Spam (officially known as Unsolicited Commercial Email - UCE) can be a problem for anyone with an email address. We have a tool to fight that, called SpamAssassin, and we can put it to work for you.

How does it work?

SpamAssassin (hereafter called SA) scans the email and assigns points to it based on certain characteristics, words, phrases, and the path it took to get here. If the number of points adds up to 5.0 or more, it is classified as spam.

Okay, now what do I do with it?

SpamAssassin adds some lines to the email header showing its scoring analysis. You can have your email client (e.g. Outlook) filter such email and put it into a "spam bucket", separating it from your normal email. We can configure your account so that it attaches the original message, untouched, to a message containing the verbose explanation of the scoring.

I don't even want to see it.

We can configure your account on the email server to summarily punt any spam scoring above some threshold (we suggest starting at 9.0). 

Why don't you throw away all spam?

You should know that the process is not perfect. Sometimes email that is ham (not spam) is classified as spam. That's called a false positive (FP). If your friend sends you an email that is incorrectly marked as spam, you really don't want that summarily punted do you?

In the same way (the more common situation) some spam fails to pick up the necessary 5.0 points to be classified as such, so it slips through. Bear in mind also that one person's spam may be another person's ham.

So it's a hopeless fight, right?

No, not at all. SA has the capability of "learning" what is spam and what is ham. For the technically minded, it's called Bayesian filtering. But it requires some help from you. We can set up a system that allows you to take an email that was classified as ham but should have been marked as spam and drag and drop it into a "should_be_spam" folder. That folder will be read by the mail server and its contents will be used to "train" your Bayes database about what you consider to be spam.

In the same way, if you get some email marked as spam, but it is really from your long lost Aunt Tillie, we can set it up so that you can drag it (after you've read, printed, or replied to it) to a "should_be_ham" folder. It will be used to train SA about what is ham.

We also have a "sledge hammer" for serious or unique problems, called whitelists and blacklists. These can be used to adjust the score by 100 points in the desired direction.

Written by Bob McClure, Consultant to CTC