« PHP 5 | Main | Questions to Ask »
July 13, 2004
Say What???
Today in conversation the topic of email spam arose once again. While typically a discussion in how annoying the feature is, the conversation actually focused on some of the techniques currently being used to combat the menace that makes email essentially useless. Tidbits of interest are that the current thought to solving spam isn't by just building filters in email clients, but rather building up a trust network based upon a collection of services and friends.
Anyways, the important point that came out of this was: if we can build a sufficently accurate statistical filter (please note it does not need to be 100% fool proof), at what point does language communication become distrupted when trying to bypass the filter?
We're already starting to see this in current crops of spam messages being sent to INBOX across the world. The subjects contain words with hypens to brace inserted random garbage (i.e. Lev-qjkladjhl-tra), the use of l33t speak (i.e. L3\/1+ra), and the insertion of complete paragraphs that make no logical sense. It hasn't reached the point of being completely unreadable yet, but how far must it go before it does become un-readable?
Posted by Dan at July 13, 2004 10:03 PM