@links = filter(@links);

sub filter {
#pseudocode
  strip off any hashes because "foo.com/bar.htm#quux" 
  is the same as "foo.com/bar.htm" 

  strip off "index.html" and so on because 
  "foo.com/index.html" is the same as "foo.com/"

  various other things ...

  NOW feed them through a hash to guarantee uniqueness

  NOW grep them against the %seen hash

  return whatever's left


}