A controversial contest at this year's Defcon hacker conference promises to reward the most successful virus writers.

It's called Race-to-Zero, and the idea is to give contestants samples of malcode and viruses that they'll modify so that they can avoid detection by antivirus software. The virus variants will then be uploaded to a closed portal. From there, the variants will be shuttled through several antivirus engines, and the first virus writer, or team, to get their variant past all AV engines wins that event of the overall contest.

This contest isn't administered by Defcon, rather it's one of those "unofficial" events tailgating on the popular conference. It was announced on a popular security mailing list.

Aside from the top winner (and it's not clear how that winner is to be determined), there are many other prizes for what the contest organizers call "notable achievements:"

-- Most elegant obfuscation

-- Dirtiest hack of an obfuscation

-- Comedy value

-- Most deserving of beer

I'm not convinced that this contest will do anything except possibly embarrass the poorly performing antivirus vendors.

First, there are already tens of thousands of new virus samples evaluated every day by the security vendors. Most of these viruses pose little threat. Second, the contest is based on existing samples of virus code, so it's not promoting the development of new viruses. Just variants. Third, the community of people who write viruses are already very collaborative. There are many virus-writing kits, public discussion forums, Web sites, and a robust exchange of ideas under way. So they're already talking and learning from each other.

Fourth: contests like these may actually motivate some young creative genius to get involved in antivirus research and solve the problem once and for all. These signature-based operating system-anchors we've been using for the past 20 years are getting tired.

I'm not losing any sleep over this contest. In fact, maybe the way to look at this is as an antivirus penetration test. Let's keep an eye out to see which AV engines are the easiest to trick.

What do you think?