John Cowan's posted some great reports from Extreme Markup 2007. Amongst them a neat summary of the key bits of FRBR (quoted here because I have real trouble keeping this stuff in my head for more than 10 seconds) :
...the FRBR model of works, expressions, manifestations, and items. For computer-document purposes, a work is the abstract work, e.g. Hamlet; an expression is a particular kind realization, like a particular edition of the text or recording of a performance; a manifestation is the expression in a particular format such as .txt, .html, .doc, or .avi; and an item is a particular copy residing on a particular server, disk, or tape. Ideally, there should be separate URIs for each of these things.
John also points to an Eric Freese paper about doing chatterbot AI with RDF. This has been lurking in the nether regions of my todo list for years, it appears now to have been done (although I couldn't find any working link to the material online). btw, if I remember correctly, although the core AIML material is sane enough, there are some entertaining Outsider Software ideas around the fringes.
Another paper he mentions talks about describing a pipe organ using topic maps. I need to read this properly sometime soon, there's an obvious overlap with what I want to do for describing the Tinocaster in RDF ( grr, I forgot, that server's not got its data back yet - link to be fixed soon). As it happens, looking at that took me (with the aid of various Talisians) down the path of using FRBR - how to do a music store, and more generally looking at describing (manufactured) products, prompting the coinage FRPR.