Taking this even further off-topic from code4lib (oh, hey, Archimedes, an institutional repository developed by Université Laval up here in Canada, ran on Apache Derby - whew, back on topic briefly!):
On Mon, 2009-04-20 at 12:49 -0400, Jonathan Rochkind wrote: > Dan Scott wrote: > > > > In the context of the Oracle-Sun and MySQL/OpenOffice/yada yada parent > > thread, Derby demonstrates that a software project can 1) go from > > proprietary to open source, 2) be contributed to by (in some ways) > > direct competitors, and once it is open source 3) lose commercial > > support from one company but gain it from another, and 4) survive for > > whoever depends on it and wants to continue using it regardless of what > > commercial entities may do. > > > > > > It also shows the dangers of this to the user community though, since by > your description Informix ended up forked into Derby and JavaDB, with > the commercial support being for JavaDB, but the open source development > taking place in Derby, and the open source development being kind of > stunted too. A situation which is not great for the users. Naw, no forks, and apologies if I represented it that way. JavaDB and IBM Cloudscape were just cuts of Apache Derby from the apache.org project that people for which people and companies could purchase honest-to-goodness support. It's a standard business model; much like multiple commercial companies exist that can offer support for Apache Solr. Nobody lost or has lost in the Derby community so far simply because some companies found a business opportunity in offering support and packaging. Mind you, there's nothing in the Apache 2 license that forces companies or indviduals to contribute modifications back to the mother project. It just didn't / doesn't make business sense to invest that amount of effort to maintain the fork. It's enough work trying to build a set of regression tests for one project without having to worry about forks. > Still, it could have been worse. To make it better would take (or have > taken) concerted effort from the user community -- which probably would > have happened if Postgres and MySQL didn't exist, making an open source > Derby more important for more people. I, uh, don't know about that; its sweet spot really is as an embedded database, although it's okay as a server for single-disk partition databases for applications that use only JDBC to connect to it. Which makes it a pretty niche database, in comparison to PostgreSQL and MySQL.