On Sun, Feb 23, 2014 at 11:21:13AM +0100, Gerfried Fuchs wrote: > * Bas Wijnen <wijnen@debian.org> [2014-02-22 17:45:34 CET]: > > Well, yes, but there's nothing wrong with that. As I wrote, the > > question is not "are we in compliance with their license?", but rather > > "is their distribution in compliance with the DFSG?" > > And it is. As I wrote, for redeclipse I agree with that. But not because we shouldn't require source; only because I believe we have source. > > My point is that its license is irrelevant for our discussion. So yes, > > I ignore it. > > And you also ignore the fact that you apply a wording onto the DFSG > which isn't there. There is a very similar wording in https://www.debian.org/vote/2006/vote_004 which was accepted. Since then, "the form that the copyright holder or upstream developer would actually use for modification" is Debian's definition of source code. > > > The appropriate definition for the code at hand is its own license, > > > not one that "is a reasonable definition". > > > > The appropriate definition would be the one from the DFSG. Which > > unfortunately doesn't exist. > > That doesn't mean that you can make up your own definition then. I > rather would go with that it is intentional that way. No. According to you, if I write a program and publish it under a license where I define "source code" as the compiled binary data, and then go and distribute only that compiled binary, it is fine to go in main. You don't actually agree that would be the case, right? > > As I wrote, I don't mind using a different definition in principle, > > but I much prefer "reasonable" over "what the authors say". > > But "what the authors say" is what we must follow, not what you > yourself consider "reasonable". No, it isn't. Of course we must follow their license terms, and if they are not following the DFSG, then the software is non-free. But whether or not something follows the DFSG is a Debian-specific question. Upstream has nothing to do with it, other than being the subject of investigation. And to answer this question, we must use Debian's dictionary of what words mean. So what definition we must follow used to be "what the maintainer considers reasonable". It now is "what the GR says". It never was "what the authors say". And that's a very good thing. > > The separation between main and the rest is not based on upstream's > > rules; it's our own rules that count here. > > Our rules based on what upstream's rules we are given. We can't ignore > them. We cannot ignore their rules, but we don't have to allow them to change our rules by redefining our words. > > I think it is a good idea to require source for art, just as we do > > for code. > > But we don't. No, we really don't. True. But the GR I linked to (which is newer than the "editorial changes" one) does mention that we strongly recommend it. So nobody will stop us if we throw blobs of binary data into the archive, as long as it isn't a programmatic work.[1] But we are strongly recommended to provide the source anyway. As games team, I think we should set an example of how things should be handled. And we should follow this strong recommendation. Thanks, Bas [1] For a game definition for an engine, this is highly debatable; I'm pretty sure we should consider all those scripts programmatic works.
Attachment:
signature.asc
Description: Digital signature