On Fri, Aug 20, 2004 at 10:10:20PM +0100, Matthew Garrett wrote: > > (To be clear, patch clauses are explicitly free, for obvious reasons--though > > as I've said I'd like that to change. I think "you must patch, *and* you > > must permit me to incorporate your patches" goes beyond the DFSG exception.) > > Right, but why?
Why does it go beyond the DFSG exception? Well, that much is probably obvious: DFSG#4 does not say "... and may allow the upstream author to require that he be allowed to incorporate all such patches" any more than it says "... and may require that the patches be unified diffs". That's possibly not what you meant to ask, though. I have a basic problem with all patch clauses: they make forking extremely difficult, and I believe ability to fork is a basic, founding principle of free software. Further, they make code reuse almost impossible, which is also a fundamental principle of free software. That's why I believe that portion of DFSG#4 is a huge error, and should be corrected. In my opinion (not the DFSG's, at present, per the exception), patch clauses are not free at all. As a result, I have a difficult time arguing the topic "are patch clauses free with an added 'you must allow incorporation' requirement?", because my mind is shouting "but patch clauses aren't even free on their own!" I'm not going to argue that patch clauses fail the DFSG, because there's an explicit exception for them; but I don't want to see patch clauses that are even stronger than the DFSG#4 exception being allowed. I do think all patch clauses would fail the DFSG if they weren't excepted by DFSG#4. So, that's a more basic reason why I, personally, find these clauses unfree: patch clauses themselves are only permitted due to an exception, and this is a super-patch clause that goes even further. (As for the give-me-a-license-that-I-didn't-even-give-to-you issue, I'm not sure. I need to think about it more.) -- Glenn Maynard
