On Fri, 27 Jan 2006, Alan Cox wrote:
> On Mer, 2006-01-25 at 17:39 -0500, Linus Torvalds wrote:
> > In other words: the _default_ license strategy is always just the
> > particular version of the GPL that accompanies a project. If you want to
> > license a program under _any_ later version of the GPL, you have to state
> > so explicitly. Linux never did.
>
> Not correct. See section 9.
Sorry, I think you're wrong.
We've _always_ said which license explicitly. It's in the COPYING file.
Even before the clarification, the COPYING file has always said
GNU GENERAL PUBLIC LICENSE
Version 2, June 1991
Copyright (C) 1989, 1991 Free Software Foundation, Inc.
51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
Everyone is permitted to copy and distribute verbatim copies
...
at the very top.
How can you say that we didn't specify a version?
If you distribute a program, and you just say "I license this under the
GPL", THEN you don't specify a verion.
But if you distribute a program, and the ONLY license that is associated
with it is a specific version of a license file, then that's what you
have, UNLESS SOMETHING SAYS OTHERWISE.
This is basic copyright law, btw, and has nothing to do with the GPL per
se. If you don't have a license, you don't have any copyright AT ALL.
Linux kernel files don't say "This is licensed under the GPL". Not mine,
at least. I don't see the point, and I never have. There's a COPYING file
that specifies what the license is, and that COPYING file very much
specifies a very _specific_ version of the GPL. Always has.
Linus
On Fri, 27 Jan 2006, Linus Torvalds wrote:
>
> This is basic copyright law, btw, and has nothing to do with the GPL per
> se. If you don't have a license, you don't have any copyright AT ALL.
This is really important, btw.
Yes, when we speak colloquially we talk about the fact that Linux is
licensed "under the GPL", but that is _not_ how anybody actually has ever
gotten a license legally. The ONLY way anybody has ever legally licensed
Linux is either with the original very strict copyright _or_ thanks to the
COPYING file. Nothing else really matters.
So the version of the GPL has always been explicit. At no point has the
kernel been distributed without a specific version being clearly mentioned
in the ONLY PLACE that gave you rights to copy the kernel in the first
place. So either you knew it was GPLv2, or you didn't have the right to
copy it in the first place.
In other words, Linux has _never_ been licensed under anything but the GPL
v2, and nobody has _ever_ gotten a legal Linux source distribution with
anything but a complete copy of GPLv2 license file.
So when I say that the additions at the top of the COPYING file are
nothing but clarifications, I'm not just making that up. Anybody who
claims that any Linux kernel I've ever made has ever been licensed under
anything else than those exact two licenses is just not correct.
And Alan, I know we've had this discussion before. You've claimed before
that my clarifications are somehow "changing" the license, and I've told
you before that no, they don't change the license, they just clarify
things that people keep on gettign wrong, or keep on being nervous about.
So people can argue all they want about this. But unless you get a real
legal opinion (not just any random shyster - a real judge making a
statement, or a respected professional who states his firm legal opinion
in no uncertain terms), I don't think you have a legal leg to stand on.
But no, IANAL. I'd be willing to bet real money that a real lawyer would
back me up on this, though.
Linus
---
PS. Just out of historical interest, the only other copyright license ever
distributed with the kernel was this one:
"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.
Mail me at "[email protected]" if you have any questions."
and that one was only valid between kernel versions 0.01 and 0.12 or
something like that.
On Sat, Jan 28, 2006 at 12:22:58AM -0500, Linus Torvalds wrote:
> PS. Just out of historical interest, the only other copyright license ever
> distributed with the kernel was this one:
>
> "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.
>
> Mail me at "[email protected]" if you have any questions."
>
> and that one was only valid between kernel versions 0.01 and 0.12 or
> something like that.
Interesting... What does that do to e.g. DVD with full (OK, modulo missing
early versions) kernel history all way back to 0.01?
Even funnier question is what does that do to full CVS including the
early versions. Can that be distributed at all and what license would
fit it? Arguing that it's mere aggregation is possible, but it's a
bit of a stretch...
On Sat, 28 Jan 2006, Al Viro wrote:
> >
> > - You may not distibute this for a fee, not even "handling"
> > costs.
> >
> > Mail me at "[email protected]" if you have any questions."
> >
> > and that one was only valid between kernel versions 0.01 and 0.12 or
> > something like that.
>
> Interesting... What does that do to e.g. DVD with full (OK, modulo missing
> early versions) kernel history all way back to 0.01?
Well, the good news is that I was the only real copyright holder back then
(there's a couple of other people who contributed to 0.11 and/or 0.12,
mainly Ted T'so with the BSD terminal control stuff - ^Z and friends).
I used to even re-write patches to suit my style (this was back then, the
patches were smaller, and I was younger and had more energy). So some
things that others sent in patches for (I think Peter McDonald did pty's)
I ended up re-writing myself (and in the process I mixed up the master and
slave pty major number, iirc ;)
> Even funnier question is what does that do to full CVS including the
> early versions. Can that be distributed at all and what license would
> fit it? Arguing that it's mere aggregation is possible, but it's a
> bit of a stretch...
I think you can take it for granted that the GPL re-licensing was
retro-active. I'm the sole copyright holder for 99% of it, and there were
no objections to the relicensing even back when it happened, so I can
pretty much guarantee that there would be none now ;)
It was a kind of strange license. I didn't spend a whole lot of time
thinking about it ;)
Linus
On Gwe, 2006-01-27 at 21:06 -0500, Linus Torvalds wrote:
> > Not correct. See section 9.
>
> Sorry, I think you're wrong.
>
> We've _always_ said which license explicitly. It's in the COPYING file.
>
> Even before the clarification, the COPYING file has always said
>
>
> GNU GENERAL PUBLIC LICENSE
> Version 2, June 1991
--
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.
--
"COPYING" is the license not the program. If it were the program you
would be violating the GPL as the GPL license text is not itself GPL
compatible. It is also not a statement of the version of the software
but a statement of the version of the document itself. Both seem quite
self evident and clear. Since the document itself provides choices for
versions to be applied it cannot be said to imply a version either.
Nowdays you specifically state version 2 then reproduce the license
document. That likewise seems self evident and clear. Section 9 probably
gives you the right to do that for any code that did not specify a
version rule already.
You might also want to ask "if the FSF COPYING text specified the
program version as you claim then how would you specify versions
differently". And likewise "How come the FSF itself, author of the
license, distributes its default COPYING file with code clearly intended
and marked to be GPL v2 or later".
In short your interpretation of the past state of affairs would not
stand up to scrutiny.
> If you distribute a program, and you just say "I license this under the
> GPL", THEN you don't specify a verion.
And merely adding the copying file likewise is still not specifying a
version. You may be *implying* one but that is not specifying.
> Linux kernel files don't say "This is licensed under the GPL". Not mine,
> at least. I don't see the point, and I never have. There's a COPYING file
The point is to avoid ambiguity. Consider the statement in clause 2
--
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 COPYING file is mere aggregation - it is a seperately licensed and
independant work to the program with incompatible conditions. In the
kernel case this does not matter as you included text with COPYING to
make the intent clear, and there is no doubt, nor alternate licenses.
Consider however taking BSD 2 clause licensed code and relicensing it
GPL with changes. If the GPL was merely placed with the code it would
not be clear that the GPL now applied to it, especially if there are
other independant GPL programs in the same archive. The advice text with
the GPL itself thus provides for the 'fail safe' worst case scenario
rather than neccessarily addressing all cases in the minimal and neatest
fashion. Lawyers dislike amibguity because it causes expensive problems
later on.
(Also historically an assertion of copyright was neccessary avoid being
public domain in the USA and some other countries.)
Alan
On Mon, 30 Jan 2006, Alan Cox wrote:
>
> You might also want to ask "if the FSF COPYING text specified the
> program version as you claim then how would you specify versions
> differently". And likewise "How come the FSF itself, author of the
> license, distributes its default COPYING file with code clearly intended
> and marked to be GPL v2 or later".
Alan, you're weasel-wording, and making up arguments that aren't valid.
You can avoid specifying a particular version of the GPL by just saying
"This project is distributed under the GPL" and pointing to the FSF. Not
everybody necessarily even includes the LICENSE file.
Please realize that the kernel is one of the largest open-source projects
in the world, if not _the_ largest. Most projects are much much smaller.
Many projects are very ad-hoc, and not all of them have necessarily been
all that careful with adding LICENSE files etc.
As to your "FSF itself" argument, the code the FSF distributes obviously
_has_ to state "GPL v2 or later" exactly because they want the "or later"
thing to be effective.
Their suggestions very much is to include the GPL license in its entirety
_and_ to say "GPL v2 or later" in the files, exactly because they need to
_expand_ the license from just the one in the license file (ie the "or
later" part is essentially an open-ended dual license).
None of your arguments in any way argue that Linux wouldn't be GPLv2.
> In short your interpretation of the past state of affairs would not
> stand up to scrutiny.
You can claim anything you like. I think you're wrong. But in the
meantime, that doesn't matter. If it ever goes to court, you'll see what a
real judge will claim.
My bet is that my interpretation is the only sane one.
> The COPYING file is mere aggregation - it is a seperately licensed and
> independant work to the program with incompatible conditions.
The fact that the COPYING file has a different copyright really doesn't
matter. It's still part of the release.
It's absolutely not different from having a separate "Release notes" file
which specifies the copyright conditions. That's how Linux-0.01 did it:
the thing was outside the actual main tar-ball, and sent out both as part
of the announcemnt and as a separate file in the same directory on the
ftp-site.
Yes, it may be "mere aggregation", but that has absolutely _zero_ impact
on my argument. It's the only license you have to copy Linux, and it very
much has an EXPLICIT VERSION. Namely version 2.
At no time has Linux ever been distributed without the version of the GPL
that it is distributed under being in any question at all. THAT is my
argument.
And as said - you can argue against it as much as you damn well please. I
simply don't care. I think you are very obviously wrong, but hey, in the
end that doesn't matter either. Take it to a judge. Arguing it to me or to
the public has absolutely zero relevance.
And regardless of what you argue, for the last 5 years there has been the
explicit explanatory note. And you can't claim that people didn't know
about it: if I remember correctly, you yourself sent updates to some of
the files you felt you had copyright on to add the ".. or any later
version" verbiage when I suggested people do so.
So why are you even arguing? It is an UNDENIABLE FACT that a noticeable
portion of the Linux kernel is version-2 only. You'd have to do a lot of
work if you wanted to re-license it - and the burden of proof is on _you_
to do so, not on me. Keeping the old license is the _only_ case that
obviously needs no proof at all, since regardless of circumstances, it's
always safe.
The fact is, the kernel is not licenseable under GPLv3 without tons of
work. Work that I'm not in the least interested in doing, or even helping
with. If you want to start such an effort, I'd suggest:
- spend hours and hours of your time talking to your lawyers, trying to
convince them that your argument has any merit at all. I doubt you'll
be able to do that.
- than start from the state 5 years ago.
Btw, at least in the US, intent actually matters. The fact that I've made
it clear that my _intent_ was always GPLv2 (and I've been very consistent
on this) together with the fact that people have accepted the addendum to
the COPYING file actually _does_ have legal meaning.
Weasel-wording and trying to work around the fact that the version has
always been explicitly mentioned is not a way to make a legal argument.
Yet that's really all your argument boils down to.
I can make very specific arguments for why version 2 ONLY is the specific
license that covers Linux. In contrast you can only make weasel-wording
"but you _could_ misunderstand it to mean xyzzy" kind of noises. That
should tell you something.
