Diebold Accuvote TSx DREs Fail Spectacularly

The Diebold AccuVote TSx, like those used here in Washington county, Wisconsin, were discovered to have a failure rate of between 23% and 38% in Montgomery County, Ohio.

The failure here is very noteworthy because the failure is the failure of the AccuVote TSx to accurately record the ballot within the invisible computer memory. This was not a failure of the poorly designed toilet paper VVPAT printer. This was not a failure within the insecure GEMS server. This was not data corruption created by the unstable Microsoft JET database used by GEMS. This was a failure of the AccuVote TSx DRE to accurately translate the screen touches of the voters (testers) into invisible, electronic ballots stored within the flash memory of the AccuVote TSx.

Here is the source: from the Daytona Daily News of March 20, 2007. The variance in the percentage of failure is whether you count the number of failed machines as 28 (initial count) or only the 14 returned to Diebold Election systems, Inc. for repair.

Of course according the ESI report on the AccuVote TSx on printed page 118 (page 123 of the pdf file), there is a 1 in 4 chance (26% actually) that an invisible ballot stored in the flash memory of the AccuVote TSx DRE will be accurately copied to the removable memory cards. It is the electronic ballots stored on the removable memory cards from which all reports (both local and later from the GEMS server after uploading) will be generated.

I will say it again, for the benefit of the municipal clerks in my county and Ms. Jaszewski, County Clerk of Washington County. I am not opposed to using technology in elections. I am opposed to using unreliable and untested technology in a something as mission critical to the American Republic as an election.

One is forced to ask the question: "Is AccuVote a proper name for this error prone product line?"