Richard Stallman <[EMAIL PROTECTED]> writes: > I don't think > > it needs to be possible to use text from manuals in a program. > > A manual is free if you can publish modified versions as manuals. > > And is a text editor free if you can only publish modified versions as > text editors -- not as manuals or tetris games or news-readers or web > browsers? > > You have to be free to publish modified versions of the program as > tetris games and news-readers and web browsers, since those are > different programs, but a manual is a different kind of thing > entirely. It is to much to ask that it should be feasible to > conveniently publish a modified version of the program as a manual.
But this is done! It *is* feasible to publish a GPL'd work as a manual. It's even more feasible to do what I'm discussing, which is to take a GPL'd manual and derive a program from it. I'm sorry for not making that clear enough in what I wrote originally -- too much rhetoric and too little meaning. So here it is more clearly: I can take a GPL'd text editor, such as GNU Emacs, and derive a tetris game or a news reader or a web browser from it. I can even derive a manual from it. Such a book could be very interesting to read, walking a technically savvy reader through the bulk of Emacs' code. Because the GPL is a copyleft, all of these derivative works will be Free Software. I think the manual will also meet your definition of Free Documentation -- right? I can take a GPL'd manual and derive a program from it -- this part is clear enough that I don't think I need to write more. I cannot take a GFDL'd manual and derive any sort of Free Software program from it. I do not think it is too much to ask that it should be feasible to conveniently publish a modified version of the manual as a Free Software program. > The GPL, for instance, does not permit this in a way that is good > for publication of books on paper. And yet there are GPL'd books on paper, and their publishers make money from them. Would Harry Potter be a good choice to put under the GPL? No, probably not. Would "On Lisp"? Well, I'd love to be able to make slides and lecture notes from it... or use some of the big-enough-to-be-copyrightable code snippets. It has significant functional content. The more you talk about this, the more I come to guess that a primary motivation behind the codified, centralized GFDL is the desire to make things convenient for publishers of printed books. Why not provide the manuals for GNU software under a dual license: GNU GPL for those who wish to treat the manuals like software programs, and GNU FDL for those who wish to print books from the manuals. The FSF has sufficient respect that it can, by example and explicit suggestion, convince most people to dual-license their similar works -- and controls the copyright on all of those manuals anyway, so can make sure things are done right at the source. This has all of the benefits of the GFDL for publishers of printed works: they can add their "value-add" contributions and publish, licensing to their readers only under the GFDL. This has all the benefits of the GPL for publishers and users of software: they can combine the manuals with other work in the Copyleft Commons of the GPL, and can derive Free Software from the manuals. Software programs would probably be distributed only under the GPL, as the GFDL is awfully inconvenient for programs. This has all the benefits of the GPL and the GFDL for the FSF and the Free Software community, as the FSF's example would encourage most people to dual-license their manuals under the GPL and the GFDL. It avoids the problems you alluded to with a conversion clause, which destroys the benefits for publishers. -Brian