Linus
>>In short your interpretation of the past state of affairs would not
>>stand up to scrutiny.
>>
>>
>
>You can claim anything you like. I think you're wrong. But in the
>meantime, that doesn't matter. If it ever goes to court, you'll see what a
>real judge will claim.
>
>My bet is that my interpretation is the only sane one.
>
>
>
And how many times have you actually stood in front of a Judge over IP
and contract issues?
The language "GPLv2 or any later version" is what it is. You can change
it moving forward, but
you cannot undo the past. You put this language in there and IT WAS WHAT
YOU MEANT AT THE
TIME. Trying to alter that would most likely result in a finding you are
acting in bad faith.
Anyway, I am not religious on it, v2 or v3 I don't care, but I like v3
and I am moving my stuff to it. It's a
good idea.
Anyway, good luck figuring it out Linus. You should listen to Alan --
he's right and is trying to help you.
:-)
Jeff
On Tue, 31 Jan 2006, Jeff V. Merkey wrote:
>
> And how many times have you actually stood in front of a Judge over IP and
> contract issues?
Well, at least I know what I'm talking about.
> The language "GPLv2 or any later version" is what it is. You can change
> it moving forward, but you cannot undo the past. You put this language
> in there and IT WAS WHAT YOU MEANT AT THE TIME. Trying to alter that
> would most likely result in a finding you are acting in bad faith.
I did _not_ put that language in, which is the whole point.
That language is in almost all GPL-licensed projects _except_ for Linux.
It's in the FSF guidelines for what they _suggest_ people will do when
they license something under the GPL. It's in all the FSF projects,
obviously, and a lot of other GPL'd projects have also just mindlessly
copied the FSF-suggested boilerplate.
Linux never did. Linux has _never_ had the "v2 or later" license wording
in general. Go take a look. The kernel on the whole tends to not mention
licenses in the individual files, preferring to instead rely on the
external COPYING file that it is distributed with. That's very much on
purpose: I personally _hate_ seeing a screenful of crapola that adds
nothing over and over again.
In short, apart from the very early code in 1991 and early -92 (versions
0.01 through 0.12), Linux has been licensed with _only_ the GPLv2 license
file, and normally no mention of "v2 or later" in the actual sources.
And the way copyrights work, you have only as many rights as explicitly
granted to you, so nobody has _ever_ had rights to re-license Linux under
any other license than the one it came with: namely the GPLv2. Alan is
trying to argue that the fact that it has been licensed under the GPLv2
would somehow "magically" mean that it has been licensed under any version
of GPL that you can pick, BUT THAT IS AN OBVIOUSLY LEGALLY FLAWED
ARGUMENT.
It is so obviously flawed that I'm surprised he continues to argue it.
There has _never_ been anything that says "any version of the GPL", or
indeed just "the GPL" without any version. The version has _always_ been
explicit: the kernel comes with the GPLv2 and no other version. If you
don't accept the COPYING file as the license, then you had no license AT
ALL to distribute Linux under.
So you have one choice, and once choice ONLY: accept the GPLv2 (as
reproduced in COPYING) or don't accept the license at all. The option
that Alan seems to want to do is "I'll take just the word 'GPL' from the
COPYING file, and then stick to that" has simply never been an option.
Now, I can't stop Alan making stupid arguments. People can argue anything
they damn well please, whether it makes sense or not. As SCO has shows us,
people can argue crap for years, even in front of a judge, without any
actual fact or paper to stand on.
And that is what Alan does. He tries to argue that the kernel has somehow
magically been released under "the GPL" (without version specifier), even
though the only license that it was ever released under (apart from the
original non-GPL made-up-by-yours-truly license) very explicitly says
which version it is, in big letters at the very top.
The fact that I made it even _more_ obvious five years ago by adding a
further explanatory notice doesn't change anything at all, except make it
more obvious.
Alan - talk to a lawyer. Really. Show him this email thread and my
arguments, and ask him what he believes. I bet you can get a lawyer to
argue your case if you _pay_ him (lawyers are whores - they are paid to
argue for their client, not for the law), but ask him what he honestly
thinks a judge would rule. THEN come back to me.
Because let's face it, the burden on proof on changing the kernel license
is on _Alan_, not me. Alan is the one arguing for change.
Now, some individual files in the kernel are dual-licensed. Some of them
are dual-licensed with a BSD-license, others are "v2 or later version".
The latter is by no means uncommon, but it's definitely in the minority.
Just to give you an idea:
[torvalds@g5 linux]$ git-ls-files '*.c' | wc -l
7978
[torvalds@g5 linux]$ git grep -l "any later version" '*.c' | wc -l
2720
[torvalds@g5 linux]$ git grep -l "Redistributions in binary form must" '*.c' | wc -l
230
ie of the C files, only about a third have the "any later version"
verbiage needed to be able to convert GPL v2 to v3 (and a small minority
look like they are dual-BSD licensed - I didn't know exactly what to grep
for, so I just picked a part of the normal BSD license, but they can
probably also be converted to GPLv3 thanks to the BSD license being a
strictly less restrictive license).
(I picked just the '*.c' files because that seemed fairer. If you could
_all_ files, the "any later version" percentage drops even further).
Linus
Linus Torvalds wrote:
>On Tue, 31 Jan 2006, Jeff V. Merkey wrote:
>
>
>>And how many times have you actually stood in front of a Judge over IP and
>>contract issues?
>>
>>
>
>Well, at least I know what I'm talking about.
>
>
Linus, it's awfully complicated when a judge has to look at this stuff.
From my experience they have
zero understanding of high tech so they always end up looking at
agreements and wording of contracts
and licenses, since they can understand this.
>
>
>>The language "GPLv2 or any later version" is what it is. You can change
>>it moving forward, but you cannot undo the past. You put this language
>>in there and IT WAS WHAT YOU MEANT AT THE TIME. Trying to alter that
>>would most likely result in a finding you are acting in bad faith.
>>
>>
>
>I did _not_ put that language in, which is the whole point.
>
>
If you can provide this beyond all doubt, then I agree you have a solid
basis to object.
However, it being there does make the whole arguments nebulous. I would
suggest
removing any such language from kernel.org and state GPLv2 ONLY.
>That language is in almost all GPL-licensed projects _except_ for Linux.
>It's in the FSF guidelines for what they _suggest_ people will do when
>they license something under the GPL. It's in all the FSF projects,
>obviously, and a lot of other GPL'd projects have also just mindlessly
>copied the FSF-suggested boilerplate.
>
>Linux never did. Linux has _never_ had the "v2 or later" license wording
>in general. Go take a look. The kernel on the whole tends to not mention
>licenses in the individual files, preferring to instead rely on the
>external COPYING file that it is distributed with. That's very much on
>purpose: I personally _hate_ seeing a screenful of crapola that adds
>nothing over and over again.
>
>In short, apart from the very early code in 1991 and early -92 (versions
>0.01 through 0.12), Linux has been licensed with _only_ the GPLv2 license
>file, and normally no mention of "v2 or later" in the actual sources.
>
>
Again, get rid of this language completely about later versions.
>And the way copyrights work, you have only as many rights as explicitly
>granted to you, so nobody has _ever_ had rights to re-license Linux under
>any other license than the one it came with: namely the GPLv2. Alan is
>trying to argue that the fact that it has been licensed under the GPLv2
>would somehow "magically" mean that it has been licensed under any version
>of GPL that you can pick, BUT THAT IS AN OBVIOUSLY LEGALLY FLAWED
>ARGUMENT.
>
>
It's not. I was also under the impression based on the language "any
later license"
and I am a very bright chap. So if I got it wrong, then imagine how many
other
folks are likely to be confused.
>It is so obviously flawed that I'm surprised he continues to argue it.
>There has _never_ been anything that says "any version of the GPL", or
>indeed just "the GPL" without any version. The version has _always_ been
>explicit: the kernel comes with the GPLv2 and no other version. If you
>don't accept the COPYING file as the license, then you had no license AT
>ALL to distribute Linux under.
>
>So you have one choice, and once choice ONLY: accept the GPLv2 (as
>reproduced in COPYING) or don't accept the license at all. The option
>that Alan seems to want to do is "I'll take just the word 'GPL' from the
>COPYING file, and then stick to that" has simply never been an option.
>
>Now, I can't stop Alan making stupid arguments. People can argue anything
>they damn well please, whether it makes sense or not. As SCO has shows us,
>people can argue crap for years, even in front of a judge, without any
>actual fact or paper to stand on.
>
>
Alan is trying to help you. I have never seen him do anything other than
support you
to the hilt. Sure, disagreements happen, but he is there for you and
Linux and has
been from day one.
>And that is what Alan does. He tries to argue that the kernel has somehow
>magically been released under "the GPL" (without version specifier), even
>though the only license that it was ever released under (apart from the
>original non-GPL made-up-by-yours-truly license) very explicitly says
>which version it is, in big letters at the very top.
>
>The fact that I made it even _more_ obvious five years ago by adding a
>further explanatory notice doesn't change anything at all, except make it
>more obvious.
>
>Alan - talk to a lawyer. Really. Show him this email thread and my
>arguments, and ask him what he believes. I bet you can get a lawyer to
>argue your case if you _pay_ him (lawyers are whores - they are paid to
>argue for their client, not for the law), but ask him what he honestly
>thinks a judge would rule. THEN come back to me.
>
>
Linus, remove all nebulous language and post a notice on kernel.org
clarifying your
position on this code, and I think the issue becomes closed. There's
still the possiblity
that under the doctrine of esstopel, someone can claim or will try to
claim conversion
to GPLv3. You will have to address this when and if it happens.
>Because let's face it, the burden on proof on changing the kernel license
>is on _Alan_, not me. Alan is the one arguing for change.
>
>
A change to GPLv3 would be a good thing for you.
>Now, some individual files in the kernel are dual-licensed. Some of them
>are dual-licensed with a BSD-license, others are "v2 or later version".
>The latter is by no means uncommon, but it's definitely in the minority.
>Just to give you an idea:
>
> [torvalds@g5 linux]$ git-ls-files '*.c' | wc -l
> 7978
> [torvalds@g5 linux]$ git grep -l "any later version" '*.c' | wc -l
> 2720
> [torvalds@g5 linux]$ git grep -l "Redistributions in binary form must" '*.c' | wc -l
> 230
>
>ie of the C files, only about a third have the "any later version"
>verbiage needed to be able to convert GPL v2 to v3 (and a small minority
>look like they are dual-BSD licensed - I didn't know exactly what to grep
>for, so I just picked a part of the normal BSD license, but they can
>probably also be converted to GPLv3 thanks to the BSD license being a
>strictly less restrictive license).
>
>(I picked just the '*.c' files because that seemed fairer. If you could
>_all_ files, the "any later version" percentage drops even further).
>
>
Given, the whole kernel cannot claim multiple licensing -- you have
convinced me
on this point.
Jeff
> Linus
>
>
>
Linus,
For what it's worth, maybe you should run some of these issues past the
Electronic Frontier Foundation (eff.org) or Robin Gross at IPJustice.org
who might be able to review these issues and help make sure you're doing
it right.
Marc Perkel wrote:
> Linus,
>
> For what it's worth, maybe you should run some of these issues past
> the Electronic Frontier Foundation (eff.org) or Robin Gross at
> IPJustice.org who might be able to review these issues and help make
> sure you're doing it right.
