In case anyone really wants to do it right, what PG needs is to have each book (and other documents) marked up semanticly. Of all of the exsting SGML/XML applications, TEI seems best for what PG is doing. Combined with SVG and X3D for graphics, xcite for any citations, etc. The best way to mark up existing PG texts may be to put the docuemnts in a wiki alongside scans and encourage the public to add the markup. Wiki-style markup seems to be easier to comprehend for most of the public. (And with reason.) In this model, incidently, each work could be served as a single file, complete with images and the like included inline. And the plain text version can be readily extracted using a stylesheet. TEI is at: http://www.tei-c.org/ -JimC -- James Cloos <cloos@jhcloos.com> OpenPGP: 1024D/ED7DAEA6