Anti-Bayesian Spam
Friday, Jun 27, 2003
The damned spammers have a new trick up their sleeve that's been foiling my mail client's smart spam filtering. They insert comment tags throughout their email, between words every few letters, so neither my email client's Bayesian logic nor my own explicit filter for 'viagra' will flag it as spam. I've literally been getting about 40 of these emails sidestepping my email program every day for the last couple weeks. Here's an example.

The two simplest solutions seem to be Apple's updating of mail to filter out comment tags in the html portions of email before running its spam filtering, or switching to an email client like Mailsmith that will let me write my own complex rules using regular expressions and perl, so I could make filters like "If the email has more than 4 comment tags" or "If, when all comment tags are removed, one of these keywords exists."

Option three is just filter out every piece of email that has a comment tag in the first place, only a lot of legitimate email has these tags (for no reason but to help the lazy programmer who didn't bother taking it out, even though no reader should ever see it).

Apple? Are you working on this problem?

If you like it, please share it.

Hi, I'm Kevin Fox.
I've been blogging at since 1998.
I can be reached at .

I also have a resume.


I'm co-founder in
a fantastic startup fulfilling the promise of the Internet of Things.

The Imp is a computer and wi-fi connection smaller and cheaper than a memory card.

Find out more.

We're also hiring.


I post most frequently on Twitter as @kfury and on Google Plus.


I've led design at Mozilla Labs, designed Gmail 1.0, Google Reader 2.0, FriendFeed, and a few special projects at Facebook.

©2012 Kevin Fox