Great advice, however, the EFF are not experts in IP law. I would
suggest a good firm in Palo Alto, like Rosati, Goodrich and Sonsini,
they are experts
in corporate IP law and Corporate law in general. They would know all
the loopholes people and companies are likely to use. Stay away from these
freebie Law Groups who are more likely to kiss your butt and tell you
waht you want to hear rather than what you NEED to hear.
OSDL should pick up the tab. It will cost around 30K for a good
analysis from them.
Jeff
>
> -
> To unsubscribe from this list: send the line "unsubscribe
> linux-kernel" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
>
Jeff V. Merkey wrote:
> Marc Perkel wrote:
>
>> Linus,
>>
>> For what it's worth, maybe you should run some of these issues past
>> the Electronic Frontier Foundation (eff.org) or Robin Gross at
>> IPJustice.org who might be able to review these issues and help make
>> sure you're doing it right.
>
>
> Great advice, however, the EFF are not experts in IP law. I would
> suggest a good firm in Palo Alto, like Rosati, Goodrich and Sonsini,
> they are experts
> in corporate IP law and Corporate law in general. They would know all
> the loopholes people and companies are likely to use. Stay away from
> these
> freebie Law Groups who are more likely to kiss your butt and tell you
> waht you want to hear rather than what you NEED to hear.
> OSDL should pick up the tab. It will cost around 30K for a good
> analysis from them.
>
> Jeff
>
Actually EFF does have some good IP lawyers on staff and Robin Gross - a
former EFF employee ([email protected]) is also very up on this. And
the prices is right too if you don't want to spend the big bucks. I
would rather get legal advice from lawyer who care about GPL than
lawyers who just want to run up a big fee and screw you in the end.
It might be worth looking into doing it right before some other SCO type
decides they can get rich by busting Linux.
Marc Perkel wrote:
>
>
> Jeff V. Merkey wrote:
>
>> Marc Perkel wrote:
>>
>>> Linus,
>>>
>>> For what it's worth, maybe you should run some of these issues past
>>> the Electronic Frontier Foundation (eff.org) or Robin Gross at
>>> IPJustice.org who might be able to review these issues and help make
>>> sure you're doing it right.
>>
>>
>>
>> Great advice, however, the EFF are not experts in IP law. I would
>> suggest a good firm in Palo Alto, like Rosati, Goodrich and Sonsini,
>> they are experts
>> in corporate IP law and Corporate law in general. They would know
>> all the loopholes people and companies are likely to use. Stay away
>> from these
>> freebie Law Groups who are more likely to kiss your butt and tell you
>> waht you want to hear rather than what you NEED to hear.
>> OSDL should pick up the tab. It will cost around 30K for a good
>> analysis from them.
>>
>> Jeff
>>
>
> Actually EFF does have some good IP lawyers on staff and Robin Gross -
> a former EFF employee ([email protected]) is also very up on this.
> And the prices is right too if you don't want to spend the big bucks.
> I would rather get legal advice from lawyer who care about GPL than
> lawyers who just want to run up a big fee and screw you in the end.
No. I think tey would salivate to be able to work on Linux stuff with
Linus. Wilson Sonsini, who was on the Novell Board at the time, directed
the lawsuit against me at TRG -- he nailed me no matter what squirming I
did. I have had other dealings with them. They would do a great job
for Linus.
>
> It might be worth looking into doing it right before some other SCO
> type decides they can get rich by busting Linux.
>
>
Not likely. Our building is right next to them and over the past few
months, there's fewer and fewer cars in the parking lot and more lights
turned off in offices in the evening. I doubt they are going to do
anything to Linux. IBM is going to get nailed on some stuff, but I think
Linux is unaffected, at least those parts RedHat, Suse, and the rest of
us need to keep going.
Jeff
El Tue, 31 Jan 2006 09:57:52 -0800 (PST),
Linus Torvalds <[email protected]> escribi?:
> Please realize that the kernel is one of the largest open-source projects
> in the world, if not _the_ largest. Most projects are much much smaller.
Obviously you've never downloaded openoffice source code :)
My kernel sources weight 259 MB after cleaning it, but according to:
http://download.openoffice.org/2.0.0/source.html openoffice 2.0
weights 260.
Openoffice is much better at bloating code, you can't win.
On Tue, 31 Jan 2006, Jeff V. Merkey wrote:
>
> > I did _not_ put that language in, which is the whole point.
>
> If you can provide this beyond all doubt, then I agree you have a solid
> basis to object.
It's really easy to prove.
Look at core kernel source code today, and look at it 10 years ago. Look
at it 15 years ago. Nothing has changed.
The really core files have copyright notices like this:
/*
* kernel/sched.c
*
* Kernel scheduler and related syscalls
*
* Copyright (C) 1991-2002 Linus Torvalds
...
with absolutely no mention of any license rights at all. Not "this is
under the GPL", not "GPLv2 or later". The _only_ license rights anybody
ever had to those files come from the COPYING file, which very clearly
states that it's "version 2, 1991"
(And yes, I'm a lazy bastard. I don't update the years. Some of the files
I wrote still say "1991, 1992" even though they've obviously been edited
since by me. If they fall into the public domain a couple of years
earlier, I really don't see myself caring, since I will have been dead for
a long while by that time _anyway_, judging by the current copyright
nonsense).
[ Side note: the _core_ kernel files are more universally GPL v2-only than
the rest of the kernel. So for example, while almost a third of all C
files have the "any later version" notice in them, when you look at just
the core files in kernel/ mm/ fs/, it's a _lot_ less rare. For example,
in fs/*.c, it's only two files out of 57, and those aren't even the most
core files.
So _qualitatively_, a lot more than "just" two thirds of the kernel are
based on my core files, and are GPLv2 _only_. The "..any later version"
stuff tends to exist mostly in drivers (and some filesystems: 9pfs,
afs, autofs, cifs, jfs, ntfs, ocfs2 have the "any later version" in
them, but the most common ones do not, and are often derived
(admittedly very indirectly, by now) from my original code. ]
> However, it being there does make the whole arguments
> nebulous. I would suggest removing any such language from kernel.org and
> state GPLv2 ONLY.
The COPYING file was edited (over _five_ years ago) to clarify the issue,
exactly because some people were confused. So the COPYING file now
explicitly says:
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.
and that has been the case for the last 5+ years.
(Another clarification is even older: the clarification that using "normal
system calls" is _not_ considered linking, and thus the GPL doesn't infect
any normal user-level programs. That one is over ten years old, since some
people seriously worried about it. Again, it was really pretty obvious
from the license itself, but the clarification made the question stop and
made some people stop worrying unnecessarily).
Alan argues that that extra notice "changed" the license (and that any
code that is older than 5 years would somehow not be GPLv2). I argue
otherwise. I argue that for the whole history, Linux has been v2-only
unless otherwise explicitly specified.
And I don't think even Alan will argue that the "v2 only" thing hasn't
been true for the last five years.
> I was also under the impression based on the language "any later license"
> and I am a very bright chap. So if I got it wrong, then imagine how many other
> folks are likely to be confused.
Exactly. That's why I added the clarification on top of the COPYING file:
people _have_ been confused.
That confusion doesn't stem from Linux, btw, but from the FSF distribution
of the GPLv2 license itself. The license is distributed as one single
file, which actually contains three parts: (1) the "preamble", (2) the
actual license itself and (3) the "How to Apply These Terms to Your New
Programs" mini-FAQ.
And that third part actually contains the wording "(at your option) any
later version.", but a lot of people seemed to not realize that this was
just part of a FSF-suggested boiler-plate on what to put in your source
files. In other words, that was never actually part of the license
itself, but just a "btw, here's how you should use it" post-script.
A lot of people seemed to be confused by that, and this is exactly why the
Linux COPYING file got the additional explanation.
(Side note: from a legal standpoint, "intent" does actually matter in the
US legal system. So the FSF can actually argue that their pre-amble and
their post-script to the license carry legal weight, because it shows
their _intent_. However, they can only argue that for programs that they
own copyright to, or when the license itself might be unclear - they can't
argue that it shows _my_ intent. I've made my intent very clear over the
years, and I've been consistent on this matter, so nobody can claim that
I've "changed the rules").
> Alan is trying to help you. I have never seen him do anything other than
> support you to the hilt. Sure, disagreements happen, but he is there for
> you and Linux and has been from day one.
Absolutely. And I actually try to be very open to changing my mind if
somebody has a valid point. Open source is absolutely not about just the
source code - it's very much about the process, and about (mostly the lack
of) control.
And hey, Alan tends to be mostly right in his concerns. Which is why he's
so respected in the community. I just think that he is off the deep end on
this one, and I have yet to see any actual convincing arguments for his
standpoint.
> Linus, remove all nebulous language and post a notice on kernel.org
> clarifying your position on this code, and I think the issue becomes
> closed.
The thing is, even the _clarification_ HAS BEEN THERE FOR 5 YEARS. At the
very top of the COPYING file.
This really is nothing new. How much more prominent can it be than be in
the top-level COPYING file that gets distributed with every single kernel
version?
> > Because let's face it, the burden on proof on changing the kernel license is
> > on _Alan_, not me. Alan is the one arguing for change.
>
> A change to GPLv3 would be a good thing for you.
A lot of people like the GPLv3. I personally don't _dislike_ the current
draft, but I don't think it's appropriate for the kernel. Part of why I
liked the GPL in the first place (v2 at that point, obviously - v3 hadn't
even been thought about) was that it put no restrictions at all on the
_use_ of binaries.
So I actually prefer the GPLv2. I don't think the current GPLv3 draft is
"evil" or "bad", or anything like that, but it's not the license I would
have selected when I started, and I don't see any reason to change to it
for the kernel.
Linus
On 1/31/06, Diego Calleja <[email protected]> wrote:
> El Tue, 31 Jan 2006 09:57:52 -0800 (PST),
> Linus Torvalds <[email protected]> escribi?:
>
> > Please realize that the kernel is one of the largest open-source projects
> > in the world, if not _the_ largest. Most projects are much much smaller.
>
> Obviously you've never downloaded openoffice source code :)
>
> My kernel sources weight 259 MB after cleaning it, but according to:
> http://download.openoffice.org/2.0.0/source.html openoffice 2.0
> weights 260.
>
> Openoffice is much better at bloating code, you can't win.
It was more a comment on the number of contributors. Not download size.
josh
Linus Torvalds wrote:
>On Tue, 31 Jan 2006, Jeff V. Merkey wrote:
>
>
>>>I did _not_ put that language in, which is the whole point.
>>>
>>>
>>If you can provide this beyond all doubt, then I agree you have a solid
>>basis to object.
>>
>>
>
>It's really easy to prove.
>
>Look at core kernel source code today, and look at it 10 years ago. Look
>at it 15 years ago. Nothing has changed.
>
>The really core files have copyright notices like this:
>
> /*
> * kernel/sched.c
> *
> * Kernel scheduler and related syscalls
> *
> * Copyright (C) 1991-2002 Linus Torvalds
> ...
>
>with absolutely no mention of any license rights at all. Not "this is
>under the GPL", not "GPLv2 or later". The _only_ license rights anybody
>ever had to those files come from the COPYING file, which very clearly
>states that it's "version 2, 1991"
>
>(And yes, I'm a lazy bastard. I don't update the years. Some of the files
>I wrote still say "1991, 1992" even though they've obviously been edited
>since by me. If they fall into the public domain a couple of years
>earlier, I really don't see myself caring, since I will have been dead for
>a long while by that time _anyway_, judging by the current copyright
>nonsense).
