The converter is actually what I would really like to have. I hate parsing. And fixing it when it breaks. It doesn't matter if the code is ugly. You are trying to parse crappy websites after all. Consider sharing.
Well, I only have parsers for slovak news site, 3 parsers for lang-8 (homepage, lang match, footprints), couple for local real estate agents, nothing reusable for english speaking folks, only thing that you could use is probably twitter parser: http://pastebin.com/wpjZ6azZ I did it just recently and I use it to parse DAWN and New Horizont feeds.
