I found this page extremely helpful in finding out EXACTLY what Lucene is doing (and how, if I wanted to, to change it). Like Erik said, it does pretty darn well just as it is. I'm not sure if anyone has already pointed you to this page yet.
You'll have to spend some time diving down in to each component of the score, as it's composed of a number of different factors, but it's all described on this page, and it tells you what the default behavior is using DefaultSimilarity. http://lucene.zones.apache.org:8080/hudson/job/Lucene-Nightly/javadoc/org/apache/lucene/search/Similarity.html Donna L. Gresh Services Research, Mathematical Sciences Department IBM T.J. Watson Research Center (914) 945-2472 http://www.research.ibm.com/people/g/donnagresh [EMAIL PROTECTED]