>
>[ Side note: the _core_ kernel files are more universally GPL v2-only than
> the rest of the kernel. So for example, while almost a third of all C
> files have the "any later version" notice in them, when you look at just
> the core files in kernel/ mm/ fs/, it's a _lot_ less rare. For example,
> in fs/*.c, it's only two files out of 57, and those aren't even the most
> core files.
>
> So _qualitatively_, a lot more than "just" two thirds of the kernel are
> based on my core files, and are GPLv2 _only_. The "..any later version"
> stuff tends to exist mostly in drivers (and some filesystems: 9pfs,
> afs, autofs, cifs, jfs, ntfs, ocfs2 have the "any later version" in
> them, but the most common ones do not, and are often derived
> (admittedly very indirectly, by now) from my original code. ]
>
>
>
OK. This one might make it.
>>However, it being there does make the whole arguments
>>nebulous. I would suggest removing any such language from kernel.org and
>>state GPLv2 ONLY.
>>
>>
>
>The COPYING file was edited (over _five_ years ago) to clarify the issue,
>exactly because some people were confused. So the COPYING file now
>explicitly says:
>
> 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.
>
>and that has been the case for the last 5+ years.
>
>(Another clarification is even older: the clarification that using "normal
>system calls" is _not_ considered linking, and thus the GPL doesn't infect
>any normal user-level programs. That one is over ten years old, since some
>people seriously worried about it. Again, it was really pretty obvious
>from the license itself, but the clarification made the question stop and
>made some people stop worrying unnecessarily).
>
>Alan argues that that extra notice "changed" the license (and that any
>code that is older than 5 years would somehow not be GPLv2). I argue
>otherwise. I argue that for the whole history, Linux has been v2-only
>unless otherwise explicitly specified.
>
>And I don't think even Alan will argue that the "v2 only" thing hasn't
>been true for the last five years.
>
>
Got it. Agreed. There was a COPYING file inside the tar.gz (you need to
check this one)
I recall seeing but I think it was 2.4 and some header files that still
said this -- better clean those up.
>
>
>>I was also under the impression based on the language "any later license"
>>and I am a very bright chap. So if I got it wrong, then imagine how many other
>>folks are likely to be confused.
>>
>>
>
>Exactly. That's why I added the clarification on top of the COPYING file:
>people _have_ been confused.
>
>That confusion doesn't stem from Linux, btw, but from the FSF distribution
>of the GPLv2 license itself. The license is distributed as one single
>file, which actually contains three parts: (1) the "preamble", (2) the
>actual license itself and (3) the "How to Apply These Terms to Your New
>Programs" mini-FAQ.
>
>And that third part actually contains the wording "(at your option) any
>later version.",
>
Take this out of the GPL license. It's granting as a term of the license
the ability to
adopt later versions. An obvious loophole -- modify the license text and
state
LINE NUMBER AND PARAGRAPH that this language DOES NOT APPLY
and BE SPECIFIC in your explanation. If it's not numbered in the GPL
text, number
it so you can make it CLEAR.
>but a lot of people seemed to not realize that this was
>just part of a FSF-suggested boiler-plate on what to put in your source
>files. In other words, that was never actually part of the license
>itself, but just a "btw, here's how you should use it" post-script.
>
>A lot of people seemed to be confused by that, and this is exactly why the
>Linux COPYING file got the additional explanation.
>
>(Side note: from a legal standpoint, "intent" does actually matter in the
>US legal system. So the FSF can actually argue that their pre-amble and
>their post-script to the license carry legal weight, because it shows
>their _intent_. However, they can only argue that for programs that they
>own copyright to, or when the license itself might be unclear - they can't
>argue that it shows _my_ intent. I've made my intent very clear over the
>years, and I've been consistent on this matter, so nobody can claim that
>I've "changed the rules").
>
>
>
>>Alan is trying to help you. I have never seen him do anything other than
>>support you to the hilt. Sure, disagreements happen, but he is there for
>>you and Linux and has been from day one.
>>
>>
>
>Absolutely. And I actually try to be very open to changing my mind if
>somebody has a valid point. Open source is absolutely not about just the
>source code - it's very much about the process, and about (mostly the lack
>of) control.
>
>And hey, Alan tends to be mostly right in his concerns. Which is why he's
>so respected in the community. I just think that he is off the deep end on
>this one, and I have yet to see any actual convincing arguments for his
>standpoint.
>
>
Alan is awesome -- period.
>
>
>>Linus, remove all nebulous language and post a notice on kernel.org
>>clarifying your position on this code, and I think the issue becomes
>>closed.
>>
>>
>
>The thing is, even the _clarification_ HAS BEEN THERE FOR 5 YEARS. At the
>very top of the COPYING file.
>
>
CITE REFERENCES. Number the GPL and Cite where it is invalid.
>This really is nothing new. How much more prominent can it be than be in
>the top-level COPYING file that gets distributed with every single kernel
>version?
>
>
>
>>>Because let's face it, the burden on proof on changing the kernel license is
>>>on _Alan_, not me. Alan is the one arguing for change.
>>>
>>>
>>A change to GPLv3 would be a good thing for you.
>>
>>
>
>A lot of people like the GPLv3. I personally don't _dislike_ the current
>draft, but I don't think it's appropriate for the kernel. Part of why I
>liked the GPL in the first place (v2 at that point, obviously - v3 hadn't
>even been thought about) was that it put no restrictions at all on the
>_use_ of binaries.
>
>So I actually prefer the GPLv2. I don't think the current GPLv3 draft is
>"evil" or "bad", or anything like that, but it's not the license I would
>have selected when I started, and I don't see any reason to change to it
>for the kernel.
>
>
Lets see where it goes. It protects the end users and gives them the
freedom to combine
GPL and non-GPL stuff, which is great.
Jeff
> Linus
>
>
>
On Tue, 31 Jan 2006, Linus Torvalds wrote:
> under the GPL", not "GPLv2 or later". The _only_ license rights anybody
> ever had to those files come from the COPYING file, which very clearly
> states that it's "version 2, 1991"
No one has disputed that I think. My understanding of that text, as
in:
GNU GENERAL PUBLIC LICENSE
Version 2, June 1991
Is that it refers as the version of the license itself. The 'revision
ID' of the published license. And that very version of the GPL has a
clause to allow version 'conversion' in section 9 to other versions.
> The COPYING file was edited (over _five_ years ago) to clarify the issue,
That wasn't clarifying the issue, that was to make sure section 9's
"any version" could no longer apply. You even seem to have
acknowledged that very text of section 9 in your announcement of
2.4.0-test8's COPYING change:
`There's been some discussions of a GPL v3 which would limit
licensing to certain "well-behaved" parties, and I'm not sure I'd
agree with such restrictions - and the GPL itself allows for "any
version" so I wanted to make this part unambigious as far as my
personal code is concerned."'
From:
http://www.uwsg.iu.edu/hypermail/linux/kernel/0009.1/0096.html
Until you made that change, a reasonable person might have presumed
that lack of version statement (no, the actual GPL license version at
the top of the GPL itself does not count ;) ) would mean section 9
would have applied.
I'm not familiar enough with Linux kernel history to know whether the
implicit "only version 2" view was widely known amongst kernel
developers back then, prior to the edit.
> and that has been the case for the last 5+ years.
Right. The "version 2 only" only has been the status quo for 5+ years
now, so most people must be content with it, grudgingly or not.
> Exactly. That's why I added the clarification on top of the COPYING
> file: people _have_ been confused.
Read your own email again. :)
> That confusion doesn't stem from Linux, btw, but from the FSF distribution
> of the GPLv2 license itself. The license is distributed as one single
> file, which actually contains three parts: (1) the "preamble", (2) the
> actual license itself and (3) the "How to Apply These Terms to Your New
> Programs" mini-FAQ.
>
> And that third part actually contains the wording "(at your option) any
> later version.",
Right.
But the actual /license/, the second part, says none specified ->
"any version ever published" in section 9. The "any later version"
text is in the 3rd part, the suggested preamble.
> files. In other words, that was never actually part of the license
> itself, but just a "btw, here's how you should use it" post-script.
Right. But there are *two* (any.*version):
1. "Any version", which is:
a) In the license itself, to cover case where programme
doesn't specify which version(s) apply, as section 9.
(ie in the "second part").
b) the text you refered to in the 2.4.0-test8 above (ie
you were looking at section 9 back then, not the preamble ;) )
2. "Any later version", which is what is mentioned in the preamble
(the third part).
Alan's reply to you was on the basis of 1. You're arguing based on 2,
having apparently completely overlooked 1 (though you apparently had
not back when you composed that 2.4.0-test8 email).
regards,
--
Paul Jakma [email protected] [email protected] Key ID: 64A2FF6A
Fortune:
God is real, unless declared integer.
Please take me off this thread.
If I want a sophomoric soporific, then I'll read it on LKML.
Linus Torvalds wrote:
> Alan argues that that extra notice "changed" the license (and that any
> code that is older than 5 years would somehow not be GPLv2). I argue
> otherwise. I argue that for the whole history, Linux has been v2-only
> unless otherwise explicitly specified.
Hope you don't mind an opinion from a bystander...
I actually believe that Alan Cox is making a fair point, when viewed
from the perspective of the strict language of the license.
The license as distributed with the kernel itself states that unless the
program specifies the version of the GPL that it's under, any version
will do. Alan makes the point that at least upto the 2.4.0-test8
addition, the program never specified the version, as the "v2" was only
in the _license_, which is not the _program_.
And this is not a bad point -- the license and the program are indeed
not the same; they are not even copyrighted by the same people. With the
addition to the COPYING file, there's something added which clearly is
yours, but before that it was just the generic GPL, copyrighted by the FSF.
As additional "proof" of the fact that the license can not be considered
part of "the program" he pointed out that the GPL document in itself is
not GPL compatible, meaning that under this strict interpretation it
could even be argued that there would be a legal problem in combining
this document with the rest of the program.
Sure, I noted all the "intent" stuff, and you may feel an interpretation
this strict is not sensible, but I really do believe he's making a fair
point if you do not want to trust the law, and all judges, to always be
as sensible as you want them to be.
> And I don't think even Alan will argue that the "v2 only" thing
> hasn't been true for the last five years.
I would not at least. The added bit needs to be considered part of the
program itself, solving the issue.
Rene.
On Tue, 31 Jan 2006, Paul Jakma wrote:
>
> http://www.uwsg.iu.edu/hypermail/linux/kernel/0009.1/0096.html
>
> Until you made that change, a reasonable person might have presumed that lack
> of version statement (no, the actual GPL license version at the top of the GPL
> itself does not count ;) ) would mean section 9 would have applied.
Actually, I don't think it was really debatable even before that, but yes,
I wanted to cover my ass. _Exactly_ because of that confusion over "any
version" in section 9.
IOW, it isn't a new confusion. It's an old confusion that I've always felt
was silly - and incorrect - and that I wanted to address explicitly,
exactly so that there wouldn't be any ambiguoity.
So I maintain that "version 2" has always been the version of the GPL as
it pertains to the kernel as far as I'm concerned, but exactly because
some other people have been confused, I wanted to make it not only
explicit, but also mention it publicly when the clarification was done, so
that people who _had_ been confused could just add in the necessary
verbiage so that code that they owned would fall under the license they
had _thought_ was the right one.
Note how I called it a "clarification" even back then. It wasn't a change
to make things GPLv2 - it was _clarifying_ that the kernel had always been
GPLv2, and acknowledging that there had indeed been confusion.
