Researchers have discovered that, by turning one of the tools used by spammers to generate difficult-to-filter mail, they can achieve a higher success rate in blocking those messages.
According to the post over at New Scientist, an awful lot of spam relies on using a relatively simple template-structure to generate a massive amount of variation across mails without a massive amount of work going into actually writing new spam messages. It’s a simple enough idea, but the new method of combating it is, surprisingly, just as simple.
Apparently, the folks at International Computer Science Institute in Berkley have hit on the idea of simply collating a bundle of mails (around 1,000) generated by such a template, they were able to effectively reverse engineer the template, using the mild variations from one spam message to another to figure out what parts of the messages were static to block any further spam from the same source with, they boast, 100% accuracy.
It’s a fairly interesting model, and one that certainly works a lot better than my own “fight fire with fire” policy (whereby I simply forward all of the spam that I receive to everything that sends me any spam, which was a great idea with post, but less good with email). Of course, the fact that it apparently requires a bank of 1,000 spam messages to work with means that there’s not really an awful lot that those of us frustrated by one or two felonious emails a week can do with it.
Still, we’ll be curious to see further developments. It’d be interesting to see a service like Hotmail or Gmail employing something similar, in which case the fact that both have such wide pools of users will stand them very well indeed.