So what would it take to create a more complex parser?  Is this something we
should think about?  We use HTTrack for LOTS of sites and would make good use
of a better JS parser.