I recently read the discussion on the GCC development mailing list related to GCC's transition to GPLv3. Despite generating 172 emails, the transition was quite smooth actually.
I decided to read about GCC's transition because I wanted to investigate a (false) claim which Slashdot featured on their front page. Slashdot's story, which I now know to be complete rubbish, claimed that key GCC developers were talking about forking GCC.
Actually, they mostly discussed version numbers, as well as the copyright status of code written by scripts, and of World War II photographs, etc.
About the version number, DJ Delorie argued against bumping the version number up to 4.3.0 for the new licence because Slashdot might print a negative story about GCC 4.3.0 not having the features previously scheduled for that version number. Which is pretty funny in hindsight because the developers finally agreed not to bump the number up to 4.3.0, and Slashdot still printed a negative story. It seems there's no getting around the fact that Slashdot just loves controversy, even when it means creating anti-free-software FUD.
So, in review, Samba's GPLv3 transition went well. Then SugarCRM ditched their custom licence and moved to GPLv3, which, as I said, is one way in which GPLv3 tackles licence proliferation. And now GCC and 286 other projects have made the transition, according to Palamida's GPLv3 counter. It seems that GPLv3 transitions are going quite smoothly.
In fact, it seems that many commentators have been surprised (and some have surely been disappointed) by the lack of major problems. There is a vacuum of bad press, which is why Slashdot has resorted to printing hallucinations, and InformationWeek can only publish bad news by inventing it. All in all, I'd chalk that down as a success for the eighteen months of hard work that was the GPLv3 consultation process.
Ciarán O'Riordan, -- Support free software: Join FSFE's Fellowship