And hey, if people want to put the confusion at my door, go right ahead.
I'm just happy I did it five years ago, so that by _now_, things are
hopefully pretty damn clear-cut.
Linus
On Tue, 31 Jan 2006, Rene Herman wrote:
>
> And this is not a bad point -- the license and the program are indeed not the
> same; they are not even copyrighted by the same people.
Umm.. I really think that is a total strawman.
_Most_ of the kernel isn't copyrighted by "the same people". Many parts of
the kernel are available from tons of different people (and are sometimes
available under different licenses). That doesn't make them less part of
the kernel program.
And btw, the GPL (the license text) is not incompatible with itself. Yes,
the license says
"Everyone is permitted to copy and distribute verbatim copies
of this license document, but changing it is not allowed."
but the fact is, the GPL also says that any license notices must be kept
intact, and that a copy of the GPL itself must be given along with the
program (in section 1).
The requirement that you be able to modify and distribute the modified
(section 2) is only _provided_ you also honor section 1. So the fact that
you're not allowed to change the license text is _not_ against the GPL
itself, and including the license as part of the program is in no way
against the GPL.
So the fact that the license has a different copyright and is written by
different people is in _no_ way different from the fact that other parts
of the kernel were perhaps originally BSD-licensed. GPLv2 can happily
cover those cases.
So go back instead to "section 0". Now THAT is I think the relevant one:
"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. ..."
Notice how the GPLv2 text says that it applies to any program that just
says it is licensed under the General Public License.
I'm convinced _that_ is how you get "no version specified" in section 9.
You have a program that just says "This is licensed under the GPL",
instead of doing the full thing.
And I say that the Linux kernel has contained a notice placed by the
copyright holder (the "COPYING" file) that says that it's to be
distributed under (and I quote from the top):
GNU GENERAL PUBLIC LICENSE
Version 2, June 1991
and that's it.
Ok. Back to work. This thread has been _waa-aayy_ too long.
Linus
On Tue, 31 Jan 2006, Linus Torvalds wrote:
> And btw, the GPL (the license text) is not incompatible with itself. Yes,
> the license says
>
> "Everyone is permitted to copy and distribute verbatim copies
> of this license document, but changing it is not allowed."
Right, that's to discourage any hasty and likely ill-advised 'local'
changes to the GPL. The FSF do grant permission to modify the GPL
though:
http://www.gnu.org/licenses/gpl-faq.html#ModifyGPL
regards,
--
Paul Jakma [email protected] [email protected] Key ID: 64A2FF6A
Fortune:
If there is no wind, row.
-- Polish proverb
On Maw, 2006-01-31 at 12:38 -0800, Linus Torvalds wrote:
> Alan argues that that extra notice "changed" the license (and that any
> code that is older than 5 years would somehow not be GPLv2). I argue
> otherwise. I argue that for the whole history, Linux has been v2-only
> unless otherwise explicitly specified.
>
> And I don't think even Alan will argue that the "v2 only" thing hasn't
> been true for the last five years.
I would argue its irrelevance
Two cases (lets call them a and b)
a) The GPLv2 only was always the case
b) There was no version so it was open to choice
Which ultimately means either
a) Linus changed nothing
b) Linus chose a version as the License allowed him to in accordance
with section 9.
So we have two legal outcomes both of which produce the right answer for
any vaguely recent source tree. At which point does it matter ?
My point was to make clear that assuming the GPL original text implies
the version of the code is wrong, and explain why the FSF recommend the
long text.
Is there doubt about the license status of the current code - not in
this area, no. The COPYING file is extremely clear on this, and more
importantly in other possible unclear and problematic areas. For example
the statement that the system calls are not derivative statement which
resolves the biggest interpretation concern of all.
Alan
On Maw, 2006-01-31 at 11:07 -0800, Linus Torvalds wrote:
> granted to you, so nobody has _ever_ had rights to re-license Linux under
> any other license than the one it came with: namely the GPLv2. Alan is
> trying to argue that the fact that it has been licensed under the GPLv2
> would somehow "magically" mean that it has been licensed under any version
> of GPL that you can pick, BUT THAT IS AN OBVIOUSLY LEGALLY FLAWED
> ARGUMENT.
Clause 9 is clear and is part of the GPL v2. The GPL v2 text gives
people (including you) that right itself if no version is specified in
the program. You and the other authors granted that right if you didn't
include a statement about version. Now as it happens various
contributors specified versions long ago and you clarified it too.
> Alan - talk to a lawyer.
Actually I did, long ago before this argument even appeared, because it
was important for another situation.
Also please get one thing straight. I'm not arguing for a license
change, I'm pointing out misunderstandings that might lead people,
particularly other GPL projects to make mistakes.
The big problem with any license change is actually the moral one, as
I'm sure you'd agree: Do you have the moral right to change the rules
(eg on DRM) when many have contributed with many differing views and in
many ways not all of them leading to them being copyright holders ?
Alan
On Maw, 2006-01-31 at 09:57 -0800, Linus Torvalds wrote:
> The fact that the COPYING file has a different copyright really doesn't
> matter. It's still part of the release.
Law is about precision and exact wording as well as intent. The exact
wording is "the Program" not "the release". And Program is capitalised
to indicate the use of the definition made earlier. That is: "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"
The COPYING file is not derivative and is not by any usual
interpretation part of the program. As I said and showed earlier the
COPYING file cannot be part of the program without contradiction.
Ask a law student if you don't believe me, or do it in equations not
words...
> It's absolutely not different from having a separate "Release notes" file
> which specifies the copyright conditions. That's how Linux-0.01 did it:
> the thing was outside the actual main tar-ball, and sent out both as part
> of the announcemnt and as a separate file in the same directory on the
> ftp-site.
The release notes are part of the program. The section starting "NOTE!"
is maybe part of the program (thats a real lawyer question). However
whether it is or not is such clear intent that nobody would win an
argument that you had not said nowdays that it is V2
> I can make very specific arguments for why version 2 ONLY is the specific
> license that covers Linux. In contrast you can only make weasel-wording
Well if you consider accurate interpretation of language 'weasel
wording' I can only feel sorry for you and be glad you write code not
English. But since I don't actually want to change the license on the
kernel and I already can if I wanted for the small percentage I wrote I
don't see the need to continue the debate.
Alan
On Tue, 31 Jan 2006, Alan Cox wrote:
>
> I would argue its irrelevance
Well, I can absolutely agree with that.
It really wasn't ever meant to _change_ anything, so it definitely should
be irrelevant in that sense.
That said, I think it actually hass one big result: we may be discussing
this for a couple of weeks, but I'm pretty sure we won't be discussing it
for months and having a huge split over it in the kernel community.
Which to me makes it less-than-irrelevant ;)
I actually tend to have two very different kinds of worries:
- the technical problem of the month. Something doesn't work, and we
don't know why, and it's been broken for long enough that it seems to
be pretty fundamental.
These things worry me, but they don't keep me up at night.
- the worry that somebody who submitted code to the kernel feels that his
code is mis-used and feels let down by the process.
This thing _worries_ me. This is something I end up losing sleep over.
The second case is unusual, but it does happen. Things like the
MMX-optimized AES routines, where Jari Ruusu ended up objecting to his
code getting used. I really don't think he had any _legal_ reasons to feel
that way, but _that_ is what really really tends to make me unhappy:
regardless of legal correctness, I want developers to feel _proud_ of what
they did in Linux, not feel like their code is being trampled on.
Do unto others.. and so on.
This is also why I don't much like BSD->GPL conversion, and try to ask the
original author to OK it (the way we did with the original AES assembly
code authored by Dr Brian Gladman - just to make sure). And regardless, I
want to keep dual licenses _actively_ dual-licensed in the kernel, just to
respect the original wishes instead of just converting them to the
(stricter) GPLv2.
So the reason I've spent a lot of time on this thread is basically that I
worry about people who would _like_ to upgrade (and incorrectly _expected_
to upgrade) the whole kernel to GPLv3 being unhappy.
So I'm spending time on this thread trying to make sure that everybody
realizes that GPLv2 was always the license of choice - people may still
have wished for something else, but at least I can do my damned best to
explain why things are how they are, and explaining that any expectations
to the contrary really were misguided.
> Two cases (lets call them a and b)
>
> a) The GPLv2 only was always the case
> b) There was no version so it was open to choice
>
> Which ultimately means either
>
> a) Linus changed nothing
> b) Linus chose a version as the License allowed him to in accordance
> with section 9.
>
> So we have two legal outcomes both of which produce the right answer for
> any vaguely recent source tree. At which point does it matter ?
Well, it does matter if somebody wants to start anew at the state we were
at five years ago in the belief that that old state is GPLv3-compatible,
and then add in the more modern part of the files that are explicitly
compatible with GPLv3...
And if somebody feels _that_ strongly about it, and feels that five years
ago there really was a real ambiguity, I'd be interested to hear his or
her arguments. And I don't see myself likely objecting to it, if only
because I'd be intrigued to see how far they get ;)
> Is there doubt about the license status of the current code - not in
> this area, no. The COPYING file is extremely clear on this, and more
> importantly in other possible unclear and problematic areas. For example
> the statement that the system calls are not derivative statement which
> resolves the biggest interpretation concern of all.
I was always surprised by how anybody could _possibly_ worry about that
one, but it did come up very early. I forget exactly when, but I think
that clarification was added way before Linux-1.0.
In general, people worry too much. And I worry about other people
worrying.
Hopefully I also worry without any real cause.
Linus
<trimming the wide Cc list due to complaints>
On Tue, 31 Jan 2006, Linus Torvalds wrote:
> Notice how the GPLv2 text says that it applies to any program that
> just says it is licensed under the General Public License.
>
> I'm convinced _that_ is how you get "no version specified" in section 9.
> You have a program that just says "This is licensed under the GPL",
> instead of doing the full thing.
> And I say that the Linux kernel has contained a notice placed by the
> copyright holder (the "COPYING" file) that says that it's to be
> distributed under (and I quote from the top):
>
> GNU GENERAL PUBLIC LICENSE
> Version 2, June 1991
>
> and that's it.
Well, most people would recognise this to be exactly the same text of
the GPLv2 as published by the FSF, as used by many software
programmes:
http://www.gnu.org/licenses/gpl.txt
and hence would have trouble recognising that you intended this to
also be the text governing exactly which version(s) of the GPL apply
linux specifically.
Note that if what you say is correct, that the immutable header of
the GPLv2 acts so as to apply a version restriction on any specific
covered programme per default, then that would mean:
1a. The "If the Program does not specify a version" text of Section 9
of the GPLv2 is utterly without effect and meaningless.
1b. Indeed, the whole of section 9 might be without effect (the
version is already specified). (It'd depend on the exact
text of the preamble maybe, but it'd be quite ambigious).
2. Hence all GPLv2 programmes which do not have include some
additional "or any later version" preamble text definitely will
not automatically upgrade to GPLv3 when it is finally published.
You've mentioned intent a few times before as being a strong factor
in interpretation[1], you have to therefore ask what RMS and the FSF
intended when they included section 9: did they really intend that
the GPLv2's 'Version 2, June 1991' would completely override the text
regarding cases where "the Program does not specify a version number
of this License"? The intent surely of section 9 was to allow for the
GPLv2 to be upgraded smoothly, even where a programme applied it
imprecisely.
Hence, logically, the intent of the framers must have been that the
'Version 2, June 1991' was to act as a version identifier for the GPL
text only, rather than to be interpreted as the version applying to
the whole programme.
