#comments

So, in the wake of the purchase of Sun by Oracle, although by no means caused by it, it seems like MySQL is at a crossroads – it’s becoming forked/branched and one wonders at what its future is going to be like. At least this one does. In the short term, obviously, not a ton, but in the medium and long run if these forks get legs and keep moving they could diverge significantly from each other and that has implications.

In that post, Jeremy Zawodny comments (and links to a longer rebuttle) that he believes that we’re heading to a world where MySQL will have many distributions, like Linux, and different distros will appeal to different people/usages. I think, though, that he is wrong. A code fork is not a distribution. Linux distros all use the same Linux kernel and change the extras that you get with said kernel – what is going on in MySQL is not simply changing the storage engine, but potentially changing the database itself. As if Ubuntu and Fedora and everyone else also forked the kernel for their distros – which I don’t think anyone would agree is a good idea.

I am not saying that this end result would necessarily be bad, I don’t know. It is probably nice to have a few different MySQL’s targetting different use cases. On the other hand, focusing the entire community on a single product has the advantage of putting a lot of momentum there. I mean, remember when there was MSQL and MySQL? Pretty similar initially and both were in wide use, but then MySQL really took off when everyone switched to start using it – momentum and popularity are extremely important in the Open Source world. I would worry that multiple diverging forks of the software could dilute the speed at which they all move forward.

And this is well under way. Drizzle was already getting a lot of buzz, people are excited. I think many folk were worried about the direction MySQL was going to take under Sun’s leadership. Now under Oracle’s which has shown hostility to the db (when it purchased Inno and Sleepycat out from under them, although, proceeded not to really do anything good or bad with it), I think people will be even more likely to look to others for development. And now one of the founders of MySQL, Monty Widenius doing his thing and starting an Open Database Alliance, which notably does not include Oracle.

While the intention seems to keep code from diverging significantly, we’ll see if this actually happens. It doesn’t seem likely that Oracle will want to lose control of one of the major assets of it’s multi billion dollar purchase. And even within the ODA, will it actually preserve code compatibility? Doing such things is really difficult. Look to all the failed attempts to enforce even just a standard file system structure in all the Linux distros. It will be interesting to see how Oracle reacts to all this and how it all plays out in the end. Perhaps this will inject a lot of innovation into the codebase – a little competition is great for development – but then down the road all the competition will ultimately result in one of them becoming the dominant, that would be a great outcome. What do you think’s going to happen?