I'd still argue that the outcome of the whole BitKeeper thing was more what Larry McVoy seemed
to fear than what anybody else seemed to fear. BitMover worked out a lot of the fundamental
misunderstandings in version control through a lot of effort and made the results clear to a
programmer who had written off the entire concept before, but turned around and, in six weeks,
wrote a better replacement, at a cost to the community of a slight delay in the release of
2.6.12-rc1. I don't think it's a coincidence that the only person able at that point to write
a better version control system than BitKeeper was also the only person trying to write a
version control system who had extensive experience using BitKeeper.
On the other hand, the middle period, where people who could abide the license were using
BitKeeper, was awkward, because not everybody had perfect access to the development process.
So I'd say that 2002 started a hard time (albeit less hard than the time before without
significant tool support) that turned out very well in the end, rather than a good time that
went sour eventually.