Spam Should Be Driving Technical Innovation

The biggest area where we need technological innovation in e-mail isn't in scalability or functionality but in spam prevention.

The biggest area where we need technological innovation in e-mail isnt in scalability or functionality but in spam prevention. Although Microsoft Corp.s Exchange Server 2003 beta (see review) has an improved virus-scanning API, virus and spam checking is an area that Microsoft leaves to third parties.

The problem is that spam isnt a peripheral issue anymore. Spam isnt just e-mails annoying little brother, who always hangs around wanting to be part of the action; its a substantial threat to enterprise productivity.

I use a Microsoft Outlook client-side spam filter, Cloudmark Inc.s SpamNet, to help block spam. According to its statistics, 27 percent of the e-mail I get is in Cloudmarks spam signature database. (Last month, that worked out to be about 600 messages.)

In addition, I still need to manually delete about 20 to 40 spam messages a day, based on a count of spam in my Deleted Items mail folder.

One way or another, organizations pay for spam filtering. Its just a question of how inefficient and costly the process is going to be.

In addition, making employees do their own spam filtering might be a source of legal liability.

Sexual harassment policies routinely forbid the distribution or display of the kind of material that regularly shows up in my in-box. Im not aware of any lawsuits over this issue, but I think IT departments that take no reasonable precautions to block this type of e-mail from being sent to staff are living on borrowed time.

On Jan. 17, the Spam Conference hosted at the Massachusetts Institute of Technology ( will provide a forum for the latest in anti-spam research, such as statistical analysis of message content using Bayesian filters.

Current techniques of keyword filtering and server blacklisting have ongoing problems with false negatives and false positives. Meanwhile, white-list approaches such as the Turing Test used by Spam Arrest LLC are another promising approach.

These techniques can be applied on a front-end filtering mail server in front of an organizations regular e-mail infrastructure, something that conscientious IT staff should certainly pursue this year if they havent already done so.

West Coast Technical Director Timothy Dyck can be reached at