|
Edited on Sat Nov-20-04 09:04 PM by Rumba
I'm working on a letter to send to all my federal and state representatives regarding making election software open source. Text is below. I need a reference to the precinct where the vote tally was observed to be running backwards to fill in here, anyone have one?
This is draft one, any suggestions or feedback are welcome. I'm staying focused on the question of human error and accuracy, and staying away from other issues like fraud, because I think it keeps the message clear.
If you want to use this in its current form or take it and modify it yourself, permission is hereby granted as long as you keep it focused on human error and accuracy. Not that I don't think there are serious questions about other issues, but please write those letters yourself. Once I finalize it I will post the copy.
-=o=- -=o=- -=o=-
The move to electronic vote recording and tabulation offers opportunities to make the American voting system more streamlined and less expensive. Being fast and cheap, however, are less important qualities in our voting system than accuracy in the recording and tabulation of votes. One potential source of inaccuracy we should address is the potential for simple human error to introduce inaccuracies.
With electronic voting systems, an innocent error on the part of a single programmer can affect the accuracy of vote recording and tabulation in hundreds or thousands of precincts. To guard against this, I support making the software that runs our voting process “open source”. The open source software movement is a well established means of ensuring that many eyes can review the software. The benefits of such review are well illustrated by considering one of the reported software errors in the 2004 election.
In the case I refer to, it was noted in INSERT LOCATION that at some point in the vote tabulation process the vote total for at least one candidate was actually decreasing as more votes were tallied.
The source of the problem? A single line of code with two problems. First, the programmer didn’t tell the computer to allow enough memory to store the tally. Like a car odometer that only allows room for three digits, once the total was more than the space to record it the number “rolled over”. Second, the programmer told the computer to allow for a negative tally. When it rolled over, it rolled over to a negative number. Like car mileage, vote tallies are never negative, so why the programmer would do this is quite perplexing.
I can’t imagine a car odometer with only three digits, or one that allows negative numbers, ever making it to market. Anyone who looked at it would recognize that an odometer needs more than three digits, and doesn’t need to record negative numbers. But because the software running on this voting machine was not open for inspection these obvious errors weren’t caught.
The American voting process is one of the most fundamental public commons on which our society is based. Making the software that runs this process open to public inspection can only increase both the accuracy and the public’s confidence in the results. I urge you to support legislation that will result in making this software available for inspection by the technical expertise of the many thousands of Americans who would consider it their civic duty to participate in ensuring that our democracy is based on an accurate election process.
Thank you for your time in hearing my concerns. Sincerely yours,
YOUR_NAME_HERE_WITH_ROOM_FOR_SIGNATURE
|