That the licence then goes on to discuss how to apply the GPL (the
third part you mentioned before), and mention how to do so (including
the "any later version" text) further suggests that simply including
a copy of the the text of the license itself does not constitute
tying down the version.
Note that I'm only arguing with you on your interpretation of the GPL
- which I believe is flawed and deserves to be contradicted lest
others assume it - not on which version of the GPL applies to Linux
in its aggregate (which definitely is "v2 only", and has been for
ages).
regards,
--
Paul Jakma [email protected] [email protected] Key ID: 64A2FF6A
Fortune:
Ernest asks Frank how long he has been working for the company.
"Ever since they threatened to fire me."
On Maw, 2006-01-31 at 16:18 -0800, Linus Torvalds wrote:
> That said, I think it actually hass one big result: we may be discussing
> this for a couple of weeks, but I'm pretty sure we won't be discussing it
> for months and having a huge split over it in the kernel community.
Agreed
> Do unto others.. and so on.
This is how you spot a lawyer from a normal human. The human agrees but
the lawyer says "but what about my dentist, will he really want me to
drill his teeth" ;)
> So I'm spending time on this thread trying to make sure that everybody
> realizes that GPLv2 was always the license of choice - people may still
The intent is clear by any view for current code, and the past is for
argument but equally its unfixable whatever the result.
> I was always surprised by how anybody could _possibly_ worry about that
> one, but it did come up very early. I forget exactly when, but I think
> that clarification was added way before Linux-1.0.
>
> In general, people worry too much.
Its the job of some of them to worry. After all many of them have
shareholders to answer for, and whether they worry about misusing your
pension fund or whether they'll still get a new yacht if there is a
lawsuit, they worry.
I've actually talked to a number of people for whom the syscall
clarification was critical to their choice to produce software for Linux
so don't underestimate your smart planning in advance ;)
Alan
Alan Cox wrote:
> On Maw, 2006-01-31 at 09:57 -0800, Linus Torvalds wrote:
>
>> The fact that the COPYING file has a different copyright really doesn't
>> matter. It's still part of the release.
>>
>
> Law is about precision and exact wording as well as intent. The exact
> wording is "the Program" not "the release". And Program is capitalised
> to indicate the use of the definition made earlier. That is: "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"
>
>
Actually the law is more precisely about the intent of the people making
the license/contract as opposed to the wording. Especially when there is
fast moving technology involved. If in 1991 Linus released his software
under GPL v2 and the wold changes in 15 years so that words mean things
differently than they did then, or newer licenses come out to deal with
new issues not contemplated back then, the meaning of the agreement is
interpreted on the basis of the intent and not on the exact wording of
the contract.
For example. I am a painter and I contract to paint your house white at
123 main street and I show up on the job and your house is really 125
main street and you change your mind and decide to have me paint it
yellow and I do that, and then you say you don't owe me because that
contract says you agreed to paint 123 main st white and you painted 125
main street yellow - you lose. The intent of the contract was that I
paint your house.
The meaning of the Linux license is what Linus intended it to be in 1991
and the way it is used in the following years. People who contribute to
the Kernel are presumed to understand the spirit of the project and if
their license says something else but they contributed it to the Kernel
then they are deamed to have waived their rights by contributing it to
the Kernel and accepting the Kernel's license.
I have way too much legal experience for someone who isn't a lawyer, but
if you don't believe me print this out and run it past a real lawyer and
see if they don't agree. The reality is that the Kernel license it what
Linus says it is and we all trust Linus as a condition for contributing
to the Kernel.
Having said that, lawyers can argue anything so if the wording and
reality were the same it makes it easier. So it is worth it to some
extent to clear up issues if we can.
On Tuesday 31 January 2006 16:45, Paul Jakma wrote:
>On Tue, 31 Jan 2006, Linus Torvalds wrote:
>> under the GPL", not "GPLv2 or later". The _only_ license rights
>> anybody ever had to those files come from the COPYING file, which
>> very clearly states that it's "version 2, 1991"
>
>No one has disputed that I think. My understanding of that text, as
>in:
>
>
> GNU GENERAL PUBLIC LICENSE
> Version 2, June 1991
>
>Is that it refers as the version of the license itself. The 'revision
>ID' of the published license. And that very version of the GPL has a
>clause to allow version 'conversion' in section 9 to other versions.
>
>> The COPYING file was edited (over _five_ years ago) to clarify the
>> issue,
>
>That wasn't clarifying the issue, that was to make sure section 9's
>"any version" could no longer apply. You even seem to have
>acknowledged that very text of section 9 in your announcement of
>2.4.0-test8's COPYING change:
>
>`There's been some discussions of a GPL v3 which would limit
>licensing to certain "well-behaved" parties, and I'm not sure I'd
>agree with such restrictions - and the GPL itself allows for "any
>version" so I wanted to make this part unambigious as far as my
>personal code is concerned."'
>
>From:
>
>http://www.uwsg.iu.edu/hypermail/linux/kernel/0009.1/0096.html
>
>Until you made that change, a reasonable person might have presumed
>that lack of version statement (no, the actual GPL license version at
>the top of the GPL itself does not count ;) ) would mean section 9
>would have applied.
>
>I'm not familiar enough with Linux kernel history to know whether the
>implicit "only version 2" view was widely known amongst kernel
>developers back then, prior to the edit.
>
>> and that has been the case for the last 5+ years.
>
>Right. The "version 2 only" only has been the status quo for 5+ years
>now, so most people must be content with it, grudgingly or not.
>
>> Exactly. That's why I added the clarification on top of the COPYING
>> file: people _have_ been confused.
>
>Read your own email again. :)
>
>> That confusion doesn't stem from Linux, btw, but from the FSF
>> distribution of the GPLv2 license itself. The license is distributed
>> as one single file, which actually contains three parts: (1) the
>> "preamble", (2) the actual license itself and (3) the "How to Apply
>> These Terms to Your New Programs" mini-FAQ.
>>
>> And that third part actually contains the wording "(at your option)
>> any later version.",
>
>Right.
>
>But the actual /license/, the second part, says none specified ->
>"any version ever published" in section 9. The "any later version"
>text is in the 3rd part, the suggested preamble.
>
>> files. In other words, that was never actually part of the license
>> itself, but just a "btw, here's how you should use it" post-script.
>
>Right. But there are *two* (any.*version):
>
>1. "Any version", which is:
>
> a) In the license itself, to cover case where programme
> doesn't specify which version(s) apply, as section 9.
> (ie in the "second part").
>
> b) the text you refered to in the 2.4.0-test8 above (ie
> you were looking at section 9 back then, not the preamble ;) )
>
>2. "Any later version", which is what is mentioned in the preamble
> (the third part).
>
>Alan's reply to you was on the basis of 1. You're arguing based on 2,
>having apparently completely overlooked 1 (though you apparently had
>not back when you composed that 2.4.0-test8 email).
>
>regards,
Well, IANAL either, but section 9 isn't the least bit ambiguous in my
opinion. To quote what everyone with a functioning eye can read:
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.
End of snip. The point is that version 2 is spelled out at the top of
the COPYING file and the word "and" preceeds the "any later version" in
the above paragraph, which it does not in the top of the COPYING file
specifying the license, means that its version 2 and thats the end of
it. There is no "any later version" in the statement specifying the
license version at the top of the COPYING file.
Why is that so hard to grasp?
--
Cheers, Gene
People having trouble with vz bouncing email to me should add the word
'online' between the 'verizon', and the dot which bypasses vz's
stupid bounce rules. I do use spamassassin too. :-)
Yahoo.com and AOL/TW attorneys please note, additions to the above
message by Gene Heskett are:
Copyright 2006 by Maurice Eugene Heskett, all rights reserved.
On Tue, Jan 31, 2006 at 11:57:31PM +0000, Alan Cox wrote:
> Clause 9 is clear and is part of the GPL v2. The GPL v2 text gives
> people (including you) that right itself if no version is specified in
> the program. You and the other authors granted that right if you didn't
> include a statement about version. Now as it happens various
> contributors specified versions long ago and you clarified it too.
Alan, Linus,
I have indicated my wish to have my code GPL-2 only in 2000. Linus,
declined to change the files. My intention has been for the
past 5 and a half years to have my code GPL-2 only, only the
files don't reflect that change yet.
REW wrote on Fri, 8 Sep 2000 23:17:09 +0200 (MEST):
> Linus,
> I trust YOU on this. Could you change
> * 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.
> into
> * modify it under the terms of the GNU General Public License
> * version 2 as published by the Free Software Foundation.
> for all the files that I own copyright on?
Linus answered on: Sat, 9 Sep 2000 09:29:01 +0200 (MEST)
> Thanks. For the moment, I'd rather like to not do any extra changes, and
> see what the reaction from the FSF and other people is - let's not hurry
> into this any faster than we have to. I wanted to make a public statement
> because I thought it needed to be done, but at the same time I do not want
> to make a rift between Linux and the FSF bigger than necessary.
>
> So let this lie for a while, and see what happens, ok?
Roger.
--
** [email protected] ** http://www.BitWizard.nl/ ** +31-15-2600998 **
*-- BitWizard writes Linux device drivers for any device you may have! --*
Q: It doesn't work. A: Look buddy, doesn't work is an ambiguous statement.
Does it sit on the couch all day? Is it unemployed? Please be specific!
Define 'it' and what it isn't doing. --------- Adapted from lxrbot FAQ
Let's give this thread a rest, OK?
Whether or not Linux is licensed under the GPLv2 only or not is
ultimately a matter for the lawyers.
The one big problem I see with the GPLv3 effort is given the
additional restrictions regarding DRM, it doesn't seem to clear to me
whether a project which has even a single line of GPLv2-only code can
accept GPLv3 code. That is, GPLv3 is designed to be compatible with
more licenses, but that doesn't matter of GPLv2 isn't compatible with
GPLv3. If that is the case, if only a _single_ person (like Rogier)
has I want GPLv2-only, the whole project is can't use any GPLv3 code
unless they are willing to track down and rewrite all of the code
written by that person or persons.
If that is true (and again, at the end of the day lawyers or more
importantly, a judge is going to have to make that call, not debating
technologists) it's hard to see the GPLv3 making any headway.
- Ted
Linus Torvalds wrote:
> _Most_ of the kernel isn't copyrighted by "the same people". Many
> parts of the kernel are available from tons of different people (and
> are sometimes available under different licenses). That doesn't make
> them less part of the kernel program.
Granted, I could have probably done without the "copyrighted by the same
people" argument. What I attempted to emphasize is that the license text
itself is an independent work, licensed under a different license than
the program.
> And btw, the GPL (the license text) is not incompatible with itself.
> Yes, the license says
>
> "Everyone is permitted to copy and distribute verbatim copies of this
> license document, but changing it is not allowed."
>
> but the fact is, the GPL also says that any license notices must be
> kept intact, and that a copy of the GPL itself must be given along
> with the program (in section 1).
The GPL text itself really is not under a GPL compatible license. The
GPL says that, under certain provisions, "You may modify your copy or
copies of the Program or any portion of it" (section 2) while the GPL
document does not allow any modification, under any provision: "but
changing it is not allowed".
The bits above (also note how you said "along _with_ the program" and
how the GPL itself says "applies _to_ the program) only further enhance
the point with which I by now fully agree that the GPL text itself is
not part of the program. As such, something other than the GPL text will
have to specify a version to have the _program_ specify a version per
section 9.
Your addition to the COPYING file certainly does just that, meaning any
possible problem has been long solved.
