Tuesday, October 31, 2006
Occam's Razor And Electronic Vote Fraud
A diarist at Daily Kos who has used ES&S touch screens to vote in Texas makes the argument that the recently reported Texas and Florida touch-screen voting incidents are not the result of intentional fraud, but of truly crummy design. His strongest argument: Who in their right mind would design vote-stealing software that let the voters see the votes being stolen? (He cites the Princeton Diebold demo as proof that if votes are going to be stolen electronically, they're going to be stolen in ways that can't be readily detected, not even by a paper trail.) And another Kossack chimes in:
Computer touch screens do not function like individual push buttons-they are scanned in rows and columns. (To simplify it) This means that in addition to the problem that Sam mentions about accidentally hitting the button above the one intended, there is also a potential for a matrix scan problem where one button is favored if two are pressed at the same time. It takes extra work in the software to determine if two buttons are being pressed or just one, and if they assume that only one button is pressed at a time they may not have included the extra code. It also depends on how well they tested their product. There is a danger that poll workers incorrectly calibrated the screens, making it harder to get the right candidate selected when the buttons are small. These are all good reasons to avoid touch screens for the present time-they are not a reliable enough system for this task at present. Some of the problems can be fixed with software changes-larger buttons an more space between buttons would solve 99% of the problems. Standards for the design and operation of the machines also need to be developed including specifications for the size and spacing for buttons and mandatory testing to ensure correct button scanning. But the problems reported are ones much more likely related to bad design and testing of the product than outright fraud. Let's fight the battle based on product operation, not on speculation of fraud unless is can be proven. Otherwise we lose credibility in the long term.Personally, I think it's easier for Republicans in charge of state elections commissions to just make sure that poor (and known Democratic) neighborhoods have fewer voting machines and polling places than rich (and known Republican) ones. That's a time-honored trick, and can be 'justified' under the 'budget cutbacks' catchall. But it wouldn't surprise me to see electronic means be used -- or rather, to not see them be used, as they wouldn't want to leave any traces.
More blogs about politics.
