ` Alan Cox
2006-09-29 18:27 ` Linus Torvalds
2006-09-29 18:40 ` Linus Torvalds
2006-09-29 19:59 ` alan
2006-09-29 20:06 ` Linus Torvalds
2006-09-29 20:21 ` alan
2006-09-29 20:32 ` alan
2006-09-29 23:12 ` Gene Heskett
2006-09-29 23:25 ` Randy Dunlap
2006-09-29 23:53 ` Gene Heskett
2006-09-30 0:31 ` Vadim Lobanov
2006-09-30 3:36 ` Gene Heskett
2006-09-30 4:37 ` Vadim Lobanov
2006-09-30 4:54 ` Randy Dunlap
2006-09-30 6:10 ` Vadim Lobanov
2006-09-29 21:11 ` Chris Smith
2006-09-29 21:33 ` alan
2006-09-29 20:51 ` alan
2006-09-29 21:25 ` Theodore Tso
2006-09-29 21:29 ` Alan Cox
2006-10-02 8:46 ` Helge Hafting
2006-09-29 19:43 ` jdow
2006-09-30 18:38 ` Thomas Gleixner
2006-09-30 20:49 ` Alan Cox
2006-09-30 20:38 ` Thomas Gleixner
2006-09-25 14:27 ` Lee Revell
2006-09-25 19:05 ` Jan Engelhardt
2006-09-25 20:58 ` Gene Heskett
2006-09-25 22:10 ` linux-os (Dick Johnson)
2006-09-26 21:32 ` Oleg Verych
2006-09-25 19:46 ` Jeff Garzik
2006-09-25 21:10 ` Gene Heskett
2006-09-25 11:11 ` Jan Engelhardt
2006-09-25 14:12 ` James Bottomley
2006-09-25 16:50 ` Linus Torvalds
2006-09-25 17:26 ` James Bottomley
2006-09-25 15:30 ` Xavier Bestel
2006-09-27 1:11 ` Sergey Panov
2006-09-27 5:55 ` Jan Engelhardt
2006-09-27 7:36 ` Sergey Panov
2006-09-27 8:58 ` Jan Engelhardt
2006-09-27 12:19 ` Alan Cox
2006-09-27 17:28 ` Linus Torvalds [this message]
2006-09-27 18:37 ` Chase Venters
2006-09-27 19:11 ` Linus Torvalds
2006-09-29 12:42 ` Pavel Machek
2006-09-27 22:58 ` Theodore Tso
2006-09-27 23:16 ` Chase Venters
2006-09-28 0:03 ` Neil Brown
2006-09-28 0:08 ` David Miller
2006-09-28 0:18 ` Linus Torvalds
2006-09-28 0:54 ` Patrick McFarland
2006-09-28 3:15 ` Linus Torvalds
2006-09-28 3:47 ` Sergey Panov
2006-09-28 4:13 ` Linus Torvalds
2006-09-28 5:05 ` Sergey Panov
2006-09-28 4:39 ` Chase Venters
2006-09-28 5:13 ` Trond Myklebust
2006-09-28 5:15 ` Jeff Garzik
2006-09-28 5:27 ` Sergey Panov
2006-09-28 5:34 ` Jeff Garzik
2006-09-28 7:30 ` Al Viro
2006-09-28 13:55 ` Lennart Sorensen
2006-09-28 14:19 ` DervishD
2006-09-28 14:40 ` Jörn Engel
2006-09-28 14:59 ` DervishD
2006-09-28 15:04 ` Linus Torvalds
2006-09-28 15:20 ` Jörn Engel
2006-09-28 15:31 ` Linus Torvalds
2006-09-28 15:46 ` Björn Steinbrink
2006-09-28 15:24 ` Linus Torvalds
2006-09-29 0:26 ` Neil Brown
2006-09-29 6:22 ` Linus Torvalds
2006-09-29 1:34 ` jdow
2006-09-29 6:08 ` Jan Engelhardt
2006-09-29 7:07 ` Linus Torvalds
2006-09-29 7:18 ` David Schwartz
2006-09-29 2:29 ` David Schwartz
2006-09-29 2:45 ` Neil Brown
2006-09-29 3:05 ` Björn Steinbrink
2006-09-29 3:31 ` David Schwartz
2006-09-29 5:37 ` Björn Steinbrink
2006-09-29 7:18 ` David Schwartz
[not found] ` <20060928225008.ded4fa2c.seanlkml@sympatico.ca>
2006-09-29 2:50 ` Sean
2006-09-29 7:18 ` David Schwartz
2006-10-02 8:55 ` Maybe it's time to fork the GPL License - create the Linux license? Marc Perkel
2006-10-02 9:14 ` Jesper Juhl
2006-10-02 9:23 ` Marc Perkel
2006-10-03 10:31 ` Jan Engelhardt
2006-10-03 15:34 ` Marc Perkel
2006-10-02 9:18 ` Dumitru Ciobarcianu
2006-10-02 9:25 ` Patrick McFarland
2006-10-02 18:26 ` James Dickens
2006-10-03 20:59 ` Ivan Dimitrov
2006-10-03 21:00 ` Chase Venters
2006-10-03 21:17 ` It's not GNU/Linux - it's jusy LINUX Marc Perkel
2006-10-03 21:41 ` Neil Brown
2006-10-04 20:09 ` Jan Engelhardt
