Saturday, September 20, 2003

The IEEE is in the process of setting standards for electronic voting machines. I've been writing about the dangers of confusing technology and social policy. This is a very real and worrisome example. As Cory notes "the standard fails to require or even recommend that voting machines be truly voter verified or verifiable". I skimmed the IEEE site and found a lot of details about programming standards but little about the larger context. It seemed to focus on the components but not the overall system design issues either technical or social. It matters less what telecommunications facilities are used than whether the messages get to the other end. That's the lesson of the Internet.

No election is perfect, after all, we are voting for people and can't be sure what they will really do and don't necessarily understand all the issues ourselves. At very least we must trust that our votes are counted fairly. I can understand the desire for the technical committee to get the job done and move on and appreciate the effort to perform a valuable service. But I worry that the endorsement of an aspect of the technology will be taken as an endorsement of the process as a whole.

You can look at the site for P-1538 and judge for yourself. The IEEE has an important role to play but the standards for voting machine are different than the standards for 802.11 because it is as much social policy as technical policy. The priority must be on giving people confidence in the overall system design and the process even if it requires taking extra time.