Hello David:

As for improving the search facilities, you have an interesting opportunity here to implement taxonomies and metadata in such a way to be understandeable by the emerging semantic web technologies.  This would involve using stuff like SKOS, OWL, and RDF. 

You may be aware of this project which is using an OWL ontology to organize SEP.

You may have already thought about all this... just thought I'd throw it out there.

let me know, if you want to discuss more.

Marcus
(the guy who emailed awhile back about helping with DIVRE)

P.S. does philpapers have any interest in building a better SEP? specifically one that used true version control under the covers...which would support things like version history and the ability to cite any article at *any* given time, and have the citation be a reference to a complete internally consistent enyclopedia

P.S.#2  I'm guessing you considered lucene and/or solr.  Just curious why you chose sphinx?