> Right now when I see HTTrack going too far, I cancel the 
> crawl, add the filter and then select 'Continue 
> Interrupted Download'.  This is best method I know of for 
> the moment.

Yep, this is exactly what I do also.  Seems to work fairly well, but of course
if the site is huge it can take a little while to cancel and again to rebuild
when continuing.