Josh's Weblog: Google Mail and [spam]

Google Mail and [spam]

I've been using Google Mail (gmail) for about a month when an offer came to beta-test moving all the mail-handling for my domain to the system. I get to keep the same email address I've been using for nearly two decades and I get to manage the accounts for my household. Sounds great, right?

Well, as you can imagine, having the same email address for that long means I'm on an awful lot of spam lists. For the most part, Google Mail handles spam pretty well. I was using a combination Amavis and SpamAssassin before, but they don't benefit as much from me training them as Google's userbase training its spam filters.

However, in one particular case it falls hard on its face: spam that has been marked as spam by the Yahoo! mailing lists. Yahoo! mailing lst owners have Yahoo! email aliases, and so they benefit from Yahoo!'s spam-filtering. What Yahoo! does is add “[spam] ” to the beginning of the subject line, allowing most mail clients to filter them accordingly. I am an owner or moderator of many mailing lists at Yahoo! and I get a lot of spam through this channel. I appreciate that Yahoo! performs the service of pre-filtering this email and marking it for me.

Aggravatingly, Google does not recognize mail with “[spam]” in its subject as spam and deal with it accordingly. For over a month I've been training their filter and it hasn't picked up on this pattern. I get 20-50 of these annoying messages a day. Not only that, because spam that gets this kind of treatment is relatively infrequent on Google, individuals training the filter are not likely to have much of an effect. (Google spam training benefits most from mass mailings that are all alike and the votes of thousands of mail readers.)

So I tried to create a filter to find these, but Google's search ignores my brackets! A quick reading of Google's basic and advanced search information doesn't lend us any help either. I tried my normal regular expression stuff I would expect to work and it doesn't work either!

I hope spammers don't figure this out. They could send targeted emails with “[spam]” in the header and escape the filter! I will look into a way to tell the google team that like “ADV:” subject lines, anything with “[spam]” should be autofiltered into the spam folder.

Josh Poulson

Posted Monday, Apr 3 2006 09:02 AM