Evan Prodromou wrote: You can develop something independently, disclaim copyright to it, and then incorporate it into a GPL project. (If that was not possible, it would mean that no GPL project can ever incorporate public-domain code.) Maybe nobody can use those public-domain pieces because they're pretty much useless without the rest that is GPL, and is thus pretty much forced to adhere to the GPL, but that doesn't mean that Erik and Lee Daniel have to keep their copyright. Timwi

Evan- I would recommend removing the word "also" from the above paragraph, but I didn't write it. Maybe Lee had a reason for phrasing it this way. Derivative copyright is a thorny issue (and underscores the idiocy of the whole concept of copyright), so I won't get into it too deeply. What it essentially is supposed to say (IMO) is, "Everything by these contributors that *can* be legally put in the public domain, is." The more other coders follow our example, the larger that body of work will be. You don't have to if you want to use the GPL, since GPL'd code can legally include PD source, just as GFDL documents can legally include public domain texts and images. If you want to be on the safe side, just adhere to the GPL. But if I added, say, a major new extension to MediaWiki, then you could use the first version of that extension that is not incorporated into a GPL work by someone else and do with it whatever you want. By the way, since >99% of our users never download or install MediaWiki (they use the wiki through someone else's web server), the difference between GPL and PD is surprisingly small for all practical purposes. You could, for example, add lots of proprietary libraries to MediaWiki or port the whole thing over to .NET, but never give anyone the source - if you don't distribute it at all, you don't have to. This is not theoretical. There is at least one person running a quite expensive wiki hosting service on the basis of a heavily modified UseMod codebase which has never been made publicly available, for instance. And the Scoop weblog/diary system has been significantly extended and localized by the people at nensch.de, but they have explicitly stated that they don't want to share the code because of the effort they have put into it. I'm sure there are dozens, if not hundreds, more examples already. The next major GPL release will supposedly address this, but existing code is not affected by any such change. Using the GPL (copyleft) to creepingly subvert copyright is at face value a good idea, but it puts you in a position where you have to actually enforce and defend copyright, and I don't want to be in that position. If copyright was abolished, then you could not force me to open the code of a proprietary MediaWiki fork, though you could reverse engineer it and copy it freely. The ultimate goal of the GPL is to replace the copyright regime on software with a copyleft regime, whereas I'd prefer to get rid of the regimes altogether. Though not typically, in this instance I find myself agreeing with the libertarian view ("Do whatever you want") more than with the socialist one ("Sir, please give us your password so that we can share your code with the community"). I advise anyone who uses the GPL to think about whether they are willing to enforce it. If so, then the GPL (especially after the web service "bugfix") is the best choice of license. In any case, this is not something we should fight about. In the short term, the differences are minor. Regards, Erik