The standard they apply, and which I like, is "software independence":
A voting system is software-independent if an undetected change or error in its software cannot cause an undetectable change or error in an election outcome.Essentially, that means that the ability of election officials to tally and audit election results cannot depend solely on the reliability of the programming of the voting machine.
News reports after the November elections indicated that some jurisdictions had issues when they went to verify results, like sometimes the machines reported different numbers.
Clearly, this is no way to run a democracy. A vote is a vote, and no matter how many times you count it, the vote should still come out the same way. I realize that in some cases (as we all learned in the aftermath of the 2000 presidential election, dimpled chads and all), there are ambiguities, and reasonable people might interpret the same data differently. But in those cases, humans should be able to examine the evidence and make a ruling, not just accept the "judgement" of a technological system.
I like this quote from the end of the Washington Post's story today:
"Why are we doing this at all? is the question people are asking," said Warren Stewart, policy director of VoteTrustUSA, a group critical of electronic voting systems. "We have a perfectly good system -- the paper-ballot optical-scan system."Anyway, enough ranting for the moment. More to come, obviously.