A look at the MySQL forks

I think the most interesting question is how much the forks diverge from each other. It sounds to me like, aside from Drizzle (which I think could be seen as a different project that shares a developer and a bunch of code), these "forks" are either subsets of the available user-contributed patches or different storage engines.

This would suggest that MySQL is currently structured sort of like the Linux kernel, except without Linus's tree. If the people making changes to the core code can reach a consensus among themselves (i.e., a patch will start in somebody's set, but eventually either everybody will decide to include it or everybody, including the origin, will decide to drop it), or at least can avoid ending up with merge conflicts where there is persistent disagreement, it's really just decentralized development.