What is does is exactly what you say it does: parse a line for the text within an <H1> tag, and frequently screw up. :-)

Slightly more precisely (though not precisely precisely), it matches a line that contains any text, an <H1> tag, possibly another HTML tag, and another arbitrary string of alphanumerics, and the does something with that last string. The circumstances under which it will fail are, alas, too numerous to go into here <wry>--I recommend looking into HTML::TokeParser if you have that option.

If you want a slightly more lucid explanation of the regular expression, you might want to try OGRE, which I believe is written by a fellow monk.