Scott, does the COMMENTS test also catch bogus HTML tags? I've seen rather a lot of spam HTML messages where there are deliberate "bogus tags" like <HUE5MTl> to throw text matching off the scent, whereas because they look like tags, the e-mail client display doesn't show them at all. Text matching on the bogus tags is a waste of energy, because the spammer changes the bogus tags too often, and sometimes every bogus tag is different inside a single message.
Andrew 8) -----Original Message----- From: R. Scott Perry [mailto:[EMAIL PROTECTED] Sent: Thursday, March 20, 2003 4:02 PM To: [EMAIL PROTECTED] Subject: Re: [Declude.JunkMail] Comments Test <snip>Specifically, 1.67 would count a comment like "<BR><!-- some comment --><H1>...", where the comment was embedded between HTML commands. 1.68 won't count those, so even 1 of the comments that the test catches in 1.68 should indicate spam (but, that's assume that "theoretical" matches "real life"). -Scott --- [This E-mail was scanned for viruses by Declude Virus (http://www.declude.com)] --- This E-mail came from the Declude.JunkMail mailing list. To unsubscribe, just send an E-mail to [EMAIL PROTECTED], and type "unsubscribe Declude.JunkMail". The archives can be found at http://www.mail-archive.com.
