> But calling the .click() through the DOM does <g>.

Hmm, I wonder why you accepted my concern as valid points and
then suggest 'blindly' clicking on links. I want to have control:
My Webcrawler class has a hook for rejecting URLs/links, eg
if they lead to another server. That would be useless if I'd
follow your suggestion.

My Webcrawler wouldn't be the best to crawl forums or sites like
ebay, but that isn't my goal, accepted?

The next steps I'd like to take is perfecting the special word
webcrawler and extracting commands/instructions/information for
bots, that can be embedded in html (eg meta tags). I want it to
be a good webcrawler, no bad email adress grabber, although I
can't prevent someone with vfp knowledge misusing/extending
the base class for that...

Bye, Olaf.