Why Torvalds is sitting out the GPLv3 process (Linux.com) [LWN.net]

Why Torvalds is sitting out the GPLv3 process (Linux.com)

Posted Sep 26, 2006 22:49 UTC (Tue) by Zack (guest, #37335)
In reply to: Why Torvalds is sitting out the GPLv3 process (Linux.com) by mingo
Parent article: Why Torvalds is sitting out the GPLv3 process (Linux.com)

>I for one dont trust Richard Stallman anymore - and apparently Linus neither. And picking the next version of the GPL is very much trust-based for me.

That's why it's *you* who picks the next version of the GPL for your software, as it specifically says "(at your option)" in:
"either version 2 of the License, or (at your option) any later version."

There is no need to put blind trust in rms as you can always choose the last version of the software that is licensed under a version of the GPL you find truly free.
And, should most people share your opinion, the hypothetical evil version of the GPL will be a stillborn.

Regardless, the GPLv3 is of no concern anymore to most of the leading kernel developers and it is already abundantly clear they will not adopt it even if they could, so why is there this interest of the linux kernel developers to put rms, the fsf and the GPLv3(process) in a bad daylight ?

I understand there is some rivalry between the Open Source and Free Software "camps", but some of the public statements towards the GPLv3 process are downright mean-spirited and/or ill-informed at times and serve no obvious purpose.

to post comments

Why Torvalds is sitting out the GPLv3 process (Linux.com)

Posted Sep 26, 2006 23:01 UTC (Tue) by mingo (subscriber, #31122) [Link] (2 responses)

It's quite simple: we very much care about free software in general.

We'd also like you to understand our position and show you what the basis of our position is.

Why Torvalds is sitting out the GPLv3 process (Linux.com)

Posted Sep 27, 2006 13:31 UTC (Wed) by dmantione (guest, #4640) [Link] (1 responses)

With all respect, but it doesn't look you are acting wisely. Wether you
like Stalman or not, in general, the GPL3 is a big improvement. If you
don't like the DRM stuff, you can always grant an exception. (Allthough
forbidding Tivoisation, just one of the intended improvements, is a wise
decision IMHO.)

Why Torvalds is sitting out the GPLv3 process (Linux.com)

Posted Sep 28, 2006 21:48 UTC (Thu) by vonbrand (guest, #4458) [Link]

"Grant an exception" that anyone can just strip out whenever they like is of no use. Plus the whole "exceptions" and "extra rights" nonsense creates hundreds of non-compatible licenses, for no good reason at all.

Why Torvalds is sitting out the GPLv3 process (Linux.com)

Posted Sep 26, 2006 23:17 UTC (Tue) by emkey (guest, #144) [Link]

but some of the public statements towards the GPLv3 process are downright mean-spirited and/or ill-informed at times and serve no obvious purpose.

All I can say is what goes around comes around. Many of us do not believe in one true way, even if we are very much believers in free software. I've been on the receiving end of about a billionth of a percent of the vitriol that Linus and others have encountered and I'll tell you I'm very tired of it. I'm surprised that Linus isn't a heck of a lot more negative given his experiences.

GPLv3 & trust?

Posted Sep 26, 2006 23:40 UTC (Tue) by mingo (subscriber, #31122) [Link] (5 responses)

That's why it's *you* who picks the next version of the GPL for your software, as it specifically says "(at your option)" in: "either version 2 of the License, or (at your option) any later version."

You are missing a very obvious thing: tens of thousands of developers have contributed under "GPLv2 or later" projects, assuming and trusting that the FSF promise that later GPL versions would be "similar in spirit" would be upheld. Furthermore, hundreds of people have contributed to the "GNU" codebase and have assigned copyrights to the FSF - again under an implicit trust attached to it.

So if, _I_, a staunch and long-time free software and GPL supporter (see this link and search for occurances of "GPL" in the article) am having doubts about the GPLv3, and if a majority of top kernel developers are having doubts about it, how many other developers have doubts about it, and isnt there a danger that their trust would be abused, if things dont change and the FSF relicenses the GNU codebase to GPLv3 and if "GPLv2 or later" projects become automatically licensed by the GPLv3 too on the day the GPLv3 is released?

