Re: Wasting our Freedom

From: Adrian Bunk
Date: Mon Sep 17 2007 - 09:38:46 EST


On Mon, Sep 17, 2007 at 11:20:19AM +0200, Hannah Schroeter wrote:
> Hi!

Hi Hannah!

> On Sun, Sep 16, 2007 at 11:13:51PM +0200, Adrian Bunk wrote:
> >On Sun, Sep 16, 2007 at 10:39:26PM +0200, Hannah Schroeter wrote:
> >> On Sun, Sep 16, 2007 at 09:59:09PM +0200, Adrian Bunk wrote:
> >> >On Sun, Sep 16, 2007 at 11:48:47AM -0700, Can E. Acar wrote:
> >> >>...
> >> >> First, these developers got questionable advice from senior Linux kernel
> >> >> developers, and SLFC (which is closely related to FSF) in the process.
>
> >> >The most questionable legal advice in this thread was by Theo de Raadt
> >> >who claimed choosing one licence for _dual-licenced_ code was illegal...
>
> >> JFTR, I do *not* think that that assessment was questionable. Unless the
> >> dual-licensing *explicitly* allows relicensing, relicensing is forbidden
> >> by copyright law. The dual-licensing allows relicensing only if that's
> >> *explicitly* stated, either in the statement offering the alternative, or
> >> in one of the licenses.
>
> >Dual licenced code by definition explicitely states that you can choose
> >the licence - otherwise it wouldn't be called dual-licenced.
>
> It does state you can choose which terms to follow, indeed, of course.
> But that does *not* imply removing the other terms altogether.

On which legal grounds do you base this statement?

And if you choose the GPL the code you distribute will be under the GPL
*only* forever [1], so what value would be in shipping terms that are
void?

And if the author intended to have the BSD licence text kept intact when
his code gets incorporated into GPL'ed code, why didn't he simply make
his code BSD-only? In fact the only difference between BSD-only code and
BSD/GPL dual-licenced code is that you can't remove the BSD licence text
for the former when incorporating it into GPL'ed code...

> >> Neither GPL nor BSD/ISC allow relicensing in their well-known wordings.
>
> >Noone said otherwise.
>
> Removing the terms you choose not to follow in one instance *is*
> relicensing.

If anything can be called relicencing, then the act of choosing one of
the licence. And this happens one level above the actual licences, and
the licence texts don't matter for this act.

> >> If you think that's questionable, you should at least provide arguments
> >> (and be ready to have your interpretation of the law and the licenses
> >> tested before court).
>
> >The licence in question was:
>
> ><-- snip -->
>
> >/*-
> > * Copyright (c) 2002-2004 Sam Leffler, Errno Consulting
> > * All rights reserved.
> > *
> > * Redistribution and use in source and binary forms, with or without
> > * modification, are permitted provided that the following conditions
> > * are met:
> > * 1. Redistributions of source code must retain the above copyright
> > * notice, this list of conditions and the following disclaimer,
> > * without modification.
> > * 2. Redistributions in binary form must reproduce at minimum a disclaimer
> > * similar to the "NO WARRANTY" disclaimer below ("Disclaimer") and any
> > * redistribution must be conditioned upon including a substantially
> > * similar Disclaimer requirement for further binary redistribution.
> > * 3. Neither the names of the above-listed copyright holders nor the names
> > * of any contributors may be used to endorse or promote products derived
> > * from this software without specific prior written permission.
> > *
> > * Alternatively, this software may be distributed under the terms of the
> > * GNU General Public License ("GPL") version 2 as published by the Free
> > * Software Foundation.
> > *
> > * NO WARRANTY
> > * ...
>
> ><-- snip -->
>
> >Theo claimed it would "break the law" [1] to choose the GPL for
> >_this_ code. [2]
>
> I re-read Theo's mail and still think the factual issues Theo states are
> probably right. Value judgements like "you should give code back" (when
> the license doesn't require it) are of course debatable (I tend to agree
> with Theo there too, but it's no mandatory requirement of course).
>
> Theo did *not* claim it breaks the law if you choose to obey by the
> terms of the GPL in said dual-licensing. Theo *did* claim (in my eyes,
> probably rightfully, and if it should ever be needed with respect to
> code related to OpenBSD, I could try to give a few bucks in support of
> having that claim legally verified) it's illegal to remove the license
> you chose to not follow in one instance of redistribution. IIRC the
> softwarefreedom.org people involved agreed with Theo's assessment in
> that instance.

You confuse two completely different situations.

The SFLC talks about how to incorporate *not* dual-licenced
BSD-only code into GPL'ed code.

> >[...]
>
> >> But the BSDl does not allow you to relicense the original code, even
> >> while it allows you to license copyrightable additions/modifications
> >> under different terms with few restrictions.
>
> >> However, you say "regarding ethics" and just go back to the legal level.
> >> Is it really ethical, if you consider both Linux and OpenBSD part of one
> >> OSS "community", to share things only in one direction? To take the
> >> reverse engineered HAL but to not allow OpenBSD to take some
> >> modifications back?
>
> >Is it really ethical to use a licence that does not require to give
> >back, but then demand that something has to be given back?
>
> IMO Theo didn't demand (as in try to enforce with legal pressure), but
> state it'd be the *morally* right thing to do even if *not* legally
> required (which isn't debated).
>
> >Why don't you use a licence that expresses your intentions in a legally
> >binding way?
>
> Because BSD people don't want to enforce it in every thinkable case. And
> BSD people don't want to enforce it using as much text as the GPL needs.
>
> But still I think it'd be the (morally!) right thing to do with respect
> to the Atheros HAL even if *not* legally bound to do so.

It's not about lazyness of BSD developers, many people who consider the
BSD licence more free than the GPL argue that the advantage of the BSD
licence is that it does not require you to give back.

Something is wrong if your licence text clearly states that you do not
require getting anything back but you then argue on moral grounds that
something has to be given back.

The majority of Linux developers might have never read the complete GPL
text, and the same might be true for some licence expressing the
intentions of the OpenBSD developers. But it would document in a legal
binding way the intentions of the OpenBSD developers.

> >[...]
>
> >But the truth is a bit less harsh:
>
> >In reality most Linux kernel developers might not mind to give back -
> >and e.g. much of the ACPI code is BSD/GPL dual-licenced, and there
> >doesn't seem to be any problem with this.
>
> *nods* Why not the same for the Atheros code?

I don't know the background of this one, but besides the already
expressed point that it wouldn't bring you much for technical reasons
(the Linux code will quickly switch to Linux conventions), there might
be better ways for starting a collaboration than Theo's rants.

> >But Theo's wrong accusations regarding dual licenced code might not be
> >the best way for starting a fruitful collaboration...
>
> As said above, the accusations, if you read them correctly, were not
> wrong, but spot on right. Unless someone proves that dual-licensing as
> in "you may follow terms A or terms B at your choice" implicitly implies
> being allowed to remove A altogether should you choose B.

As said above, keeping void terms neither makes sense, nor does there
seem to be any legal requirement.

> >[...]
>
> Kind regards,
>
> Hannah.

cu
Adrian

[1] the original code is of course still dual licence

--

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

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/