Guidelines for GPLv3 Process Released

Justin Baugh writes "The Free Software Foundation and the Software Freedom Law Center have released a document detailing the guidelines and the process that will be used for revising the GNU GPL, and have launched a new website related to the V3 process. It was announced in a press release this morning that the FSF will be releasing the first discussion draft of the new license for comments at the International Public Conference for GPLv3 at MIT on January 16 and 17, 2006."

Comments

Do no harm?

[guysmilee]

Can some one explain the "Do no harm" bit? I am hoping it means GPL'd software can't be used in a weapon ... but i admit i just skim read ...

Re:Still a bit wary of one element of the GPL

[Ckwop]

A clause of "NAME OF FOUNDER OF PROJECT is free to upgrade this license to any future version of the GPL at his/her discretion" might be a better idea. This way, you CAN switch to new versions of the GPL even though you have thousands of contributors each with individual copyright on bits of the code, but you can also refuse to license the software under a future version of the GPL if it is not in your interest.

This doesn't work. Suppose I fork your project. Now I am the project leader for my fork. Should I be allowed to relicense everyone elses under any GPL version because I forked?

This is actually worse than the existing clause because that means I can relicense your code to any GPL version instead of just later versions.

I actually question why we need a GPL V3 in the first place. The current GPL works perfectly well and I don't see any reason to bother using the new license at all.

Simon

oooops...

[hummassa]

1. There is no clause obligating GPL licensers to license a work under the GPL version X "or any subsequent versions". As a matter of fact, most of the Linux kernel is licensed under the GPL v2 ONLY; arguably, as most of linux is a derivative work of the initial GPL v2 ONLY Torvalds job, they have to abide to the terms of the GPL v2.

2. It's a matter of trusting the FSF (which I don't (*)) or trusting [INSERT PERSON HERE]; my proposal is simpler: Trust the LAW. It's the only thing that (much like a computer) will do exactly as told to (and, much like a computer, you may be saying different from what you mean).

So, my suggestions will be (I'm signing in right now):

1. keep all rights given by the GPLv2, so GPLv3 will be compatible;
2. eliminate the "nonexistent linking clause", by:
    2.1. stating what is NOT TO BE CONSIDERED a derivative work;
    2.2. leaving all other definitions to the law.
3. don't try to mess with patents or trademarks because this would be incompatible with #1.
4. don't try to mess with choice of law or choice or venue because this would be incompatible with #1 AND non-free.
5. change the language so it's more comprehensible and leave no ambiguity behind.

Re:oooops...

[Misch]

Are you going to explain the asterisk in your statement It's a matter of trusting the FSF (which I don't (*)), or are you going to be like a certain credit card company that I've dealt with... (changing terms from "90 Days Same As Cash" to "90 Days Same As Cash*")

Re:Still a bit wary of one element of the GPL

[Tim C]

While I don't doubt what you say, and don't suppose for a moment that the FSF would ever go rogue, here's a little thought experiment - would it be possible to slightly alter section 9, perhaps over a number of revisions of the GPL, such that eventually the wording allowed for wholesale changes to the licence? As long as each revision is itself similar in spirit, they should be okay, and of course with each passing revision the amount of similarity required would drop.

Not saying they'd ever do this, just wondering if it would be possible.

Re:Still a bit wary of one element of the GPL

[Anonymous Coward]

But what happens when someone combines code from two GPL projects, one with the license upgrade rights held by Linus Torvalds, another with the same rights held by Steve Ballmer? What if you have code with 172 different license upgraders, each with rights to some parts of the code? Maybe everyone is still using GPL v2 when you start, but years later one guy upgrades all his code to GPL v3, and another to GPL v4. Now you've got a mix of GPL versions 2, 3 and 4, and maybe they aren't all quite compatible...

Tell me why oh why?

[mu22le]

What's wrong with version 2?
Why do we need v3?
Expecially since a lot of people seem scared by what could come out this time.

Sorry about that ...

[hummassa]

(*) MHO:
1. I respect RMS for his role in setting in motion the Free Software movement.
2. (corollary) I respect RMS for his role in starting GCC (which I use) and EMACS (which I don't use).
3. I respect the FSF and RMS for their roles in furthering the cause of Free software.
4. I don't trust the FSF and RMS to know that sometimes even well-intentioned people do things contrary to their own beliefs.
5. (corollary) The GFDL is non-free and sucks. Freedom should be equally applicable to all kinds of software: programs, libraries and other content types.
6. (corollary) The whole "GPL forbids (dyn)linking with GPL-incompatible works" party line is an outrageous lie, and it sucks. And, even if it seems to boost the FS movement (GPL'd Qt! Hey!) I think it does more harm than good.