On Wed, Aug 25, 2004 at 09:29:43AM -0400, Brian Thomas Sniffen wrote: > > The FSF could release a GPL version 3 which has completely arbitrary > > terms. If control of the FSF had passed to someone unscrupulous, these > > terms might be proprietary. [I'm not saying this is a likely scenario, > > just a possible one -- I hope this hypothesis seems particularly > > outrageous.]
Well, in theory not--"such new versions will be similar in spirit to the present version". That vague limitation isn't particularly reassuring, of course. > This is where you lose me. The FSF releases their GPL v3, which is > suspiciously similar to a Microsoft EULA. Now what? The change I > submitted, which is distributed with GCC, is licensed only under GPL > v2. Earlier, I wrote a reply asking about things like "v2 vs. v2-or-greater compatibility" and so on; but after thinking about it for a while, and rereading the GPL, I realized this is a very common mistaken idea: you *can not* release your work under "GPL v2, not greater". GPL#9 says "if you release under v2, upgrades are allowed". If you want to release under v2 without allowing upgrades, you'd have to revoke clause 9--which would be GPL-incompatible, so you can't do that to your gcc contribution. The practice of writing "GPL v2" in their licenses when they mean to deny upgrades is wrong--"or greater" is implied by GPL#9. (Raul said a similar thing, but I thought I'd explain it from the perspective of one who, until yesterday, held the same belief as you.) -- Glenn Maynard
