Ilya Grigorik has written up what a few people are probably doing in Ruby Screen-Scraper in 60 Seconds.
The combination of Firebug and Hpricot is power, simple, and fast.
The power comes in not having to come up with the XPath to get that piece that you see in your browser. No more walking through the DOM working out the XPath from that. No more looking at the source. Find it. Inspect it. Copy it.