Diebold Tabulation Software Drops More Votes

Discussion in 'Politics' started by Principessa, Dec 12, 2008.

  1. Principessa

    Gold Member

    Joined:
    Nov 22, 2006
    Messages:
    19,494
    Likes Received:
    28
    Gender:
    Female
    Premier/Diebold Tabulation Software Drops More Votes -- This Time in Ohio

    Officials in Montgomery County in Ohio discovered this week that touch-screen voting machines used in the presidential election failed to tabulate five votes in the city of Trotwood. The machines in question are made by Premier Election Solutions (formerly Diebold Election Systems). The machines print out a voter-verifiable paper audit trail as voters cast their ballots.

    Montgomery County officials discovered that although the five votes were recorded to a memory card inside the voting machine, the votes weren't counted by the tabulation software when the memory card was uploaded to the tabulation server. Premier's Global Election Management System (or GEMS) is the tabulation software that counts votes from memory cards.

    The company's GEMS software is currently at the center of an investigation into dropped votes in a California county and was also the source of a previous problem found in Ohio in Montgomery and Butler counties during the May primary. Officials discovered then that the GEMS system dropped votes if officials tried to load too many memory cards at once. The problem turned out to be a sharing violation on the Premier election servers set up in eleven counties. No votes were lost during the primary as a result of this problem, since officials were aware the machine was rejecting votes when it occurred, but the state sued Premier in August over this and other issues.

    Unlike the primary, officials had no indication in the November general election that the machine was dropping the five votes that were discovered missing this week.

    The votes were not included in the official count that county officials certified to the state. Montgomery officials discovered the missing votes only through a special hand audit they were conducting on order from Secretary of State Jennifer Brunner.

    In an effort to restore public confidence in the voting machines the state uses, Brunner ordered all of Ohio's 88 counties to conduct a graduated audit by randomly selecting precincts and conducting a hand recount of the presidential race on 5 percent of the ballots cast in order to compare the votes against digital tallies and ensure that electronic voting machines are counting votes accurately. If the results of the 5 percent audit differ by more than two votes of the machine count, county officials are required to audit an additional 3 percent of ballots. If this sampling also differs by more than two votes, then officials are required to hand count all of the ballots cast in the county.

    Montgomery County is the first county to have uncovered a specific problem with the voting machines as a result of the audit.

    According to the Dayton Daily News, officials discovered the missing votes when they put the memory cards back into the voting machines to conduct the manual audit. The director of the county board of elections told the Daily News that there is no circumstance, under normal election conditions, that would cause officials to put the memory card back into a voting machine once it's been taken out. They put the memory card back into the machine this time only to conduct the mandatory partial audit required by Brunner.

    I'm a little unclear to me why they needed to put the memory card back into the machine to conduct the audit since they already had paper trails of the votes printed out when voters cast their ballots and they already had the digital tallies from the machines as well. Neither county nor state election officials were available to answer questions about the issue at press time.

    The Premier touchscreen machines are used in 44 Ohio counties.

     
  2. D_Tintagel_Demondong

    Joined:
    Dec 9, 2005
    Messages:
    4,055
    Likes Received:
    8
    This seems to be a server problem and not a problem with the voting machines themselves. I think that touch screen voting machines are clear, quick, and simple and should be employed more. I hope that these glitches don't scare people away.
     
  3. B_spotted_duck

    B_spotted_duck New Member

    Joined:
    Jun 1, 2008
    Messages:
    2,455
    Likes Received:
    1
    Gender:
    Male
    Location:
    west coast liberal-friendly zone
    presumably the same vendor makes the server and the voting machines. think of it all as one big system. i don't trust anything from these guys... they're not exactly neutral:

    Diebold - dKosopedia
    iebold political connections

    In 2003, Walden ODell, chief executive of Diebold, announced that he had been a top fund-raiser for President George W. Bush. In a letter dated August 14, 2003 in which he invited guests to a $1,000-a-plate fund-raiser at his suburban Columbus mansion, O'Dell stated that he is "committed to helping Ohio deliver its electoral votes to the president next year."
    In the invitation, O'Dell also asked guests to consider donating $10,000 each to the Bush reelection campaign. The letter went out the day before Ohio Secretary of State Ken Blackwell, a Cincinnati Republican, was to qualify Diebold, as one of three companies eligible to sell upgraded electronic voting machines to Ohio counties in time for the 2004 election. (The Plain Dealer, August 28, 2003, Voting machine controversy Head of firm seeking Ohio contract committed to Bush victory, Julie Carr Smyth.)
    ...
     
