Spam, Spam, Spam, Spam…

How many spam messages do I get per week? Hundreds… sometimes thousands and it makes me mad. Very mad. Built in mail filters used to work for me, but then some idiots at some college did some research about how “If you rearrange the lterets in wrods tehy are stlil readable.” Shazam! The flood gates of spam were opened and my poor inbox was flooded. Then came the V1AGRA people and although some people got hard, my inbox went flaccid.

So, donning my military helmet I marched into war to find out what I, a private citizen, could do.

*Don’t give your address out.*

If you have a website, you have probably put your e-mail address on it so nice people can contact you. The problem is there is a multi-billion dollar business that sell addresses like yours to the highest bidders. So take them off an replace them with a link the spammers cant see.

“Encoder”:http://automaticlabs.com/products/enkoderform/ Takes your email address and encodes it so that address harvesters can’t see it. With no exceptions.

NEVER REMOVE!

That’s right, don’t remove yourself from the spammers lists. doing so will only confirm that your e-mail address is “active” and more spam will come streaming in. Spammers are not, regardless of what they say, honest or ethical people. If you push remove you will receive 10 more spam messages.

GA-FILTER FISH!

Good, but what about those of us that have already been sold-out? What can we do? We can add more sophisticated space-age filtering is what we can do, damn it! Filters that learn over time instead of get bypassed by spammers. I am using “SpamSieve”:http://www.c-command.com/spamsieve/ and it is ridiculous how much spam it catches.

Filtered Mail
65 Good Messages
2830 Spam Messages (98%)

SpamSieve Accuracy
1 False Positives
19 False Negatives (95%)
99.3% Correct

Corpus
411 Good Messages
2861 Spam Messages (87%)
62291 Total Words

Showing Statistics Since
6/2/04 9:15 AM

It let 18 spam messages in the first week and since then only one got by. SpamSieve has a complex Bayesian statistical filter (based on an “article”:http://www.paulgraham.com/spam.html by Paul Graham) that learns from all your e-mail for words that spammers use and words your good messages contain and figures out mathematically if it’s spam or not. Ss you can see by the “99.3% Correct” it’s doing pretty well at catching the spam. It only once thought a non-spam e-mail was spam and that was because it was from a website that sold sandals and since it had no data on sandals it assumed it was spam. Honest mistake. I mean how many times have you gotten e-mailed about your flip-flops?

PC users can try to use “SpamBayes”:http://spambayes.sourceforge.net/ but I have no idea what that does, or how it works.

In the end it comes down to how much spam you have. Most people get a few messages a month and that is alright by them to delete, but for us web old-timers, spam can bring your day to a grinding halt and make you wish you never gave your real address on that contest to win a free iPod the year before.

Comments are now closed.