Monday, February 11, 2013

SkySQL Java Connector vs DrizzleJDBC

A few months ago SkySQL announced their client libraries, and the JDBC driver is based on my DrizzleJDBC. A lot of people have asked me what i think about this and this post aims to set some things straight. (Note that i might be convoluting SkySQL and Monty Program Ab in this post, it is a bit unclear to me where the boundaries between the companies are)

About a year and a half ago SkySQL contacted me and told me they wanted to build their own version of MySQL Connector/J and they asked me if I wanted to build a few features in DrizzleJDBC and do a bit of consulting when they continued the development. They gave me a good offer, but I felt I didn't have time since I had just started a new job. So, they just did what anyone would have done, they took the code and started hacking anyway.

I have no hard feelings against SkySQL, towards me they have handled this in a good way, I totally understand the reason they wanted their own fork - it is easier to tell the bigco enterprises they are targeting that "we built this driver" than "we contribute alot to this hippie opensource project".

A few recommendations to SkySQL if they do anything like this in the future though; be more transparent in announcements and communication - tell people where the code came from, it is not like it is a secret anyway. Don't claim that the forked project is dead, when they forked my driver, a lot of companies were using it in production without issues. Don't relicense, it will only make people think you are evil.

SkySQL have offered to contribute their code base back to DrizzleJDBC, but I think i will keep DrizzleJDBC a small minimal implementation, I never intended for it to be a full Connector/J replacement. Drizzle JDBC is not dead, I fix bugs whenever there is a report and merge pull requests.

Oh! I fully expect a cake whenever they make their first connector-related sale!

Update: Appears SkySQL made their first sale, gigantic cake, thanks!