why should indexer/searcher be coupled to email files in particular, instead of dispatching on MIME types? why should the datamodel be something ad-hoc instead of RDF?

http://blog.whats-your.name/post/2011/02/14/sifting-thru-you...