Here's some candy: we've built simple sample gateways to the USDA Plants site and the Simthsonian Mammal Species of the World site. Server-side perl makes up cgi requests to the site, parses the returned html and emits XML. Of course, it is a very unstable solution since changes to the site's html will break it. But we did it because we experiment with federating heterogeneous XML biodiversity databases.

You can nibble at http://www.cs.umb.edu/efg/xml/junwan/DEVELOP/gateway/gateway2.html

Bob Morris Jun Wan