Together with gsingers and rmuir we’re submitting a paper to SIGIR-2012 on Lucene 4. This is pretty exciting – it’s my first “formal” paper submitted to such an important conference.
We realized that the IR community at large is no longer familiar with Lucene as it exists today – the name is being referenced, there were some (1 or 2 as far as I know) papers that presented the original design ca. Lucene 1.x, and there have been some evaluations. All of this stuff is pretty old, important from historical point of view, but completely at odds with the current architecture, so it may actually be doing Lucene a disservice in the eyes of IR researchers. And yet there’s plenty of cutting edge stuff in Lucene today, both in terms of applying modern IR algorithms and addressing engineering challenges.
So, we figured that it would be great to present Lucene as it is today (i.e. the 4.x branch), the motivations behind the re-design and do it in a way that is understandable to the academic IR community, in an attractive way.
Time is short, the deadline is July 2, but we’re off to a good start.