[EMAIL PROTECTED] (Thomas Bushnell, BSG) writes: > > Why is it so different to a published library function? > > Apart from convenience of argument, that is. > > Libraries are much more tightly integrated with their callers, for > example.
Oh, and you ignored my stressing the importance of considering intent and common practice. It is the intent of kernel designers that the kernel is not one program together with the programs that it loads. It generally *is* the intent of library authors that the library is part of the programs with which they are linked. The nature and structure of the situation is important, as are the intentions of the parties. But the real point, which people frequently seem to miss, is that only this way can the integrity of the copyleft be maintained. If the use of dynamic loading were sufficient to get around the GPL, then its copylefting provisions would be entirely a dead letter, and it would be no better that the BSD license. For that very reason, the FSF and GPL-users in general staunchly stick to the importance of this rule, because it is the only thing that preserves the integrity of the copyleft itself. Now, if you want to argue the point further, *please* take it up with RMS or with Eben Moglen. There's really no advantage to wasting debian-legal with it. That's what I was trying to get across to [EMAIL PROTECTED], and now he's suckered us into a stupid argument. Thomas