I'm using Tidy now, and it's been working pretty well. What I do is use Tidy to convert mal-formed HTML to XHTML and then use the Simple XML methods for my parsing.

Anyone know if this new library is better than that approach?