Well, I'm making a bit of progress on the author reading/signing front. Instead of generating .ics files directly, the HTML reader is now feeding a database, and the .ics (and soon RSS) files are generated from the database. I'm now extracting book titles from Elliott Bay Books; all I had before was author names (the two bits of information are hiding on different pages). I'm also using one of Amazon's web service interfaces to search for books matching the author and title listed, and turning that into an ISBN number.
This buys us a few things; first, we have better data, because we can cross-reference author and title information with Amazon. Amazingly enough, I found at least one mis-spelled author name. Second, because all of the data is sitting in an easy-to-query form in a database, it's a lot easier to build an RSS feed from it. In fact, I already have a working RSS 1.0 writer, but it's still a bit rough, so I'm going to hold off announcing RSS feeds until I get it working correctly. Finally, by having the ISBN number, we're set to tie into other book-related services, like All Consuming, so we can do things like genre- or recommendation-driven lists of upcoming events. That's still a ways away, though.