In addition to the page-processing you mentioned (Strip parts of pages, or extract only certain content, the way MobiPocket's .enews XML parser works), I would like the ability to alter urls before following them. Some sites carry different versions of a story with nearly the same url. For instance, the New York Times accepts ?pagewanted=[1|2|all|print] at the end of a story url to get that version of the article. The "print" version is quite palm-friendly and it would be great if I could have iSiloX add that flag to all urls on that site, especially if combined with the nearly palm-friendly index page
here.