spam filtering software

 

Filtering spam - Art or Science 

Tis a fine balance, Shakespeare would have said. It is easy, after all, to block 100% of spams. The tricky part is in convincing a dumb machine what is spam and what is ham. And then preventing spammers from convincing your machine otherwise.

There are literally dozens of techniques for filtering spam that will covered at this site. Each and every one of them has their strong points and their weakness. Science and technology are at the roots of every one of those techniques. Very boolean - it either is spam, or it is not. The problem is that no one single technique is without enough weakness to become the single spam killer method.

Again, blocking 100% of spam is quite simple. But the net result offered is unworkable. We can block spam based on the senders address. But what about address spoofing?. Simple enough to block any email with the word viagra - but what about v1agra or vi^gra or any of another dozen variations. True to form, some poor saps are ordered by the boss to block "any and all variations". Instead of insisting that such an effort is a waste of time, the poor wanker takes his orders and heads down a road of miserable failure.

He turns the knobs on whatever they have filtering spam to try and block more. Instead of analyzing the spams to find a weakness, he just starts lowering the threshold - the bar, above which an email is considered spam, and below which is not. Quickly he is issued complaints by people who are not getting emails they expected. And as one might expect, it's old dinosaur brains, the boss, waiting on a confirmation email for his upcoming weekend at the dude ranch.

The poor admin checks the server, only to find dozens of legitimate hams that were blocked by the new spam filter settings. To be sure, there are times when a hammer will solve the problem, and other times, when a little smarter analysis is due. And sometimes a problem is so small that it's worth ignoring all together. It never makes sense to invent hours of work for an admin just so Mrs. Bluehair doesn't mistake an enl^rgment spam as an investment opportunity - nary her eyes have seen, after all.

The point to be made here, is that filtering spam is done with dumb machines that make decisions based on scientific means. The art of it all is in striking the very fine line between letting hams get through, and blocking the spam. At this site, we will cover both the old tried and true methodologies, as well as the new techniques, both solid and burgeoning. Strengths and weaknesses of each, everything that is available today, to help strike that fine balance between letting hams through, while blocking spams.