no title

Thanks for your text-parser feedback, everyone. Tonight I got through about two thirds of the first wave and made these changes:

• Restyled blockquotes to be full-width, separated from the surrounding text by thin gray horizontal lines.
• Made the “is body text?” decision algorithm more accurate and strict by increasing the weight of a key heuristic.
• Fixed relative links.
• Fixed processing of pages where the entire article is inside a form element. Why do developers do that?
• Fixed specific issues with WSJ, MetaFilter, WikiHow, Asia Times, F.A.Z., Gamasutra, Microsoft TechNet, Slashdot, AppleInsider, and The Globe and Mail.

Please keep the feedback coming. Thanks.