Paul Nathan Puri <[EMAIL PROTECTED]> writes: > I would make that argument, but that's just me. RMS made a big mistake by > not defining all the legal terms. One huge risk is that 'derivative' and > 'copyright' will be divergently interpreted in various countries.
That is probably intentional. RMS seems to want "deriviative" to be interpreted in as wide a sense as possible, so that as many other programs as possible will be forced to be GPLed. Offering a definition of "deriviative" would only serve to narrow the definition in comparison with what the court would decide without a definition: I don't think a license on a program can validly claim to influence other programs that the judge would not, without consideration to the license, deem covered by the original author's copyright. I might be misreading you, but it seems you're saying that I could write a program and distribute it with a notice saying: This program, or any of its deriviatives, may only me copied if you pay me $100 per copy. For the purposes of this note, a "deriviative" means any program whose source code include the use of the word "foo". It's relatively clear to me that if I sued someone who dared to write foo in his program, I would be told by the judge that I cannot make other people's program deriviatives (in the copyright law's sense) of mine simply by claiming so. There must be some kind of intrinsic legal limit on which kind of connections between programs that give the author of program A right to set conditions on the distribution of program B. As I see the discission, we're trying to figure out where that limit is in connection with header inclusions. -- Henning Makholm
