eVoting

[andygates]

Here's a nice piece describing one of the current US electronic voting systems' apparently weird vote-flipping behaviour. In a nutshell, "We should not blame voters when the real problem lies with poor usability engineering of electronic voting machine."  Jeez, don't these guys test their kit with non-computer users?  *facepalm*

Comments

[andygates.livejournal.com]

They don't test them at all? Dear hooting gods, no wonder they're fubared. I'd be insisting on a representative sample and getting that sample in the street (slow and expensive, but reliable).

The audit trail is scandalous and exposes them to cries of foul -- again, if I was in the voting machine market, I'd be getting people like Rebecca Mercuri in to consult, and then trumpeting that as my USP. "MunkyVote[TM]! The one IT people trust!"

It can't be hard to do right. Under the hood you've got a disk-encrypted device in a locked box, it does write-only transactions based on open-source checksummed code so we know it's running what it says it's running. Etc, etc. And part of that is that you test the screen templates with the general public so that you know they understand it!