On Tue, Jul 27, 2004 at 01:09:48PM -0400, Brian Thomas Sniffen wrote: > > I had thought from previous GPL discussions that "distribute the source > > and let users link it" was not a reasonable way to sidestep license > > compatibility issues, because the source was still a derived work. Does > > this mean that one can distribute the source (or object files, even) of > > a program that links to a GPLed library, and just let users link it? > > That seems like a rather large loophole. > > No. The FSF's claim is that the source for some program using, say, > the GNU Readline library, is essentially bundled with instructions for > building this into a binary which incorporates GNU Readline. Anybody > distributing such sources is very clearly suggesting that users do the > final assembly, and has certainly built it themselves to test it. > > So there's somebody building a copyright-violating work, and > distributing copies of it in a strangely compressed form. That > doesn't make it any *less* infringing.
No, there's somebody building a derivative work (the binary) and not distributing it (which the GPL allows), and distributing a work that is not a derivative (the source) and suggesting that others build it (not infringing, as long as it's not distributed). I agree that it goes against the spirit of the license, and that it wouldn't be something to try in court--but I don't think the argument is so clear cut. I hope that the FSF wouldn't want strengthen the idea that telling people *how* to violate copyright should be illegal (eg. DeCSS, "contributory infringement"). -- Glenn Maynard
