Skip to main content

View Diary: 20,000 more votes than ballots (Waukesha, 2006) [updated-Tue] (272 comments)

Comment Preferences

  •  I'm dealing with something similar (1+ / 0-)
    Recommended by:
    HudsonValleyMark

    in a  completely different world. The embedded computer and the web app have two different numbers for what should be the same value. The PLC is considered the "authoritative" value - but, but, but, ... why? And if the PLC is authoritative, why isn't the web app simply querying the PLC and reporting the same number? So now the entire historical record contained in the database is "wrong"... It's just insane.

    Causation was, is, and ever shall be a slippery bitch, so we're best sticking with noting the facts

    by jam on Tue Apr 12, 2011 at 09:08:25 PM PDT

    [ Parent ]

    •  yes, probably quite similar (0+ / 0-)

      Even before one gets to the EMS "back-end," there are all sorts of interesting ways in which voting systems can give internally inconsistent results, innocently or through subversion. Little hardware glitches, thread bugs, faulty assumptions buried deep in the source code, user settings that aren't clearly or correctly documented.... And then the EMS can go wrong or get bent in all sorts of ways.

      From memory... you may remember a firestorm of controversy about a precinct in Gahanna, Franklin County, Ohio that reported thousands of votes although it had hundreds of voters. That precinct was using an old push-button Direct Recording Electronic system. Supposedly, the board consulted with the manufacturer and found that the votes were intact on the original system, but that it was possible to replicate the reporting failure through some sequence of actions. I think it was that if you queried for the vote totals and then did something else too soon, one or two bits that were "supposed" to be zero would get set to one, and you would end up with vote counts from Bizarroworld.

      The details (which are out there in a county report somewhere) had a strong ring of "you can't make this sh*t up." However, they were fundamentally unverifiable to anyone without the skill and access to take apart the voting machine, and not entirely verifiable regardless. Super. Neat little case study in the trouble with DREs: maybe you can convince yourself that you got the right answer -- maybe you're even right -- but why would anyone else believe you?

      Maybe I should say that's the problem with elections. Some of our present election procedures give unwarranted confidence; others give warranted confidence to some people, but not necessarily to people who don't participate in them. A county may have dreadful chain-of-custody procedures or brilliant ones, but we can't really tell the difference.

      Blah blah blah. Thanks for humoring me while I wax philosophical, or whatever this was. It seems to me that the problem is much harder and much more interesting than some people who accuse me of dismissing it realize.

Subscribe or Donate to support Daily Kos.

Click here for the mobile view of the site