YEP*.* NOT possible. Not only are various items that are in spam possibly in legitimate mail, and have millions of combinations, but some spammers have horrible spelling. I have heard several people say that this is intentional to bypass content checkers.
I and several others in this group have realized filtering of "wanted" mail due to strict rules on SPAM. You can't even rely on domain names, because many companies buy up "clean" domain names(When I cleaned up some DMOZ data, I found that one company bought THOUSANDS of those domain names. (Historically, and apparantly a VB site, turns out to be a PORN site. Some people have NO conscience) ), and some sullied domain names may be bought up by clean companies. The same has happened with IP addresses.(and new name resolution requirements have just about obliterated IP based checking anyway, some IPs are used by thousands of companies). HECK, I got one message that required me to go to the site and look around before determining it was spam. Steve >-- Original Message -- >From: "Oscar Rylin" <[EMAIL PROTECTED]> >To: <[EMAIL PROTECTED]> >Subject: RE: OT - Re: New Client Code 2.6 >Date: Tue, 26 Nov 2002 21:27:25 +0100 > > >Not possible? ;) >http://www.google.com/search?hl=sv&ie=UTF-8&oe=UTF-8&q=bayesian+filterin >g&lr= > >/me goes back to lurking >------ >>>> From: > [EMAIL PROTECTED] > On Behalf Of [EMAIL PROTECTED] >Sent: den 26 november 2002 16:29 >To: Eric L. Howard; [EMAIL PROTECTED] >Subject: RE: OT - Re: New Client Code 2.6 > >As for it tripping off spam filters? >It isn't possible to determine if something is spam via content, and you >should probably rethink it.
