On Tue, 2004-07-13 at 19:18, Brian Thomas Sniffen wrote: > Matthew Garrett <[EMAIL PROTECTED]> writes:
> > Right. That's the sort of conclusion I'm coming to. If it /is/ actually > > effectively another criterion, then the right way to go about it is by > > changing the DFSG. I don't think it really exists there at the moment, > > and so I think trying to test packages against it is unreasonable. > > I wouldn't object to calling it out in a specific point, but I think > it's already a necessary conclusion from the existing points of the > DFSG. I disagree. > You can push the privacy requirement off of one existing DFSG point > only by putting it onto another. For example, you can push it off of > point 1 by saying that it's not a distribution problem, only a > modification problem. And then I say: I don't see why it's necessary to "push" it off DFSG #1 - the only real argument I've heard for thinking it conflicts with that at all is your funny "fee" one, and I don't think that's going to fly with a wider audience. > I have to be able to prepare derivative works. Allowing me to do so > only for certain derivatives (Say, those with Changelogs) is Free. > Creating other conditions which I must fulfill is not Free -- whether > those conditions require sending a dollar to the author, putting my > name on a Wiki, publishing my changes, or sending my work to the > initial author. Why is that not free? I'm actually inclined to agree that private modification (with no distribution whatsoever) should never result in extra requirements being made, but in several jurisdictions attempting to restrict that would be limiting fair use. I agree with Branden about this - licenses that attempt to restrict things that you could do under plain copyright law are almost certainly non-free, because the whole free software movement is founded on the opposite being the case. I think that's a far more powerful argument. > So you say it doesn't kick in until there's distribution, at which > point I go back to saying it's a fee. But if you insist on the right > to freely derive new works and to freely distribute them, then you > have to have the right to not distribute them, nor to do things which > might prevent you from being able to make modifications -- like > publishing that you're about to do so. So your only real argument with this is the "fee" thing? Nobody's been arguing that a requirement that non-distributed modifications must be published would be free, so the only issue here is whether the DFSG says anything about the freeness of licenses which require you to publish distributed modifications. If the only way we can claim it does is to make slightly dubious claims about language, then I don't think there's going to be a great range of support. -- Matthew Garrett | [EMAIL PROTECTED]