> The requirement that you be able to modify and distribute the
> modified (section 2) is only _provided_ you also honor section 1. So
> the fact that you're not allowed to change the license text is _not_
> against the GPL itself, and including the license as part of the
> program is in no way against the GPL.
To be GPL compatible, you need to be compatible with all requirements of
the GPL, one of which is section 2. The licence the GPL text is under
is not compatible with this section.
No, it's not against the program's license (the program being Linux and
the license the GPL v2) to include the license with the program. There
would only be a potential problem if the GPL text were to be considered
_part of_ the program since then combining the GPL licensed program and
the license text would require the license text to be under a GPL
compatible license. As it is, it is included as a mere aggregation, and
this is certainly fine.
Hope I'm not annoying you -- these are obviously points normally only a
lawyer would enjoy arguing. If I am, maybe pointing out the thoroughly
amusing recursion of having to distribute the GPL alongside the GPL if
the GPL text itself would be licensed under the GPL itself (not even
just a compatible license) makes it better? (<-- joke...).
> I'm convinced _that_ is how you get "no version specified" in section
> 9. You have a program that just says "This is licensed under the
> GPL", instead of doing the full thing.
>
> And I say that the Linux kernel has contained a notice placed by the
> copyright holder (the "COPYING" file) that says that it's to be
> distributed under (and I quote from the top):
>
> GNU GENERAL PUBLIC LICENSE Version 2, June 1991
>
> and that's it.
I just saw Paul Jakma make a very strong point here. Section 9 would be
useless if having the version in the header, of the same document, would
be all that is required to have the program specify a version.
Having said all that... everybody agrees that the bit you added to the
COPYING file is enough to have the kernel under the GPL v2 only. So,
there's no need for this, but do you think it could be a good idea to
add a file LICENSE, copyrighted by you, to the kernel that says that the
kernel is licensed under the GPL v2, as supplied in the file COPYING
(and moving the syscall bit as well, and restore COPYING to pristine)?
No, this would not change anything other than provide for an even _more_
unambiguous situation. In that case, noone would have to argue whether
or not anything added to the COPYING file was part of the program
proper. On the other hand, if doing so would make some people believe
that anything changed, it could be counter-productive as well.
Wanted to suggest it though...
Rene.
On Wed, 1 Feb 2006, Rene Herman wrote:
> >
> > but the fact is, the GPL also says that any license notices must be
> > kept intact, and that a copy of the GPL itself must be given along
> > with the program (in section 1).
>
> The GPL text itself really is not under a GPL compatible license. The GPL says
> that, under certain provisions, "You may modify your copy or copies of the
> Program or any portion of it" (section 2) while the GPL document does not
> allow any modification, under any provision: "but changing it is not allowed".
You continue to ignore the big important fact that:
- section 2 is _conditional_ on section 1
- section 1 (and FSF guidelines) _requires_ you to leave copyright
notices intact and give the license out along with the program.
IOW, your claim that the GPL requires you to be able to make changes is
INCORRECT.
The GPL requires you to be able to make a certain _class_ of changes, but
other changes (like modifying the copyright license of notices or removing
the license) are expressly _forbidden_ both by the GPL and by copyright
law.
Think about it. The GPL is _not_ incompatible with including the GNU
General Public License as part of the program.
This can, btw, also be shown independently by the fact that the FSF
clearly _intended_ the license to be actively linked into the program:
they ask you (in the "How to Apply These Terms to Your New Programs") to
have commands to view parts of the license if your program is interactive.
In other words, your claim that the license file itself is somehow
"incompatible" with being included as part of the the GPL'd Program is
simply not possible.
Btw, try this:
shell$ gdb
(gdb) show copying
and if you strace it (or do "strings" on the binary), you'll even notice
that the license is very much part of the binary image itself. How much
more "part of the program" can it be?
So if you seriously suggest that the GNU General Public License file is
incompatible with the GPL, then that ignores the fact that
- the license file itself assks you to combine it with the program
(much more tightly than "mere aggregation")
- the GPL license itself has the requirement that copyright licenses be
left intact (actually, the license itself seems to say just keep intact
all notices that "refer to this License", but the FSF has made it clear
that it covers _any_ copyright license notes, and that the GPL is
compatible with the BSD license which requires the same retention of
copyright notices).
or alternatively:
- the FSF actively _encourages_ people to violate their own copyright
(which means that they can't sue, under the doctine of estoppel), and
your point is moot _anyway_.
So. Claiming that the GPL license text itself cannot be part of the
program is disingenious. According to your reading, the modified BSD
license wouldn't be compatible with the GPL either, because it requires:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions, and the following disclaimer,
* without modification.
yet the FSF has clearly stated that this is perfectly fine, even though it
also disallows modifications to the license text.
Quod erat demonstrandum.
Linus
Linus Torvalds wrote:
> but the fact is, the GPL also says that any license notices must be kept
> intact, and that a copy of the GPL itself must be given along with the
> program (in section 1).
The GPL says you must keep intact notices in the Program that refer to the
license. That logically cannot include the license itself. For one thing,
the license is not in the program. For another thing, the license does not
refer to itself.
> I'm convinced _that_ is how you get "no version specified" in section 9.
> You have a program that just says "This is licensed under the GPL",
> instead of doing the full thing.
You are essentially arguing here that the FSF screwed up. They intended the
normal case to be that you got a copy of the GPL with the program but could
substitute a later version at your option. Why would they have made it
almost impossible (or even extra-difficult) to get the result that they
preferred?
A much more logical reading is that the License and the Program are two
distinct things. When it says "notices in the Program", that's precisely
what it means, notices that are in the Program. When it says "the Program
specifies a version number", that's also what it means, a version number
specified by the Program.
Not only does this make it easy to get the result the FSF preferred, but ot
is supported by the License's internal definition of the term "Program".
I don't think you can lock a person down to a single version of the GPL
unless you explicitly say so in each individual file. Otherwise, nothing
prevents someone from separating out one file and distributing it, kicking
in clause 9's rule about a Program that does not specify a version number.
The requirement to keep license notices intact does not mean you must add
them or maintain them. Just that you can't remove or modify them.
It has always been well-understood on this mailing list that a file can be
excerpted from a GPL project and inserted into another one. Surely the FSF
didn't intend to make the default position be that GPL'd projects would be
mutually incompatible.
DS
Linus Torvalds wrote:
> You continue to ignore the big important fact that:
>
> - section 2 is _conditional_ on section 1
I'm not ignoring this -- I really just don't see it to be relevant. All
sections have to be obeyed the same, and in this sense all sections are
conditional on all others. Violate one, and the license is not available
to you. This goes without saying.
> - section 1 (and FSF guidelines) _requires_ you to leave copyright
> notices intact and give the license out along with the program.
>
> IOW, your claim that the GPL requires you to be able to make changes is
> INCORRECT.
There are two things here that need seperating. On the one hand, we have
"the program". On the other we have "the license" (and license notices).
The GPL requires you to be able to make changes to the program, not the
license. In the context of the Linux kernel, Linux is the program, and
the GPLv2 is the license. In the context of the GPLv2, the GPLv2 is the
program, and:
Everyone is permitted to copy and distribute verbatim copies
of this license document, but changing it is not allowed.
is the license. If you couldn't change just this bit (and the FSF
copyright) then things would not conflict. But this very license says
you can't change _the program_ (the license document). This conflicts.
> This can, btw, also be shown independently by the fact that the FSF
> clearly _intended_ the license to be actively linked into the
> program: they ask you (in the "How to Apply These Terms to Your New
> Programs") to have commands to view parts of the license if your
> program is interactive.
This, in fact, seems to be a good point. This one wants an FSF lawyer.
> So. Claiming that the GPL license text itself cannot be part of the
> program is disingenious. According to your reading, the modified BSD
> license wouldn't be compatible with the GPL either, because it requires:
>
> * 1. Redistributions of source code must retain the above copyright
> * notice, this list of conditions, and the following disclaimer,
> * without modification.
>
> yet the FSF has clearly stated that this is perfectly fine, even though it
> also disallows modifications to the license text.
But please note that this is indeed perfectly fine, and does not
contradict anything I say. This requirement only disallows modifications
to the copright and license notices, not modifications of the
_program_ and the GPL is perfectly fine with that.
This is the point I've tried to make a number of times now. The program
is not the same thing as the license and allowing or disallowing
something being done to the license is not the same thing as allowing or
disallowing something to be done to the program. And in the case of the
GPL license text, the license text _is_ the program. Not the license.
Note, we are arguing here over whether or not the GPL document itself is
GPL compatible. I feel it obvious that it is not, you do not agree. What
we _do_ agree on though is that it's wholy irrelevant for the kernel at
least since 2.4.0-test8.
Ever since then, the kernel as a whole has been V2, not a doubt about
it. You feel that even before that, it was V2 only but then you are
ignoring the very strong point Paul Jakma made and which I repeated: if
the "Version 2" in the GPL header were enough to have the program
specify a version, section 9 would be utterly useless, and as such it's
obvious that at very least the intent of the GPL authors here was that
it was _not_ enough.
Hey, you still replied, so you probably don't think I'm completely full
of it. Did you think there was any merit in the suggestion of seperating
the GPL and the NOTE into separate files, so as to leave even fewer room
for people trying to milk the argument?
Rene.
On Thu, 2 Feb 2006, Rene Herman wrote:
> > So. Claiming that the GPL license text itself cannot be part of the program
> > is disingenious. According to your reading, the modified BSD license
> > wouldn't be compatible with the GPL either, because it requires:
> >
> > * 1. Redistributions of source code must retain the above copyright
> > * notice, this list of conditions, and the following disclaimer,
> > * without modification.
> >
> > yet the FSF has clearly stated that this is perfectly fine, even though it
> > also disallows modifications to the license text.
>
> But please note that this is indeed perfectly fine, and does not contradict
> anything I say. This requirement only disallows modifications to the copright
> and license notices, not modifications of the _program_ and the GPL is
> perfectly fine with that.
You continue to make this totally arbitrary distinction between "copyright
and license notices" and the "program".
That distinction does not make any sense.
The "license notices" _are_ an integral part of the program. Nothing else
makes sense. They are often in the same files (that's the most common case
by far), and as I showed you, they are often (as in the case of the GPL
license notice itself) actually _linked_in_ into the actual binary itself.
I find absolutely zero legal ground for your distinction that they are
somehow a totally independent entity from "the program", and it is in
fact in total disagreement with GPLv2 section 9 where they talk about the
"Program" specifying a license version.
You seem to be saying that no copyright license or notice is ever part of
the "Program", but that's totally idiotic. It would mean that you can
_never_ specify a "license version" in the Program, because by your
totally inconsistent rules, that copyright license note would
automatically somehow magically not be part of "Program".
So your reading simply DOES NOT MAKE SENSE.
I claim that Copyright notices and licenses very much _are_ part of the
program. I claim that because (a) the license says nothing to the contrary
and (b) the license clearly _does_ talk of copyright notices within the
Program (since the "Program" can specify a license version as per section
9: how the hell could you specify a license version "in" the Program, if
all copyright license notices were somehow magically "outside" the
Program?)
You cannot have it both ways. You can't say that copyright license notes
are sometimes outside the Program, and sometimes part of the Program.
The _only_ consistent reading is that copyright notices are _always_ part
of the Program, and that you must leave them intact as per section 1. This
also makes it clear that the "you must be able to modify" (section 2) does
_not_ cover copyright notices and licenses, even though they _are_ part of
the Program.
