I don't get it. How can Diebold be consistently so lax? This is not rocket science -- but if they'd treat it as seriously as manned rockets, medical devices or even banking records, we would not be having these problems. Or if we did, we would actually be having followup.
These folks simply are not serious.
Unlike some people, I do beleive in electronic voting, conceptually. I believe the audit trail can be far superior with an electronic record. I believe with an electronic record, each voter should be able to audit their vote online, verifying it was counted. And it should be possible to verify that each counted vote corresponds to a registered voter, and that your vote is one of those registered voters.
The standard for electronic voting should be such that major alarm bells go off at the slightest discrepancy, with redundancy and backup at every stage. If a machine catches fire while you are casting your vote, you should be able to quickly determine whether your vote was tabulated -- and be able to recast if not.
In other words, we can and should expect superior reliability and accuracy over manual efforts, and we should be able to make fraud both virtually impossible, and highly detectable.
This doesn't even seem to be on our radar. We seem to look at the lowest bidder, generally Diebold, and a metric of meeting lax standards -- and then are extremely lax in validating against those standards.
Most importantly, we tolerate a degree of secrecy that I think is unacceptable in this role. If ever Open Source should be a requirement, this is where.