  4. mindseye

    Gold Member

    Joined:
    Apr 9, 2002
    Messages:
    5,685
    Likes Received:
    1
    Gender:
    Male
    No tabulation system is going to be perfect: we've seen well-intentioned human beings spoil (or worse, lose) hand-counted ballots as well. But a tabulation system should be transparent.

    I have confidence in the recount process going on in the Minnesota senate race because of this transparency: the ballots themselves are open, and the counting is open. There are some decisions of the canvassing board with which I disagree (and not always for partisan reasons), but the process by which those decisions are made is clear and open.

    By comparison, I have no confidence in the recount scheduled for this week in Virginia's 5th Congressional district. Since some precincts in this district used paper ballots, and other precincts in the district used touch-screen machines with no paper trail, there can't be any consistency in the recount process, and worse: there can't be any openness. The ballots counted by the touch-screen machine can't be individually examined for voter intent or even to see if the vote was recorded correctly.

    This isn't a problem specific to Diebold (other companies have done this too), but they've contributed to the lack of openness by selling machines without paper trails, and also by hiding their source code as a "trade secret".
     
  5. D_Tintagel_Demondong

    Joined:
    Dec 9, 2005
    Messages:
    4,055
    Likes Received:
    8
    Having worked in data management for years (farming, harvesting, grooming, disseminating), I can say that there is something worse than missing data and that's false data.

    At least these people found out that the data was missing, and can possibly correct the situation--unless some lame "dimpled chad" policy will prevent the votes from counting.

    I'm sure that the conspiracy theorists are all abuzz knowing that the same company runs the whole system. As mindseye said, if it's transparent then it's safe. The software should also do real time error-checking, which it clearly does not. This sort of data should never be queued because all sorts of problems can occur while the data is in stasis. It should be stored and counted as soon as the person votes, and the feedback to the voter (telling them that their vote has been successfully accepted) should be immediate. If this were the case, then this problem wouldn't have happened.
     
  6. Calboner

    Gold Member

    Joined:
    Aug 16, 2007
    Messages:
    9,026
    Albums:
    5
    Likes Received:
    2,465
    Gender:
    Male
    Location:
    USA
    The main point about touchscreen voting machines is that their tabulations are unverifiable. There is no record against which to check the vote count. The mistabulations are only discovered when, e.g., the machines give someone a negative vote or count more votes than there are registered voters. It is like a voting system where you hand your ballot to someone who may or may not count it and who is answerable to absolutely no one. It is simply not reliable, because there is no way to check its accuracy.
     
  7. Hockeytiger

    Hockeytiger Active Member

    Joined:
    Sep 14, 2004
    Messages:
    680
    Likes Received:
    82
    Gender:
    Male
    Location:
    Midwestern US
    The point is not whether these systems are perfect, but rather, are they better than other systems. I honestly have no idea of they are or not, but from my own experience they nearly eliminate voter error, which is a huge step in the right direction. Creating a zero error rate system ought not be the goal here.
     
  8. JustAsking

    Gold Member

    Joined:
    Nov 23, 2004
    Messages:
    3,249
    Likes Received:
    3
    Gender:
    Male
    Location:
    Ohio
    I have been in the software development business for 30 years. I have done everything from realtime embedded control stuff to large scale concurrent data systems.

    Software bugs are a fact of life in the business, especially in large scale systems. However, I believe that it is very possible to make an electronic voting system that has virtually no bugs. The requirements for functionality are simple enough to the point where special techniques and exhaustive testing can be employed to reduce the defects to close to zero.

    The defects I hear about in these systems are understandable in that they appear in areas where software problems are likely to occur. For example, uploading the data from many memory cards at once into a server is a place where people would expect trouble.

    On the other hand, any system designer worth his salt would immediately identify areas like this (heavily loaded concurrency conditions) as a likely place to use proven techniques to mitigate problems.

    So, although I see that the problems occur in the likely areas, I also maintain that those problems will only occur in systems designed either by amateurs or by a company that just doesn't care about quality.

    My conclusion is that given the responsibility and trust we engender on a company like Diebold to make a reliable voting system, they have irresponsibly failed to live up to that trust.
     
  9. Calboner

    Gold Member

    Joined:
    Aug 16, 2007
    Messages:
    9,026
    Albums:
    5
    Likes Received:
    2,465
    Gender:
    Male
    Location:
    USA
    How can you possibly say that they are "better" than other systems for recording votes when there is absolutely no way of verifying whether they have recorded votes accurately or not? That is like saying that your marksmanship is better than mine when you are firing at a target that nobody but you can examine, and we are supposed to take your word for it that you have hit the bullseye every time (or 98% of the time, or whatever). One of these machines could produce a completely bogus tally and there would be no way of proving it.
     
Draft saved Draft deleted