MySQL's non-GPL Adventure
I’m a bit late on this piece of news, and it’s been treated by more or less everyone who owns a blog. The news in short is that MySQL has decided to not release their online backup tool/driver to the masses. It will only be available to paying Enterprise customers. A lot of people have been writing about it, Jeremy Cole, Vadim Tkachenko and Don MacAskill and they all have valid points in both directions.
The reason why I’m rather late with this is that I’m not entirely sure what I think about it. On one side, I can see their point since they after all are a revenue driven company. MySQL/Sun employees need food on their table as well. Selling additional services and to a certain degree also selling additional software is a model which has been around for a long time. That said; it’s not necessarily a good model by default. The Fedora/RedHat model is frequently referenced as a successful one. I believe this would be ideal for MySQL as well. Let the masses break it, fix it and hand it over to customers willing to pay for the reassurance and increased stability.
On the other side, I’m really pro-opensource and any diversion away from it worries me. The main problem with this approach is that since the backup tool will not have been tested by the general masses and only by the QA team. While the QA team might do their job very well, they couldn’t come close to the quality of testing that hundreds of thousands of people can do together) I’m not particularly fuzzed with not being able to use the backup tool everywhere. I tend to roll my own, or use something a brilliant mind has GPLed. My biggest concern is the quality of the software and the potential of taking this further, extending the witholding of features into the more critical parts of MySQL (storage engines, further developed optimizers etc.) Let’s hope that day never comes!
A mid-way compromise would be to release one version of the tool under a license which allows closed forks, get the bugreports, fix it and release the finalised version to the enterprise customers. However, this is probably the least attractive option even though it’d give the highly skilled community a base to fork the software from and apply our own fixes.
All in all, considering what MySQL still is contributing to the community I will continue to be a fan.