You dont seem to be willing to stand up for their moral rights (in fact you dont even seem to accept that _I_ am in honest disagreement, although i wrote about my reasons in many comments) - but i think they should not be forgotten either. They wrote real free software and thus made this world richer.

Trust what they promise, not what you wish they promised

Posted Sep 27, 2006 0:27 UTC (Wed) by GreyWizard (guest, #1026) [Link]

I can respect that the "no technical restrictions" clause is not to your liking, that you probably won't adopt GPLv3 for software that you write and that your goals are not the same as those of the Free Software Foundation. However the Free Software Foundation has clearly explained their position in numerous documents published over many years so your claims about betrayed trust are ignorant at best. http://www.gnu.org/philosophy/ Have you read any of that? How can you honestly claim that the GPLv3 draft is something other than an extention of that position (whether or not you agree)?

GPLv3 & trust?

Posted Sep 27, 2006 0:27 UTC (Wed) by Zack (guest, #37335) [Link]

>isnt there a danger that their trust would be abused, if things dont change and the FSF relicenses the GNU codebase to GPLv3 and if "GPLv2 or later" projects become automatically licensed by the GPLv3 too on the day the GPLv3 is released?

Basically, no. The worst that could happen is that you'll be stuck with GPLv2 forever.

Basically the developers you are worrying about are the same developers who are distributing their software under the GPL. If the GPLv3 comes around and everybody hates it, the developers of these projects would naturally fork.
This remains even true if the copyrights are assigned to the FSF. Even though they would be the sole copyright holder of a certain program A that was licensed under "GPLv3 or higher" it would be a stale project if it had no developers left and last free version would become the main tree.

The only catch would be that the FSF could always incorporate the changes from the now-main "GPLv2 or higher" into their hypothetical onerous (or even proprietary) "GPLv3 or higher" version. But with no developers left and all goodwill destroyed it would quickly become a futile effort.
This would be in a worst case scenario.

The "at your option" is there for a reason. And that reason is that you should never have to blindly rely on any person or any institution, no matter how benevolent they may seem at the moment.
It's there to mitigate the worst case scenario and enable GPL-licensed Free Software to survive under very hostile conditions.

GPLv3 & trust?

Posted Sep 27, 2006 0:38 UTC (Wed) by pinky0x51 (guest, #40742) [Link]

>You are missing a very obvious thing: tens of thousands of developers have contributed under "GPLv2 or later" projects, assuming and trusting that the FSF promise that later GPL versions would be "similar in spirit" would be upheld.

Where is the problem? The spirit is still the same.

Look at the GNU Manifesto, read the GNU philosophy or just the preamble of the GPL.

The spirit was always the everyone who gets software which is licensed under the GPL should have the right to (0) use it for any purpose, (1) study and modify it (and of course than go back to freedom (0)), (2) share it and (3) copy modified versions. If you want to remove this rights through license-changing, technology (DRM), software patents or through any other technology or law wich could appear in the future than you act against the spirit of the GPL.
GPLv2 protected this freedoms really good in the age before DRM and software patents, GPLv3 will protect this freedoms really good in the age of software patents and DRM and i hope GPLv4 will protect this freedoms in the future if new technology and/or laws appear to remove this freedoms.

As long as the GPL (never mind in which version) will secure this freedom the GPL is in the spirit of it predecessors.

GPLv3 & trust?

Posted Sep 27, 2006 7:06 UTC (Wed) by job (guest, #670) [Link]

How could "GPLv2 or later" become "automatically licensed by the GPLv3"?
Who has the right to relicense your code?

And how would that be possible, without redefining basic language words
such as "or"? (Which by the way would make programming a bit
difficult! ;) )

These fears borders to irrational bashing. No one except you has the
right to relicense your software, not now and not with the GPLv3.

GPLv3 & trust?

