I'm heading down the HTML::TokeParser road right now.

Perl Monks do it again.

Thanks very much, Diz.