Kernel developers' position on GPLv3

Posted Sep 23, 2006 8:58 UTC (Sat) by timtas (guest, #2815)
Parent article: Kernel developers' position on GPLv3

I too think that, if you replace GPLv2 with BSD and GPLv2 with GPL, this position might well be put on any randon BSD advocacy list. GPL always had the spirit of "you may use it, but you have to play fair", while BSD rather goes for "do what you want, we don't care about the freedom of the end user, we just want as many people as possible to use our work, thieves included".

While I don't agree with this position, I still understand people thinking that way and don't think it is fundamentally wrong to do so. I just wonder why Greg and all like the GPLv2 in the first place. Maybe because they have to, as Linux is GPLv2?

What I really find appalling are all the FUD threats, like "it would potentially jeopardise the entire patent portfolio of a company simply by the act of placing a GPLv3 licensed programme on their website." This is a totally speculative exaggeration out of any proportion. I'm sure you are well aware of that, since you wrap this in the usual "well, some lawyer may see it this way".

You might not have been forced by your employers to write this, but I'm sure that IBM and Novell will really like you a lot for this piece of work.

Tim

to post comments

Kernel developers' position on GPLv3

Posted Sep 23, 2006 19:25 UTC (Sat) by sepreece (guest, #19270) [Link] (1 responses)

What you seem to be missing is that they still support GPLv2. The complaint is that GPLv3 asks for more restrictions than GPLv2 did, and the kernel developers apparently think they go too far.

As to the questions about the patent clause, I don't think it's FUD. I don't think anyone but a lawyer could tell you how serious the danger would be, but there definitely is some risk there.

I doubt that any of the kernel developers would want to apologize for hte fact that their employers are sinking millions of dollars into supporting and developing Linux and other OSS projects. You seem to be unwilling to admit that anyone could possibly just disagree with you...

Kernel developers' position on GPLv3

Posted Sep 26, 2006 13:47 UTC (Tue) by timtas (guest, #2815) [Link]

I have not missed that they still support v2, I just wondered why...

> I doubt that any of the kernel developers would want to apologize for hte
> fact that their employers are sinking millions of dollars into supporting
> and developing Linux and other OSS projects.

Sorry, but I don't quite get this sentence. I have no problem that they work for companies supporting Linux, I even have no problem with those companies at all. I only doubted they speak for themselves, because their language and their arguments sounded very much like business talk and not like personal opinions on things.

> You seem to be unwilling to admit that anyone could possibly just
> disagree with you...

Oh, really? Would I participate in this discussion, then? A lot of people disagree with the current draft of GPLv3, I have no problem with that at all. I'm not even suggesting in the least the kernel should move to GPLv3. What I didn't like was the tone of the position and generally that people put their "own" position while constantly referring to companies, distributors and all.

I think the companies should speak for themeselves in that matter, their view is certainly valid to consider.

Kernel developers' position on GPLv3

Posted Sep 23, 2006 21:44 UTC (Sat) by sepreece (guest, #19270) [Link] (1 responses)

Here's what I think is of concern to people who hold patents: "If you convey a covered work, you similarly covenant to all recipients, including recipients of works based on the covered work, not to assert any of your essential patent claims in the covered work."

Note that you are giving this covenant for a work based on the work you created, without knowing up-front what is in that derived work. There is nothing here that says that work could not infringe on patents you hold that were not used in the work that you released.

That is, you release work A, granting a covenant to allow downstream users to use any patented technology in that work. Downstream user creates a derivative work AB that adds to it an extension that infringes a patent that you hold, but that was not used in the work A. The current language seems to say that you are giving away the right to use ANY patents you hold that anyone uses in such a derived work.

[And, yes, I have filed a version of this at gplv3.fsf.org]

The question is how a court would read "the covered work" - whether it would take it to mean the original covered work or any downstream modification based on the original work. Most people would read it to mean the original work, but there is at least a modicum of ambiguity, and that's where lawyers live. If they mean it narrowly, they should rewrite it to make the language clearer.

Kernel developers' position on GPLv3

Posted Sep 26, 2006 1:44 UTC (Tue) by sanjoy (subscriber, #5026) [Link]

As you quoted from the GPLv3:

If you convey a covered work, you similarly covenant to all recipients, including recipients of works based on the covered work, not to assert any of your essential patent claims in the covered work.

This language answers your objection, I believe. Here's how. Suppose that you distribute a modified version of GNU diff v8 and GNU diff v8 uses one of your patent claims. Then, yes, you've given up the right to assert that patent claim against a derivative work of GNU diff v8. And that's fair: You shouldn't be able to sue your recipients over what you distribute. You also give up the right to sue them for using your patent in GNU diff v8a, a derivative work of GNU diff v8.

But now one of the recipients makes GNU diff v9 by implementing a tricky compression method on which you hold patent claims. You can still sue him or her because you promised only "not to assert any of your essential patent claims in the covered work" (emphasis mine). You did not make the broader promise: "I will not assert any of my essential patent claims in the covered work or in any of its downstream versions."