Maryland Scraps Diebold Voting System
beadfulthings writes "After eight years and some $65 million, the state of Maryland is taking its first steps to return to an accountable, paper-ballot based voting system. Governor Martin O'Malley has announced an initial outlay of $6.5 million towards the $20 million cost of an optical system which will scan and tally the votes while the paper ballots are retained as a backup. The new (or old) system is expected to be in place by 2010 — or four years before the state finishes paying off the bill for the touch-screen system."
Perhaps the state could sell some of the Diebold devices to help pay the bill that they're stuck with. They may garner a cult following(like the iPhone) of hackers and tinkerers. The devices are worthless as voting machines but they may be coaxed into second lives as kiosk-style internet machines, etc.
Optical scan ballots really aren't a dramatic improvement in reliability.
They can be. Have the touchscreen device print the vote onto the paper ballot, and a barcode with a checksum. Scan it optically and verify it against the checksum.
Anyone can verify their ballot - they simply look at what is marked. Misreads simply don't happen - if the two don't match, there is a problem. Give they guy a new ballot (replacing the old one), and have him do it again.
If the hand recount doesn't _exactly_ match the automated totals, it can be scanned in batches (any size). Count X ballots, scan X ballots. If they don't match, there is a problem.
As a nice side effect, machines don't have to be trusted, and don't have to have a network connection either. The machine can't screw up your vote without marking the wrong thing (or the CRC would be wrong), and you can check that yourself before it's counted. Recounts can be done by hand, and in the event of total system failure, you can still mark the silly thing by hand.
As a nice bonus to this, you get the benefits of touchscreen voting - secret ballots for the blind (audio), multiple language support, pictures, the ability to offer more in-depth descriptions of line items, etc.
It's not exactly rocket science.