voting fraud

"Inside A U.S. Election Vote Counting Program" dissects the Diebold voting kiosk software (used in 37 states) and presents some evidence that it intentionally keeps two sets of books -- one in which the votes are recorded, and one from which the results are generated -- in order to intentionally facilitate fraud.

You might look at it like this: Suppose you have votes on paper ballots, and you pile all the paper ballots in room one. Then, you make a copy of all the ballots and put the stack of copies in room 2.

You then leave the door open to room 2, so that people can come in and out, replacing some of the votes in the stack with their own.

You could have some sort of security device that would tell you if any of the copies of votes in room 2 have been changed, but you opt not to.

"The Truth About the Rob-Georgia File" (same author) has an interview with one of the folks who was responsible for deploying this system, who tells a story that lacks basic security measures in just about every way you can imagine. This story pretty well undermines the first story, since clearly this company couldn't find its ass with both hands: they don't sound smart enough to rig an election.

It is somewhat suspicious that the author of these two articles is also hyping his own book on the topic. But here's the (less accusatory, but also far more vague) take on it from EFF and NYT.

Tags: , , , ,

9 Responses:

  1. anonymous says:

    I don't think incompetence is any sign that there isn't fraud. If they were rigging the elections, why would they care about security? The only security that matters in that case, is that they can securely, reliably, and secretly create their own vote totals. Who cares about accuracy when you know the results before the voting begins?

    Incompetence and irregularity greatly benefits anybody who wants to fix the vote -- it creates a cover. It makes any audit trail hard to follow, and creates a large number of false leads for anyone trying to detect fraud. The general sense of chaos within the organization means that employees will be less likely to notice the deliberate irregularities, as they must also hide the vote fixing from their own employees.

  2. anonymous says:

    I remember stories from my grandmother; she'd volunteer every election when she was young, and help run things.

    They had the mechanical machines that, at the end of the night, would show their final tallies on the back. Come night's end, they'd add up the numbers from all the machines and that would be that.

    Only, there were always a bunch of people doing this, since each party needed to have its own people to ensure they weren't getting screwed. And every single time, every single person would get different totals. All using brand new calculators. Invariably, as the story goes, when all was said and done, my grandmother had the right numbers.

    Growing up hearing that shook my faith in any voting system actually working.

    • jwz says:

      Sure, but centralized ballot stuffing is qualitatively worse than simply having a large margin of error.

  3. anonymous says:

    Analysis of the source code from Diebold's publicly available ftp. Linked from NYT.

    Passwords are sent in the clear between the smartcard and terminal. Nice design choice!

    Another choice tidbit:

    All of the data on a storage device is encrypted using a single, hardcoded DES [NBS77] key:

    #define DESKEY ((des_key*)"F2654hD4")

    - mang.

    • harryh says:

      That's only 32 bits of key.
      Aren't DES keys 56 bits?

      • anonymous says:

        Four bits-per-character would only allow sixteen possible options. That doesn't allow much of an alphabet, even less if you include numbers. (No, I haven't forgotten about hexadecimal.)

        I suspect each character in the key represents 8 bits, meaning this:

          #define DESKEY ((des_key*)"F2654hD4")

        represents sixty-four bits.

        Does this remind anyone else of the DeCSS situation oh-so-many years ago? Weak crypto, keys stored improperly, etc... the DVD Content Scrambling System used 40-bit crypto (as required by US export laws at the time), and the key was stored as a 5 letter/number sequence.

        We're going to an electronic voting system here in Washington State soon. I guess we'll know there are problems when His Billness wins the governorship on write-in votes... from one machine... and the votes were all placed within a few seconds of one another.

      • anonymous says:

        The 'h' in there would seem to indicate that it's not really hex. I have a horrifying suspicion that that's the raw binary data for the key, not just a text representation of it...

        -- cnh

      • jautero says:

        It's not hex. There are 8 ASCII characters in the string. That makes 7*8=56 bits.

  4. violentbloom says:

    this is why I stopped voting.

    though I am thinking of going back to vote for governer in california. :) just because it's all so ludacris.