[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <alpine.LFD.0.98.0706150748180.14121@woody.linux-foundation.org>
Date: Fri, 15 Jun 2007 08:21:28 -0700 (PDT)
From: Linus Torvalds <torvalds@...ux-foundation.org>
To: Carlo Wood <carlo@...noe.com>
cc: Alexandre Oliva <aoliva@...hat.com>,
Jeremy Maitin-Shepard <jbms@....edu>,
Sean <seanlkml@...patico.ca>, Adrian Bunk <bunk@...sta.de>,
Valdis.Kletnieks@...edu, Daniel Hazelton <dhazelton@...er.net>,
Alan Cox <alan@...rguk.ukuu.org.uk>, Greg KH <greg@...ah.com>,
debian developer <debiandev@...il.com>, david@...g.hm,
Tarkan Erimer <tarkan@...one.net.tr>,
linux-kernel@...r.kernel.org,
Andrew Morton <akpm@...ux-foundation.org>, mingo@...e.hu
Subject: Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3
On Fri, 15 Jun 2007, Carlo Wood wrote:
>
> Therefore, it seems pretty weird to me that LONG before version 3
> was written and released, someone purposely would choose to freeze
> their software at version 2. Why make it "impossible" to use this
> safe-guard?
Because what you call "safe-guard" is just call "idiotic".
The fact is, anybody who cares about the copyright license he uses on
software (and I sure do care!) should *never* give a blank check to
somebody else that they don't trust to change that license for them.
In other words, the whole "v2 or later" language *only* makes sense if:
- You are the FSF, and you *decide* what "or later" actually means
In this case, you obviously don't have to trust "somebody else". You
just trust yourself.
- You don't care about your choice of license.
In this case, you might as well let somebody else make that decision
for you, although quite frankly, you might as well use something like
the BSD license and let many *more* people make that decision to
relicense for you.
- ..or you trust the FSF implicitly.
In this case, you're not an independent entity, you're just a lackey of
the FSF when it comes to the license.
If none of those conditions are true, you'd be *incompetent* to leave the
"v2 or later".
And I'm not incompetent. None of the above conditions holds true for me,
so the "v2 or later" would be totally idiotic. Which is why Linux has
*never* had that statement for *any* code I have ever written!
So the fact that *you* (and the FSF) call it a "safe-guard" by no means
makes it so.
> I never knew it was possible to change the "version 2 or higher" into
> "only version 2", but I am not a laywer and not into licenses at all,
> and I am sure Linus had laywers look into this, so we can take this
> as a fact.
What you state above is that you didn't even *READ* the GPL, yet you make
some arguments about it as if you had!
So let me educate you. You may not want to be educated, but I'll try
anyway:
- Please *read* the license that you discuss. If you don't read it,
there's no point in you making any arguments about it.
And by reading it, I don't mean just mechanical "reading" in a
technical sense. I mean "read and think about it"
- Once you have actually read it, you'll notice that there *is* no
"either version 2 of the License, or (at your option) any later
version" in the GPL v2 license AT ALL.
Really. That phrase DOES NOT EXIST in the GPLv2.
- That phrase exists *outside* the license, in the *suggestion* of how
you migth want to use it! Notice how that license comes *after* the
"END OF TERMS AND CONDITIONS" thing, and in the section that describes
how you might *use* the license.
In other words, you are a perfect example of somebody who argues about the
GPLv2 without ever having read or understood it!
I'm really tired of people claiming that the kernel license is a
"modified" GPLv2 to remove the "v2 or later" language. It's *not* modified
at all: that language simply DOES NOT EXIST in the GPLv2. It exists
outside the license, in a suggestion on how you _might_ use it, and it's a
suggestion that my code has *never* *ever* actually followed.
>From the very first version of Linux that used the GPLv2 (that would be
Linux-0.95, iirc), Linux has *always* been under a unmodified GPLv2. Not
"any later". Not "modified". It has *always* stated the version
explicitly (by *including* it verbatim - there was certainly never any
ambiguity about versions), and it has *never* stated "v2 or later" in any
source code I wrote!
Yes, afterwards, I have added clarifications, because it turns out that
lawyers (and non-lawyers, for that matter) actually prefer having certain
things clarified and written out, so that nobody can later claim that they
misunderstood certain things.
But even those clarifications NEVER MODIFIED THE ACTUAL LICENSE. They are
literally just clarifications. The license didn't change.
The clarifications are
- the obvious *fact* that user-mode code using standard system calls is
not a derived work of the kernel.
This is so obvious to anybody with any knowledge of what "derived work"
means that there is absolutely no legal question about this being a
fact, but the FSF had made such a name for the GPL being "viral" that
some people were still worried that I was crazy and would sue. They
probably weren't actually worried that they'd lose, but a lawsuit is
potentially quite damaging (to PR and in just legal expenses) even if
you do end winning, so to make it clear that I wasn't crazy, I added a
clarification about this very early on.
Side note: while it's just a clarification, and as such doesn't
actually alter the license in *any* way, it's *also* a "public
statement of intent", and as such has legal meaning: it means that if I
were to stop taking my meds and go crazy, I *still* couldn't sue,
because of the so-called "doctrine of lathes" (and probably other
legal issues), since I have made it clear that you cannot be *expected*
to be sued over it.
So please do realize that while a clarification does not change the
license in any way, shape, or form, it does actually have some real
legal implications as to whether you can effectively sue somebody. You
cannot first claim that certain uses are ok, and then suddenly change
your mind when people have made decisions based on your claim.
- the above *fact* that the source code never followed the FSF suggestion
of "v2 or later" was clarified and made explicit.
Again, this in no way changed the license itself (that "v2 or later"
language simply _does_not_exist_ in the GPLv2), but it *also* didn't
actually change the licensing of the code itself. Because the whole "v2
or later" language (in the "how to use") only takes effect
*if*it*exists* in the source code, and the whole "any version of your
choice" language (which *does* exist in the GPLv2) only takes effect if
there wasn't some explicit license version that was obvious (and
dammit, it doesn't get much more obvious than a COPYING file that
includes the whole license *verbatim*)
So I am 100% sure that the whole Linux kernel is "v2 only". It has
*always* been "v2 only" since very early in 1992, and before that, it
wasn't GPL'd at all.
Certain individual *parts* of the kernel can be used under other licenses
(BSD, Mozilla, and yes, GPLv3+), but that is very much a localized issue,
and does not in any way change the fact that the Linux kernel itself in
its entirety is GPL version *2*. And nothing else.
And I'm also 100% sure that anybody who claims otherwise is totally
ignorant of either copyright law, the GPLv2, or *both*.
Linus
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists