----- Forwarded message from Bob Beck <[email protected]> -----
From: Bob Beck <[email protected]>
To: [email protected]
Subject: I respect the GPL immensely, really I do - but I believe this type of action weakens us all.
X-Spam-Checker-Version: SpamAssassin 3.1.8 (2007-02-13) on bofh.cns.ualberta.ca
X-Spam-Level:
X-Spam-Status: No, score=-2.8 required=3.0 tests=AWL,BAYES_00,
DNS_FROM_RFC_ABUSE,FORGED_RCVD_HELO autolearn=no version=3.1.8
Date: Sat, 1 Sep 2007 16:22:43 -0600
User-Agent: Mutt/1.5.12-2006-07-14
[ A copy of this is going to the linux kernel mailing list, regarding the
recent license modifications to reyk's files]
>Oh, and if you look at the OpenBSD CVS you see versions 4 months old
>with dozens of contributions by Reyk and with:
>
>/* $OpenBSD: ath.c,v 1.63 2007/05/09 16:41:14 reyk Exp $ */
>/* $NetBSD: ath.c,v 1.37 2004/08/18 21:59:39 dyoung Exp $ */
>/*-
> * Copyright (c) 2002-2004 Sam Leffler, Errno Consulting
> * All rights reserved.
Of course you do! because some of reyk's work used some of Sam's
work, and unlike what it seems a portion of the Linux community seems
to be willing to do in their Zealotry for the GPL, reyk is not
*removing and modifying* the licenses granted by the original authors.
That's the point. He's not saying he wrote this piece, and he's not
*changing* the conditions under which Sam distributed the code in the
first place. However what scares me more is the seeming willingness to
make the authors of a derivative work appear to be the primary authors
of something, and a willingness to change an authors copyright
statement (on reyk's code) without his permission.
I have always immensely respected the GPL - it has very noble
goals, they are very appropriate in some cases, they don't happen to
be mine, but that's fine, I don't release my code under it - but
that's fine, it's my choice. Just like many smart people who I know
and respect do their work in GPL land, and this is great too. However,
when it comes time to be looking at someone else's work above all we
have to respect the various authors choice of how they want their hard
work shared with the community.
To me, this seems like a portion of the Linux community seems to be
wanting to make their own rules, chosing to rewrite a license at any
time they choose without the original author's agreement. This appalls
and scares me. Why? not only does it show a huge lack of respect for
someone who has worked very hard to produce something the whole
community can use, but seriously undermines software freedom as a
whole. This is a slippery slope. If one community starts modifying the
others licenses for no purpose other than zealotry, I see only two
consequences:
1) a hugh rift of mistrust between the developers of both camps,
meaning no cooperating to make the world a better place.
2) A weakening of the respect for licensing on all sides of the
community, which weakens the credibility of both BSD *AND* the GPL
license when tested from the outside. Frankly, this scares the hell
out of me and dismays me.
I seriously hope that saner more mature and forward thinking heads
inside the Linux community will stop bashing the things that Theo and
the rest of our community is saying just because it's coming from
Theo, and he's a great target to bash, and start thinking about what
you are doing to free software as a whole. I think you are on the
verge of doing irreparable damage that will seriously weaken the
ability for all of our projects to move forward, and protect our
rights as code authors in the future.
-Bob
----- End forwarded message -----
On Sat, Sep 01, 2007 at 04:41:12PM -0600, Bob Beck wrote:
> ----- Forwarded message from Bob Beck <[email protected]> -----
>
> From: Bob Beck <[email protected]>
> To: [email protected]
> Subject: I respect the GPL immensely, really I do - but I believe this type of action weakens us all.
> X-Spam-Checker-Version: SpamAssassin 3.1.8 (2007-02-13) on bofh.cns.ualberta.ca
> X-Spam-Level:
> X-Spam-Status: No, score=-2.8 required=3.0 tests=AWL,BAYES_00,
> DNS_FROM_RFC_ABUSE,FORGED_RCVD_HELO autolearn=no version=3.1.8
> Date: Sat, 1 Sep 2007 16:22:43 -0600
> User-Agent: Mutt/1.5.12-2006-07-14
>
> [ A copy of this is going to the linux kernel mailing list, regarding the
> recent license modifications to reyk's files]
>
> >Oh, and if you look at the OpenBSD CVS you see versions 4 months old
> >with dozens of contributions by Reyk and with:
> >
> >/* $OpenBSD: ath.c,v 1.63 2007/05/09 16:41:14 reyk Exp $ */
> >/* $NetBSD: ath.c,v 1.37 2004/08/18 21:59:39 dyoung Exp $ */
> >/*-
> > * Copyright (c) 2002-2004 Sam Leffler, Errno Consulting
> > * All rights reserved.
>
> Of course you do! because some of reyk's work used some of Sam's
> work, and unlike what it seems a portion of the Linux community seems
> to be willing to do in their Zealotry for the GPL, reyk is not
> *removing and modifying* the licenses granted by the original authors.
> That's the point. He's not saying he wrote this piece, and he's not
> *changing* the conditions under which Sam distributed the code in the
> first place. However what scares me more is the seeming willingness to
> make the authors of a derivative work appear to be the primary authors
> of something, and a willingness to change an authors copyright
> statement (on reyk's code) without his permission.
>...
You miss the whole point of dual licencing:
Sam has stated in the licence that the code can be distributed under the
terms of the BSD licence, or alternatively it can be distributed under
the terms of the GPLv2.
Noone removed Sam's licence.
Sam has offered a choice, and if you choose one of the two offered
licences when distributing his code that complies with what he stated
in his copyright notice.
IANAL, but if reyk contributed to dual licenced code keeping the file
dual licenced it's hard to argue that he did not make the changes he
made available dual licenced.
> I seriously hope that saner more mature and forward thinking heads
> inside the Linux community will stop bashing the things that Theo and
> the rest of our community is saying just because it's coming from
> Theo, and he's a great target to bash, and start thinking about what
> you are doing to free software as a whole. I think you are on the
> verge of doing irreparable damage that will seriously weaken the
> ability for all of our projects to move forward, and protect our
> rights as code authors in the future.
The funny thing is that it seems Jiri's patch contained copyright
violations - and the parts of his patch Theo attacks are the parts
that are OK...
> -Bob
cu
Adrian
--
"Is there not promise of rain?" Ling Tan asked suddenly out
of the darkness. There had been need of rain for many days.
"Only a promise," Lao Er said.
Pearl S. Buck - Dragon Seed
> You miss the whole point of dual licencing:
>
> Sam has stated in the licence that the code can be distributed under the
> terms of the BSD licence, or alternatively it can be distributed under
> the terms of the GPLv2.
>
> Noone removed Sam's licence.
>
> Sam has offered a choice, and if you choose one of the two offered
> licences when distributing his code that complies with what he stated
> in his copyright notice.
>
> IANAL, but if reyk contributed to dual licenced code keeping the file
> dual licenced it's hard to argue that he did not make the changes he
> made available dual licenced.
I think that there needs to be some general understanding about what the policy has to be when dual-licensed or BSD-licensed code is incorporated into the Linux kernel. Unfortunately, I don't think it's practical keep dual-licensed or BSD-licensed code in the Linux kernel because it prevents GPLv2 code from being incorporated into that file.
For example, if I make a file 'foo.c' available under a dual license, and it winds up in the Linux kernel, what happens? Suppose someone decides my locking is not optimal, and they cut/paste a few lines of code from another file, GPLv2 only, into 'foo.c'. Did they just put someone's GPLv2 code under a dual license? Of course not.
I think the policy should be:
1) If BSD-only code is incorporated into the Linux kernel, the BSD license must not be removed, as this is prohibited by the BSD license. However, a note should be included that only the original work is offered under the BSD license and that the license does not apply to the new elements that may have been added to that file. The file, as it presently sits, is likely only offered under the GPLv2.
2) If dual-licensed code is added to the Linux kernel, one of two things has to happen.
A) If the code is relatively self-contained, it should stay dual-licensed. A warning should be placed in the file (with a pointer to a more detailed explanation) that GPLv2-only code cannot be spliced into this file without removing the dual-license.
B) If the code is not relatively self-contained, or can no longer stay as A above due to the desire to splice in code from other modules are add GPLv2-only code, the GPL license should be removed.
I honestly wish that it would be possible to offer more code back to the BSD-licensed versions. If you know who I am, you know I'm a strong critic of the GPL license and much prefer the BSD license. However, I think Linux as a project suffers if it includes files available under different licenses. It's much harder to work with the code, modify it, and contribute to it.
Almost any consistent license is better than different files being under different licenses such that you are burdened with compliance issues when you try to move improvements from one file into another.
Imagine if every file system were under a different license and someon adds a new file system hook that every filesystem benefits from implementing. Suppose the original author submits an implementation for ext2 and it's under GPLv2. Think about the burdens if you then tried to implement that hook for three other filesystems, each with their own license, and you can't cut/paste the ext2 hook in.
DS