"Rockville's voting was complicated by a glitch. Thousands of residents who had not yet voted were mistakenly listed as having already cast absentee ballots because of a state database problem ... The state's [voter registration] list inadvertently marked as absentee the names of voters with a home address that begins with the number 5."
My home street address is actually 5, so I might have been affected by this, if we had had an election this week.
While this is an inexcusable occurrence, it is not really that surprising. The greater the complexity in a system, the more likely it is that unexpected glitches such as this will occur. I don't know if this was a software error, a programming error, a configuration error, human error, or something else. And the point is that it does not really matter. Complex software-based systems, especially ones developed the way voting machines today are built and tested, are likely to result in unforeseen problems. In the 2006 primary in Rockville, there were even worse problems with the electronic poll books. In 2008, the problem will probably be something new.
What worries me are not only the problems that we can observe, such as the ones last year and this week. I'm worried about the problems that might result in the wrong votes being totaled, without us ever knowing it. It is foolish to think that only really obvious errors will occur. Maryland's voting system, using only DREs statewide, is vulnerable to catastrophic error but also undetectable errors. And, although the state passed a measure to move to optical scanners in 2010, it now appears that the funding for this move may not be available.