Wednesday, May 19, 2010

Derby 10.6 is released!

See Knut Anders's weblog for the scoop! I was part of several major elements of the 10.6 release, and helped out with a number of smaller pieces, so I'm very pleased to see this work released. Rick, thanks a million for putting out the release!

Sadly, after I left my previous job, the new owners of the system almost immediately dumped Derby and transitioned to a Industrial Strength DBMS. Sigh. Derby was in 24x7 production use for almost 4 years at AmberPoint, and I think it did a wonderful job. But I've moved on, and they've moved on.

I still think Derby is a wonderful piece of software, and contains some of the best Java code I've seen anywhere: clearly written, efficient, well-maintained, thoroughly tested, elegant.

I'm not finding much time to spend on Derby right now; working on my mentoring projects for the Google Summer of Code is about all I have time for.

But Derby is a good project and I hope to stay connected to it, and I hope that it finds the success it deserves.


  1. Job well done Bryan & the Derby crew.

    Thanks for the effort you have put on.

  2. Even before you left we saw very bad performance in many of the queries. We seemed to have reached a tipping point after some number of rows. The new industrial DB (my new employer) is
    doing a much better job on the problematic queries. I think the Derby problem may have been an optimizer bug like you saw before. The system would go to 100% CPU during the entire query.