Kirk and I are having a lot of fun with this RSS issue. Here's an updated version of my previous example that copes with Unicode BOMs and pre-validates the fetched RSS before it attempts to store it. It demonstrates that typed XML really does validate everything...
Continues here.