On Tue, Feb 01, 2000 at 12:35:15PM -0800, David Johnson wrote: > Oh, but it does! I'm sorry that I can't quote the relevant law to you, That's all right, I know what you're talking about and you're right it has been tried many times and has failed each one so far. > not being a lawyer or anything. But there have been court cases in the > past that have determined that APIs cannot be copyrighted. A footnote > containing a chapter or page reference is the prose equivalence to an > API. For a program, a reference to call a function elsewhere is > analogous to a footnote and the #include <qapplication.h> statement is > analogous to any entry in a bibliography. Okay, I will accept this premise as true. > The GPLd portions of the KDE source code, as long as it remains source > code, contains zero elements of Qt. It contains only references to the API. Since we have accepted that API's are not Copyrightable, this is a non-problem. Source distribution therefore cannot be considered a problem. > After it is compiled into a binary, > there is only a miniscule amount of Qt code, and what there is consists > of the translated/compiled API. Anything within KDE that belongs to Qt > is easily covered under copyright law as a reference or fair use. Ahh, but a representation of an API _can_ be Copyrighted. That is, Motif's headers cannot be used by the Lesstif project, they would be infringing Copyright if they used them because the documented API is different than the source code which implements it. The "Public Domain headers" suggestion could possibly be applied here, but that is a cheap way to get around the above problem and I don't think anyone involved is seriously considering it because that is precisely what it would be taken as by everyone else. > Now a license may impose further restrictions to the user beyond what > copyright disallows. Only as conditions for allowing of distribution. Shrinkwrap licenses in proprietary software are not under this category because they first make you agree that you do not in fact have in your posession a copy of the software. If you don't agree to these terms to install the program, you're free to do whatever you want under the Fair Use provisions of Copyright law. This includes reverse-engineering a binary, etc, etc, that the licenses do not permit you to do once you've agreed to them. Whether these are or even should be allowed to be completely legal is a matter of some debate. So far nobody has had the funds to stand up to a real legal battle with someone over the legality of a license which you're not allowed to read until you've already paid for the product (and in almost all cases cannot get a refund for) so who knows how that would really turn out. > But I don't see the GPL doing this. It only refers > to "all the source code for all modules it contains" plus scripts and > makefiles and such. I don't see Qt as being contained within KDE (except > for the API) as it is external to it. Of course, RMS (politely) > disagrees with me on this, but as it is written, it is not clear at all > that Qt is a module of KDE as defined in GPL version 2. As source, that is the case. As binary, the code is a derivative of the source and Qt. Whether that happens at compile time (the headers), link time (linking to the shlib), or at runtime (loading the shlib) is a matter of some debate. The real question is how would a court decide. The answer is that nobody knows, it hasn't gone to court yet. > > [Also, there's the Objective C situation > > (where gcc mods were distributed as ".o" files and the end user was > > given instructions on how to build the modified gcc)... > > Statically linking NeXT's object files into gcc in the user's private > environment is a far cry from dynamic linkage, and has little bearing > upon KDE/Qt. Yes, but is actually something that you can do legally under the GPL. -- Joseph Carter <knghtbrd@debian.org> Debian Linux developer http://tank.debian.net GnuPG key pub 1024D/DCF9DAB3 sub 2048g/3F9C2A43 http://www.debian.org 20F6 2261 F185 7A3E 79FC 44F9 8FF7 D7A3 DCF9 DAB3 * Knghtktty is not going to ask how zucchini got into the discussion ...
Attachment:
pgpXqm99n2DGy.pgp
Description: PGP signature