Batch convert Word documents to OpenOffice documents (with unoconv maybe? I can't remember), convert them to XHTML with odt2xhtml, run them through Tidy, import them to Drupal, mine them with a slew of regular expressions, import tags from last.fm and you end up with this: a database of artists featured on past Mixing It/Where's The Skill In That? radio shows (where playlists existed, at least).