I finally got to bed this morning at 2:30am. I figured if I was going to set up a RSS feed I might as well update the software that I run the blog on. There were a few glitches, hopefully none of you were online at that time to see them. Let me know if anything is broken.

BTW, technology is important to the Psychology of the Deal when it improves your effectiveness. Playing with shiny new toys may or may not add to your net worth.

Sim’ shared some experiences on fighting spam. Manual spam filters take too much work. Sim’ wouldn’t have a problem whitelisting my many wmail addresses if he let technology set him free.

For smart spam filtering I’ve used POPFile for about a year now. From the english-language page

POPFile is an automatic mail classification tool. Once properly set up and trained, it will work in the background of your computer, scanning mail as it arrives and filing it however you wish. You can give it a simple job, like separating out junk e-mail, or a complicated one – like filing mail into a dozen folders. Think of it as a personal assistant for your inbox.

It uses Naive Bayes filtering, which I first heard about at Paul Graham’s website. In simple terms, it is a content filter that rapidly learns what the spam you receive looks like. The only way spammers can get by these sorts of filters is to change the way they write — to make it indistinguishable from your normal email. If that were to occur it would no longer be spam you might actually want to read it. So if you have an interest in say, film investments, any non-hyped unsolicited email about film investments would actually be interesting to you.

An beneficial side-effect occured for me. I subscribe to a couple of e-newletters that contain some interesting and useful information. The publishers also promote offers to their list and occassionally try a hard sell. When they send a “sales-speak” email, my filters put it in the spam folder. I manually glance at the spam folder to ensure accurate classification.

I have 93.34% accuracy since 16 December 2002. That understates accuracy as I also use POPfile to classify my inbound email into 10 categories (spam and nine others). Manually reclassifying a film investment email from film to invest counts as a classification error. However the filter now sorts all film investment emails as investment related and leaves the rest as film related 100% accurately.

10 Comments