The pylucene project page has the usual information including a Subversion repository, mailing list info, and a bugzilla component. We're experimenting with the use of Subversion as a version control system.

A lot of people have asked for pylucene as a separate project. I hope that this will be useful to folks out there in the broader Python community. I also hope that some folks will get involved in helping to improve pylucene. There are a bunch of things that need doing, a distutils based install being high on my list.

From my perspective, this a good study in looking at how to leverage other projects. We were interested in full-text indexing from Python. We had a few options that included Managing Gigabytes, Lupy (a pure python port of Lucene), CLucene (a C++ port of Lucene) and the original Java port of Lucene. When comparing Managing Gigabytes versus any of the Lucene related libraries, most of the Lucene ports are being actively maintained, while the Managing Gigabytes software hasn't been updated since 1999. That leaves you with a Lucene variant. Lupy is much slower than either Lucene or CLucene. So that rules out Lupy. Only Lucene and CLucene are left. Coming from Python, it would seem that CLucene is the obvious choice since it's written in C++. The problem with this is that it's a port, and you are always trailing Lucene. Further, the real search engine expertise lies with Doug Cutting, who is working on Lucene. So if you have a search related problem in CLucene, it might have to go all the way back to Lucene anyway. The biggest problem with Lucene is that it's written in Java, and the last thing we want in Chandler is to have people also install Java. Enter gcj. Thanks to the hard work of the gcj team, you can compile Lucene into native code, which can then be wrapped with SWIG (just like CLucene needs to be for Python). As both gcj and gcc get better, pylucene will benefit (of course CLucene would also benefit from improvements to gcc).