I haven't made up my mind which side I'm on, but here are some more random criticisms.
Raul Miller <[EMAIL PROTECTED]>: > Then: P is the source code for a program which includes L. A is > instructing people to download L (if necessary) to compile P. > > This is a clear attempt to evade the copyright on L. A went to > considerable effort to design this situation. It didn't "just > happen". It certainly is a clear attempt to evade the copyright on L, but that doesn't mean it's not a successful attempt. Just because A's actions go against the spirit of what F intended doesn't mean it's illegal. Law has to be more precise than that. > > As far as I understand you RMS's claim would be that A is somehow > > violating the GPL even though he never copies the library it applies > > to. > > Oh, come on -- he's copying it by reference, in the compilation > instructions. But that's really irrelevant: > > All that's needed for "contributory infringement" is that A make it easy > for people to make illegal copies. And, any distribution which results > in working copies of P clearly involves illegal copies of L. What are these "illegal copies of L" and why are they illegal? Are you talking about the memory image of the process? Creating a memory image is what you normally do to run a program, so I'm not sure that this counts as copying. Or you do you mean the copy on disc, which is totally unchanged since before B loaded P onto his sytem, so why should it suddenly become illegal when used with P? I am opposed to allowing things like memory images count as copies with respect to copyright law as that would allow copyright owners to restrict how data is used. I think I tend to believe that the GPL might be practically equivalent to the LGPL in the hands of someone who is willing to resort to a little trickery. However, I am willing to believe this may be a bad thing, so I am very open to coherent arguments that justify the claim that "if you tell people to use a program FOO and BAR linked together, you're combining them into a larger program". I suppose that means that someone who publishes the glue code and the instructions for linking a GPL program with a GPL-incompatible program is violating the copyright on the GPL program EVEN IF THEY NEVER HAD A COPY OF IT LET ALONE MADE A COPY OF IT! I have in the past suggested that it might be possible to combine and generalise the GPL and the LGPL into a single licence that talks about functional units with well-defined interfaces rather than about "programs" and "libraries", however, I don't think I ever got as far as a precise formulation of this idea. Edmund
