On Wed, 2004-02-18 at 20:01, Roy T. Fielding wrote: > Allow me to make a less convoluted translation: > > "Code incorporating patents, when the code and contributors' patents are > licensed solely under the MIT license, cannot be incorporated into a > derivative work distributed under GPLv2, because any recipient who > receives a copy of such a derivative work has no rights to use any of > the patents incorporated into the original MIT code." > > Why, then, is the MIT license compatible with the GPL?
Reason 1 of 2: -------------- Because the MIT license is a blanket grant of permission, almost without restriction: | ...Permission is hereby granted, free of charge, to any person | obtaining a copy of this software and associated documentation files | (the "Software"), to deal in the Software without restriction, | including without limitation the rights to use, copy, modify, merge, | publish, distribute, sublicense, and/or sell copies of the Software, | and to permit persons to whom the Software is furnished to do so, | subject to... It out and out grants you "permission" (!!!) ... "without restriction" ... "including without limitation" ... [all this stuff] ... "subject to" ... [hardly anything]. It doesn't just grant you a copyright license, but grants you out and out permission. (I just noticed that in rereading the license for this email--shocking! IANAL, so I may be very confused here--can a Real Lawyer comment on this? I was quite surprised to notice the apparent breadth of the grant.) So the fact that the grant doesn't mention patents doesn't make me think patents aren't included, just as the fact that the grant doesn't mention copyrights doesn't make me think a copyright license isn't included. (Also note that the license even uses those magic patentspeak words "make, use, and sell".) So, I would *assume* that a contributor submitting software initially under just the MIT license, (where the code incorporates patents they own), is in effect allowing their incorporated patents and incorporated copyrights to be used anywhere with hardly any limits or conditions. It would make no sense to me for someone to hand the world a piece of code containing his patents and say "I hereby give anyone permission to do anything with this, without restriction and without limitation", and later come back and say that he wasn't licensing the incorporated copyrights or patents. So it looks to me that an MIT-licensed patent, (to use sloppy wording), is in effect also automatically compatibly GPL-licensed, (to continue the sloppiness), as well as being available for proprietary works. So I don't think your less convoluted translation holds, as the MIT license couldn't cause a patent to be licensed solely under itself. (Now, granted, IANAL, and IANAPL, so I may be totally off-base here. And I'm assuming of course that the licensor/contributors have full copyrights and patent rights to the code, there are no disputes on ownership, and that sort of thing.) Reason 2 of 2: -------------- However, even I'm totally wrong in the above, and the MIT license *doesn't* implicitly include a patent license when the owner of a patent distributes code under that license, there's still nothing in the MIT license that *retracts* any [patent] grants under any conditions in which the GPL would not also retract/disallow its granted rights. (For instance, lying about author attributions breaks both licenses.) So, the MIT license isn't adding any restrictions not already in the GPL, meaning the MIT license is GPL-compatible. (This is a simpler claim than the above; "IANAL" still applies of course, and I'm still assuming there aren't weird corner cases going on with IP ownership disputes and such.) (Side issue: I just read John Cowan's very similar but impressively far more concise response. I've got to learn to write more concisely!) -- Mark Shewmaker [EMAIL PROTECTED] -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
