A rather quaint trait of the Apache Lucene project's homepage is that the search is actually powered by Google. Shouldn't Lucene -- an open source search technology project -- be able to improve on that?

The short answer is: yes. At the ApacheCon Lucene meetup in Amsterdam, Erik Hatcher gave a demo of LucidFind, which is based on Solr -- one of Lucene's more productized flavors. (Incidentally, this was one of the fastest demos I've ever seen -- it may have lasted for all of two minutes). LucidFind searches most of the public sources for Lucene (including wikis and mailinglists) and related projects (such as Solr, Nutch, and Mahout), and then allows you to drill-down into the results using faceted navigation. You can narrow results down on project, source, and author.

Of course, Erik Hatcher is one of the founders of Lucid Imagination, and LucidFind is a showcase for the company. Other search vendors have done the same before them, notably Vivisimo with its federating/clustering Clusty, and Exalead with its public website search. The clever thing here, of course, is that those are interesting examples, but LucidFind takes it to a new meta-level -- actually indexing search so you can find it.

So certainly in this scenario, Apache's Solr proves more useful than Apache's Google Site Search. If you ever need to find any technical details on Lucene, this is the place to go. Now if only it could find you the answer to the question whether Lucene would work for you, too. But that's another story.