Is Fighting Spam A Lost Cause?
Not really. I believe that for every problem, there's always a solution. It's just a matter of how effective those solutions are.
Today, the most popular anti-spam measure would be the RBLs (Realtime Blackhole Lists). ISPs and end-users now use tools that scan the source of each e-mail and check if the originating IP is listed within any RBL. If the IP is listed with any RBL server, the e-mail is automatically considered as SPAM.
Although using RBLs looks somewhat like an anti-spam measure, it really is not one. RBLs only list IP addresses of e-mail servers that can be used as an open relay. Open relays here refer to mail servers that don't require authorization to receive and send out e-mails.
I can still remember back to the good old days where there were a few public SMTP servers that anyone can use to send out e-mails. However, most spammers have abused these servers and now, there is no such thing as a free SMTP service any longer. However, open relays still exist and spammers now exploit such servers.
Therefore, when a mail server is listed in a RBL, we can only assume that the e-mail that originate from such a server is probably SPAM. Probably, but it is also possible that the e-mail is NOT spam at all!
And even with a number of servers maintaining a comprehensive database of open relays, the amount of spam is still stiffling. Spammers get better everyday at spoofing their e-mail identities. The spoofing technique instantly renders RBLs useless because it replaces the originating e-mail server IP address with that of a legitimate and secure e-mail server. So, spam will still get past the vaunted RBLs.
So, instead of just relying on RBLs alone, some spam filters filter out e-mails that contain specific words or phrases that are commonly found spam mails. This method is a little better compared to just using RBLs. But these simple filters are far too primitive and spammers easily circumvent this countermeasure.
The spam situation today is so bad. It really seems like a lost cause.
But It's Not Over YET!
Yes! It is not a lost battle yet!
Adrian recently sent me a link to a comparison of a few spam filters and one showcased a pretty interesting technique of filtering spams.
The technique was simple yet ingenious.
There is actually a number of spam filters now using this technique. They are called Bayesian Spam Filters and they all make use of a branch of statistical mathematics called Bayesian Analysis to analyse e-mail contents and give an overall score on its contents. That score is used to determine the probability of the e-mail being spam or otherwise.
But that's not the most impressive thing about Bayesian spam filters. Believe it or not, they actually learn and get better and better as they filter out your spam mails!