Posted Oct 7, 2006 23:22 UTC (Sat) by anton (subscriber, #25547) [Link]

I am one of those "people have contributed to the "GNU" codebase and
have assigned copyrights to the FSF". I don't feel that the FSF is
betraying my trust with the GPLv3 process, and the GPLv3 looks to me
like it will be a fine license. But even if I did not like it, there
would be various alternatives for me; they would not be pretty, but
they exist.

So, thank you, Ingo, for so valiantly standing up for my defense, but
that is really not needed.

Why Torvalds is sitting out the GPLv3 process (Linux.com)

Posted Sep 27, 2006 1:02 UTC (Wed) by k8to (guest, #15413) [Link] (2 responses)

> > I for one dont trust Richard Stallman anymore - and apparently Linus
> > neither. And picking the next version of the GPL is very much trust-based
> > for me.

> That's why it's *you* who picks the next version of the GPL for
> your software, as it specifically says "(at your option)" in:
> "either version 2 of the License, or (at your option) any later version."

This kind of misses the issue. If I write some code under the GPLv2 (or later), and then GPLvLater comes out and changes the rules of the road in very significant ways that I do not accept, it's not _my_ choice as to whether people license y code under GPLvLater, anyone can apply this new license even if I find it totally unconscionable. I can of course continue to license changes to my code under GPLv2 (or later), but the unconscionable license may become the one generally used.

Now, I'm not claiming (in this post) anything in either direction about whether GPLv3 will substantially change the rules in objctionable ways, but the issue remains that changing it in ways that GPLv2(or later) coders do not appreciate would be a seriously unpleasant business for programmers who have done so.

Why Torvalds is sitting out the GPLv3 process (Linux.com)

Posted Sep 27, 2006 1:13 UTC (Wed) by mingo (subscriber, #31122) [Link] (1 responses)

This kind of misses the issue. If I write some code under the GPLv2 (or later), and then GPLvLater comes out and changes the rules of the road in very significant ways that I do not accept, it's not _my_ choice as to whether people license y code under GPLvLater, anyone can apply this new license even if I find it totally unconsciable. I can of course continue to license changes to my code under GPLv2 (or later), but the unconscionable license may become the one generally used.

Exactly. For example, i might become supportive of the GPLv3 later on if it's modified sufficiently - but still if i saw people (contributors) coherently arguing that it's unacceptable to them then i'd be uneasy to force it on them - because once introduced, the GPLv3 license is constructed in a way so that code would "gravitate" towards the one pure GPLv3 license. We really, really have to be careful and unify a large majority of current contributors under the GPLv3 umbrella. Saying that "oh it was all written into the GNU Manifesto, sorry" or "you can list extra permissions/restrictions" is not enough to keep a community unified.

Why Torvalds is sitting out the GPLv3 process (Linux.com)

Posted Sep 27, 2006 12:27 UTC (Wed) by robilad (guest, #27163) [Link]

That is exactly the same way with GPLv2, though.

You can, if you feel so inclined, strip the additional permissions to you granted in libgcc's license, or strip those granted to you for g++'s standard libraries, or even those in gcj.

You can even take LGPLd code, and convert it to GPLv2 licensed code, and accordingly turn the glibc into a GPLd work, or GNOME, or whatever LGPL licensed piece of code you find out there.

If there was merit to your claim, we would be seeing that LGPLd projects 'gravitate' towards GPLv2, and that GPLd projects strip off their exceptions, and all that.

I haven't seen any of that over the past years. In fact, libgmp went from GPL to LGPL with FSF's blessing. Ogg ended up being BSD licensed with FSF's blessing. Most of KDE libs, like the HTML component, are LGPLd and happily staying that way. And so on.

Why Torvalds is sitting out the GPLv3 process (Linux.com)

Posted Sep 28, 2006 3:29 UTC (Thu) by AJWM (guest, #15888) [Link]

> And, should most people share your opinion, the hypothetical evil version of the GPL will be a stillborn.

Well, no, the entire body of FSF code _will_ be moved to v3, barring some spectacularly unforseen circumstance.

Now true, if it's really that bad then that code could all be forked at the latest v2 version, but that'd be a major commitment. Hardly making v3 "stillborn".