Richard Stallman <[EMAIL PROTECTED]> writes: > 1) Because the borders between the cases are ambiguous and uncertain. > > I sent a message a day or two ago (perhaps after you sent this one) > which addresses that issue. > > 2) Because we want to be able to combine works from different sources, > > As I explained, this desire is usually impossible due to > incompatibility of licenses. To reject the GFDL on these grounds and > accept some other GPL-incompatible license is a double standard.
Wrong, wrong, wrong. We want to be able to combine works. I keep repeating this, you keep saying "I understand you", and then you keep proceeding as if I hadn't repeated it. The desire expressed in number two is *not* about the desire to combine any two arbitrary works. We have that desire, but we agree that it isn't currently met, and it doesn't impact freeness. So let's think about *extension* rather than combination. Can I extend work X and have it be a piece of free software? Any free software license must meet that test. The GFDL does not meet that test. There is no double standard, because you have misconstrued the standard. The standard is NOT whether X meets: "For all works A, A can be combined with X and still be free." If that were our test, it would be inconsistent to be bothered that the GFDL fails, since of course most free software licenses fail. We are concerned with whether X meets: "There exists a work A of free software, such that A can be combined with X and still be free software." It is this test which the GFDL fails, and it fails not because it has terms which happen to be inconsistent with this or that free software license, but rather because it contains terms which are fundamentally inimical to the very concept of free software itself. It might be sensible not to care about this if documentation and programs were radically different things, but they simply are not. Thomas