See? My reading is perfectly self-consistent, and doesn't need any
artificial separation of "copyright notice" vs "Program". Yours is not.
Your notion of copyright notes and licenses somehow being separate from
"Program" simply cannot be reconciled with reality, common sense _or_ the
wording in section 9.
The fact is, the copyright notices and licenses really _are_ a very
integral part of "Program". You cannot separate one from the other - not
conceptually, and not physically - and trying to do so would in fact be a
clear violation of section 1.
Linus
> The "license notices" _are_ an integral part of the program. Nothing else
> makes sense. They are often in the same files (that's the most common case
> by far), and as I showed you, they are often (as in the case of the GPL
> license notice itself) actually _linked_in_ into the actual binary itself.
The COPYING file is not part of the program. I'm sorry you can't get
that into your brain but continually bleating that it doesn't make sense
won't actually get you anywhere. Please go and talk to a law student or
just let the subject drop.
> The fact is, the copyright notices and licenses really _are_ a very
> integral part of "Program". You cannot separate one from the other - not
> conceptually, and not physically - and trying to do so would in fact be a
> clear violation of section 1.
Is it your position that a program distribution that includes a copy of
GPLv2 and contains no other information about licensing is licensed *only*
under GPLv2 and you can not substitute a later version at your option?
It may help to review section 9:
"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."
Do you see in here *any* reference to a way to limit a program so that only
one particular version of the GPL applies to it?
In fact, the kernel's notice that you may only use v2 and no later version
of the GPL appears, to me, to be an additional restriction that is not
permitted by the GPL. That is, unless you can find a GPL section that
permits that type of restriction. I was unable to find one. None of the
examples included with the GPL include any such restriction of that type.
I previously wrote that I thought that you could do so by including a
notice to that effect in every file. I now do not believe that is the case.
You cannot include any other type of restriction in a file under the GPL,
and I don't see why that type of restriction would be an exception.
DS
On Wednesday 01 February 2006 18:29, Rene Herman wrote:
> There are two things here that need seperating. On the one hand, we have
> "the program". On the other we have "the license" (and license notices).
> The GPL requires you to be able to make changes to the program,
... Subject to the terms of Section 1. It says so explicitly:
"You may ... under the terms of Section 1 above, provided that you also ..."
The GPL requires you to be able to make changes to the program _except_
for changes that would not "keep intact all the notices that refer to this
License" (+ a couple of other exceptions). Editing the text of the GPL in
the COPYING file would certainly violate this. However, copying COPYING
to a new file (license_for_foo, say) and editing that copy wouldn't. The
GPL allows this, the GPL's license does not, ergo, in the final analysis
I think you are right that the GPL's license is not literally
GPL-compatible.
> > This can, btw, also be shown independently by the fact that the FSF
> > clearly _intended_ the license to be actively linked into the
> > program: they ask you (in the "How to Apply These Terms to Your New
> > Programs") to have commands to view parts of the license if your
> > program is interactive.
>
> This, in fact, seems to be a good point. This one wants an FSF lawyer.
There is a very strong implication there that the GPL can be incorporated
into GPLed code. I'd read it as allowing the GPL v 2 to be included in
GPLed source code notwithstanding Section 2. It'd be nice if that were
explicit and part of the Terms and Conditions, but the verbiage is part of
the License, and I suspect counts for something. But IANAL.
As a practical matter, even if the GPL is technically GPL-incompatible,
the chances of anyone objecting to their GPLed code rubbing shoulders
with the GPL is remote.
Andrew Wade
On Feb 02, 2006, at 01:39, Andrew James Wade wrote:
> ... Subject to the terms of Section 1. It says so explicitly:
> "You may ... under the terms of Section 1 above, provided that you
> also ..." The GPL requires you to be able to make changes to the
> program _except_ for changes that would not "keep intact all the
> notices that refer to this License" (+ a couple of other exceptions).
But see, even assuming the really odd case of a project consisting of
one file, the GPL, that project would be completely GPL compatible.
As the license specifies the licensing terms for the project (IE: the
GPL), it may not legally be modified _even_ _under_ _copyright_ _law_
(because it's the project license). As a result, it is that the GPL
document may also be GPL licensed (because the only restrictions
therein are automatically implied by copyright law in the first
place). The FSF also mentions in their "deriving from the GPL" page
that you may create any derivation you want, but you may not
reference the GPL or FSF in such a revised license. That doesn't
affect or infringe on the GPL either, because such misuse would be
trademark infringement (outside the GPL) and not copyright
infringement (covered by the GPL), and therefore still doesn't affect
the terms of the GPL itself.
Cheers,
Kyle Moffett
--
Somone asked me why I work on this free (http://www.fsf.org/
philosophy/) software stuff and not get a real job. Charles Schulz
had the best answer:
"Why do musicians compose symphonies and poets write poems? They do
it because life wouldn't have any meaning for them if they didn't.
That's why I draw cartoons. It's my life."
-- Charles Schulz
Kyle Moffet wrote:
> But see, even assuming the really odd case of a project consisting of
> one file, the GPL, that project would be completely GPL compatible.
> As the license specifies the licensing terms for the project (IE: the
> GPL), it may not legally be modified _even_ _under_ _copyright_ _law_
> (because it's the project license). As a result, it is that the GPL
> document may also be GPL licensed (because the only restrictions
> therein are automatically implied by copyright law in the first
> place).
That's just not true. There is no reason under copyright law why the author
of a program could not modify its license. However, the GPL explicitly
prohibits *anyone* from modifying it.
The reason you can't modify the GPL, even we assume the GPL is licensed
under the GPL, is because the GPL says you can't modify the GPL.
Andrew Wade wrote:
> As a practical matter, even if the GPL is technically GPL-incompatible,
> the chances of anyone objecting to their GPLed code rubbing shoulders
> with the GPL is remote.
It is logically impossible for the GPL to be GPL-incompatible. To be
GPL-incomptabile, a license would have to contain requirements or
restrictions not found in the GPL. How could the GPL possibly do that?
DS
Linus Torvalds wrote:
> That distinction does not make any sense.
I'm afraid I have by now run out of steam. I you still feel the subject
interesting, please consult a lawyer.
Rene.
"David Schwartz" <[email protected]> writes:
> I don't think you can lock a person down to a single version of the GPL
> unless you explicitly say so in each individual file. Otherwise, nothing
> prevents someone from separating out one file and distributing it, kicking
> in clause 9's rule about a Program that does not specify a version number.
Nope. If you don't have a licence to use that single file, you can't
use it (distribute etc.) at all. So you basically choose between that
single version of GPL or nothing.
--
Krzysztof Halasa
Jeff V. Merkey wrote:
> Got it. Agreed. There was a COPYING file inside the tar.gz (you need to
> check this one)
> I recall seeing but I think it was 2.4 and some header files that still
> said this -- better clean those up.
You can't really do this part. If you make changes to the GPL, you're not
allowed to call it the GPL anymore... (See the FAQ)
--
[Name ] :: [Matan I. Peled ]
[Location ] :: [Israel ]
[Public Key] :: [0xD6F42CA5 ]
[Keyserver ] :: [keyserver.kjsl.com]
encrypted/signed plain text preferred
On Thursday 02 February 2006 01:57, Kyle Moffett wrote:
> But see, even assuming the really odd case of a project consisting of
> one file, the GPL, that project would be completely GPL compatible.
> As the license specifies the licensing terms for the project (IE: the
> GPL), it may not legally be modified _even_ _under_ _copyright_ _law_
> (because it's the project license).
Well ok, but modified copies of it (that aren't the license) could be
made. For the sake of argument, imaging copying the preamble into
another file and adding a few paragraphs there, with a prominent notice of
your modifications. This is allowed by the GPL (the License is the same,
and you're not violating the other terms either as far as I can see). But
it is not allowed by the license for the text of the GPL. Hence the
problem. It's an odd squirrelly corner case of the sort that cause licenses
to bloat up with verbiage.
The out is if the license file is not a portion of the program, in which
case section 2 of the GPL doesn't apply. I think this is what Alan Cox
is arguing. However, while it would be convenient if the License text
wasn't considered part of the program, I'm not sure that reasoning
would hold up in court.
It's also a complete non-issue: both because no-one is going to actually
object to the text of the GPL being there, and because it can be replaced
with a simple link in the extremely unlikely event that the need arises.
Andrew Wade
On Thursday 02 February 2006 02:11, David Schwartz wrote:
> The reason you can't modify the GPL, even we assume the GPL is licensed
> under the GPL, is because the GPL says you can't modify the GPL.
So you're saying that the two lines that are the GPL's copyright
license, are also part of the terms and conditions of the GPL, and act in
that capacity as a restriction of what you can do under Section 2? That's
an interesting argument, I might agree.
> It is logically impossible for the GPL to be GPL-incompatible. To be
> GPL-incomptabile, a license would have to contain requirements or
> restrictions not found in the GPL. How could the GPL possibly do that?
It's not the GPL doing that, it's the GPL's license. But perhaps the
restrictions of the license are also restrictions of the GPL, they are part
of the text of that document after all.
Well enough pedantism for me.
Andrew Wade
> On Thursday 02 February 2006 02:11, David Schwartz wrote:
> > The reason you can't modify the GPL, even we assume the GPL
> > is licensed
> > under the GPL, is because the GPL says you can't modify the GPL.
> So you're saying that the two lines that are the GPL's copyright
> license, are also part of the terms and conditions of the GPL, and act in
> that capacity as a restriction of what you can do under Section 2? That's
> an interesting argument, I might agree.
Yes, the GPL is the license agreement for both the GPL and anything
licensed under the GPL.
> > It is logically impossible for the GPL to be GPL-incompatible. To be
> > GPL-incomptabile, a license would have to contain requirements or
> > restrictions not found in the GPL. How could the GPL possibly do that?
> It's not the GPL doing that, it's the GPL's license. But perhaps the
> restrictions of the license are also restrictions of the GPL,
> they are part
> of the text of that document after all.
I have no idea what you mean by the "GPL's license". If it's contained in
the GPL, it's part of the GPL. The GPL is all of a piece.
DS
> Well ok, but modified copies of it (that aren't the license) could be
> made. For the sake of argument, imaging copying the preamble into
> another file and adding a few paragraphs there, with a prominent notice of
> your modifications. This is allowed by the GPL (the License is the same,
> and you're not violating the other terms either as far as I can see). But
> it is not allowed by the license for the text of the GPL. Hence the
> problem. It's an odd squirrelly corner case of the sort that
> cause licenses
> to bloat up with verbiage.
I totally don't understand why anyone thinks this is a problem. The GPL
restricts copying and modification in various ways. One of them is that you
can't modify the GPL itself. Another is that you can't remove license
notices or disclaimers of liability.
There is no problem and there is no conflict. The GPL is licensed under the
GPL, it's self-licensing. The GPL specifies the terms of use for the works
covered by it, including the GPL itself.
The GPL is not GPL-incompatible because something can only be
GPL-incompatible if it contains restrictions not found in the GPL. It is
logically impossible for the GPL to do so.
DS
On Friday 03 February 2006 00:35, David Schwartz wrote:
> I have no idea what you mean by the "GPL's license". If it's contained in
> the GPL, it's part of the GPL. The GPL is all of a piece.
"Everyone is permitted to copy and distribute verbatim copies
of this license document, but changing it is not allowed."
Those two lines look like a copyright license for the GPL. It makes sense
for the GPL to be licensed under the GPL, but it's less than obvious to
me that that is actually the case.
Andrew Wade