2006-10-04 20:53 ` linux-os (Dick Johnson)
2006-10-03 21:41 ` Adam Henley
2006-10-03 21:47 ` Marc Perkel
2006-10-03 21:58 ` Chase Venters
2006-10-03 22:00 ` Hua Zhong
2006-10-04 1:17 ` Patrick Draper
2006-10-04 2:06 ` Patrick McFarland
2006-10-04 15:16 ` Patrick Draper
2006-10-03 22:10 ` M4y3c0
2006-10-03 22:02 ` Maybe it's time to fork the GPL License - create the Linux license? Daniel Barkalow
2006-09-28 14:51 ` GPLv3 Position Statement Simon Oosthoek
2006-09-28 15:07 ` DervishD
2006-09-28 15:38 ` Linus Torvalds
2006-09-28 18:34 ` Linus Torvalds
2006-09-28 17:16 ` linux-os (Dick Johnson)
2006-09-28 18:59 ` Segher Boessenkool
2006-09-28 19:34 ` linux-os (Dick Johnson)
2006-09-28 20:01 ` Oleg Verych
2006-09-28 23:12 ` Gene Heskett
2006-09-29 8:04 ` Jan Engelhardt
2006-09-28 1:35 ` Al Viro
2006-09-28 3:13 ` Sergey Panov
2006-09-28 3:36 ` Linus Torvalds
2006-09-28 1:53 ` Alan Cox
2006-09-28 9:41 ` Jörn Engel
2006-09-28 9:55 ` Jeff Garzik
2006-09-28 14:45 ` Linus Torvalds
2006-09-28 15:19 ` Alan Cox
2006-09-28 14:40 ` Linus Torvalds
2006-09-28 2:34 ` Gene Heskett
2006-09-28 8:04 ` Jan Engelhardt
2006-09-28 13:50 ` Christer Weinigel
2006-09-28 20:43 ` Linus Torvalds
2006-09-27 18:01 ` Theodore Tso
2006-09-27 12:32 ` Theodore Tso
2006-09-27 21:05 ` David Schwartz
2006-09-29 12:47 ` Pavel Machek
2006-09-27 17:00 ` Linus Torvalds
2006-09-28 8:18 ` Jan Engelhardt
2006-09-28 9:39 ` Samuel Tardieu
2006-09-27 16:08 ` Greg KH
2006-09-25 10:02 tridge
2006-09-27 9:43 Nicolas Mailhot
2006-09-27 17:51 ` Linus Torvalds
2006-09-27 20:34 ` Krzysztof Halasa
2006-09-27 20:41 ` Linus Torvalds
2006-09-27 21:01 ` Krzysztof Halasa
2006-09-27 23:01 ` Alan Cox
2006-09-27 23:04 ` Linus Torvalds
2006-09-27 23:07 ` Thomas Gleixner
2006-09-28 7:36 ` Andi Kleen
2006-09-29 3:01 James Bottomley
2006-09-29 4:40 ` Neil Brown
2006-09-29 6:56 ` James Bottomley
2006-09-29 7:48 ` tridge
2006-09-29 9:37 ` David Schwartz
2006-09-29 10:08 ` Samuel Tardieu
2006-09-29 12:59 ` Alan Cox
2006-09-29 14:36 ` Linus Torvalds
2006-09-30 15:32 ` Andrea Arcangeli
2006-09-30 16:59 ` Linus Torvalds
2006-09-30 22:16 ` tridge
2006-09-29 11:59 ` Sanjoy Mahajan
2006-09-29 14:54 ` Horst H. von Brand
2006-09-29 21:46 ` tridge
[not found] ` <20060929180323.d2c0d2ee.seanlkml@sympatico.ca>
2006-09-29 22:03 ` Sean
2006-09-29 22:20 ` Jeff Garzik
2006-09-29 5:51 ` tridge
2006-09-29 7:31 ` James Bottomley
2006-09-29 12:08 ` Sanjoy Mahajan
2006-09-29 20:50 ` James Bottomley
2006-09-29 7:31 ` James Bottomley
2006-09-29 8:09 ` tridge
2006-09-29 20:28 ` James Bottomley
2006-09-29 8:53 ` tridge
2006-09-29 9:37 ` Jeff Garzik
2006-09-29 9:52 ` tridge
2006-09-30 0:56 ` James Bottomley
2006-09-29 19:50 ` James Bottomley
2006-09-30 7:05 ` tridge
2006-09-30 15:06 ` James Bottomley
2006-10-01 6:28 ` tridge
2006-10-01 15:45 ` James Bottomley
2006-10-01 16:48 ` Valdis.Kletnieks
2006-09-30 21:11 ` David Lang
2006-09-30 22:29 ` Michiel de Boer
2006-10-01 6:52 ` tridge
2006-09-29 7:32 ` David Woodhouse
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=Pine.LNX.4.64.0609271000510.3952@g5.osdl.org \
--to=torvalds@osdl.org \
--cc=James.Bottomley@SteelEye.com \
--cc=alan@lxorguk.ukuu.org.uk \
--cc=jengelh@linux01.gwdg.de \
--cc=linux-kernel@vger.kernel.org \
--cc=sipan@sipan.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).