Anthony DeRobertis <[EMAIL PROTECTED]>: > > However, you could certainly distribute P on its own if > > you could reasonably claim that P is useful without GPLLib. > > I'll further argue that P is not based upon GPLLib in any meaningful > manner; it includes absolutely no part of GPLLib.
If P is useless without GPLLib, then it might be argued that by distributing P you are encouraging people to link it with GPLLib and are thus indirectly distributing a combined work "P+GPLLib" which infringes GPLLib's licence. That's why the existence of alternative implementations of GPLLib is important. (Even the existence of alternative GPL implementations might help.) However, if Debian were to distribute P and GPLLib in such a way that P uses GPLLib by default, then I would guess there is potentially a problem even if there are alternative non-GPL implementations of the library. > > which is probably doable if the > > script makes relatively minor use of grep, etc > > I think you'd have a very hard time finding scripts which make minor use > of GPL utilities. Even our cat program (like everything else in > coreutils) is GPL. So is our echo program. I suppose I should explain what I mean by "minor", though I'm not quite sure myself. Perhaps one could compare with the situation where someone distributes a summary of someone else's novel, compared with where someone distributes a criticism of the novel that also summarises it in the course of criticising it. I don't have any legal evidence for this idea, but I suspect that in addition to how much is taken from or used from another work, what else a work contains may be relevant in deciding whether it is a derived or an independent work. Edmund
