I think the only way to detect unwanted links is to capture the page output and parse it. It's just too easy to obfuscate, use filters, do complicated stuff that a code parser will never find. Ozh -- http://planetozh.com/