As I mentioned last week, I attended the meeting of the Guilford County commissioners in an attempt (my third) to get them to wake up and reel in George Gilbert, the county election director.
I have spent almost two year tangling with Mr. Gilbert over the issue of touch-screen electronic voting machines (which they call DREs). First, as a Guilford voter, then as a member of the NC Select Committee on Electronic Voting which drafted a law requiring paper ballots be produced by electronic voting machines that can be verified by the voter when their vote is cast.
Mr. Gilbert is not a fan of mine, having at one point publicly accused me of knowingly making false statements about electronic voting (he later apologized, but he continues to imply that I do not know what I am talking about). Mr. Gilbert feels that because I am not an expert on NC election law, my expertise as a computer systems engineer of 18+ years, does not qualify me to pass judgement on the security and reliability of computer voting machines.
This is, of course, nonsense. An emergency room doctors does not have to be a hospital administrator to recognize a faulty piece of medical equipment.
Mr. Gilbert harbors a deep-seated suspicion of paper ballots and an almost blind faith in digital ballots. His philosophy seems to be that while faulty, error-prone humans cannot be trusted to accurately count pieces of paper, they can be trusted to design incredibly complicated computers software with complete confidence. He often opines the fact that the paper ballot has a 200+ year history of fraud and that we need something new. This ignores the fact that after 200+ years we have gotten pretty damn good at spotting voting fraud and that introducing digital ballots will mean another 200+ years to learn all the ways digital ballots can be tampered with.
Mr. Gilbert has fought tirelessly to prevent paper ballots, which are cheaper and more secure than digital ballots, from being used in our county. He has cited numerous arguments about why touch screen voting machines (which I call "black box voting", or "faith-based voting") are superior to paper-based optical scan. All of these arguments were refuted categorically by computer scientists and computer security professionals who testified before the Select Committee. Mr. Gilbert has now turned to specious claims about cost and plans to buy a completely unreliable hybrid system rather than cheaper, reliable optical scan system.
Mr. Gilbert's claims about how much it will cost the county to buy new voting systems have been all over the place.
When Mr. Gilbert appeared before the the Select Committee (11/23/04), he testified that it would cost $3.4 million to install voting machines that provided a voter-verified paper ballot. On 1/19/05, he told the High Point Enterprise the cost was $9 million. In a News & Record article the cost fell to $8 million. A 8/17/05 High Point Enterprise article has Gilbert claiming $6 million.
I don't know where Mr. Gilbert pulls these numbers from, but I suspect they may come from estimates created by the odious Electon Center, a "non-partisan" organization which takes money secretly from voting machine companies, and holds secret meetings the same companies to help them hire a lobbyist to fight reformers and critics. If the Election Center had anything to do with producing these numbers you can bet a proctologist was involved.
How bad is the system that Mr. Gilbert wishes to foist off on Guilford voters? Pretty bad.
On July 29, 2005 it was reported that certification of the Diebold TSx GEMS v. 1.18.22 had been denied by the Secretary of State, Bruce McPherson. The initial report told of a 10% failure rate due to jammed printers and computer "crashes".
Just 5 days later, the newspapers reported that the failures were twice as bad as originally reported, and the failures were not centered in the printers but were instead software issues. Of the 96 voting machines tested, 19 failed with a total of 21 crashes resulting in a blue screen and messages about an "illegal operation" or a "fatal exception error." Also, 10 machines had a total of 11 printer jams. Nearly one-third of the test machines failed in one way or another.
Now this is not the exact model Mr. Gilbert is planning on buying, but it is a similar model from another vendor who has yet to field test the model. It issues the "toilet paper" roll for storing ballots (instead of printing discrete ballots), which is a major headache when they jam or must be counted by hand.
A North Carolina election director with 17 years experience and well regarded by his peers explained the problems caused by this type of DRE/paper hybrid to the NC House Committee on Election Law back in August:
I want you all to at least see one time what you’re looking at. (He holds up long tape.) This is a voter verifiable paper tape produced from one of the vendors that is currently having their system certified. Here’s a single ballot. This ballot over here contains 55 votes, and so a fairly good-sized roll of paper. I’m not sure now we would handle it if we had to do a manual recount, but I’m sure we would find a way. I would have 3,545 of those tapes that size if I had had this system in place
Can you see the headache that would be caused by 3500+ rolls of tape in the event of a recount?
We are going to pay good money for a seriously defective system? A system that
will cost more money to operate than the more reliable optical-scan systems?
Why would Mr. Gilbert not listen to an election director with so much experience and forgo such a troublesome and expensive system in favor of a proven system like optical scan? I don't know, perhaps Mr. Gilbert can ask the Mr. Gilbert I quoted above that question?
He was the one testifying before the Election Law Committee. Yes, Mr. Gilbert made that statement, opposing the very system he now wishes Guilford tax payers to buy.
It seems clear to me that having lost the fight to keep electronic voting machines paperless, Mr. Gilbert is going to do all he can to thwart the intent of the law by buying a system that will fail badly, in hopes of discrediting paper once and for all. Some people ascribe sinister motives to Mr. Gilbert's actions, but I discount this view utterly. It seems to me that Mr. Gilbert has invested way too much of his ego in being right and is prepared to be right no matter what it takes.
This cannot be permitted. I do not know what Mr. Gilbert's phobia about paper is all about, but he must get over it, or turn his job over to someone not afraid of paper.
It seems to me in the best interest of the voters of Guilford County to have the County Commissioners appoint a special panel to examine Mr. Gilbert's proposal to buy new voting machines for 2006. Such a panel must have election offcials from OUTSIDE Mr. Gilbert's department and must also include computer scientist(s), computer security expert(s) and possibly a citizen's advocate. Actually, this panel should examine not only Mr. Gilbert's proposal, but the entire security auditing procedure used to test and protect voting machines.
Our vote is the currency of our democracy, and like currency, it should be printed on paper.
Update: Speaking of voting machine problems, poll watchers are
reporting machine failures in Lucas and Wood County, Ohio.
Background reading
Use, Security, and Reliability of Electronic Voting System
SAIC Security assessment of Diebold voting machines (purged from the Maryland Sec of State's web site)
RABA Technologies Security Assessment (Purged from the Maryland web site)
Compuware Security Evaluation (Ohio)
Computer scientist Justin Moore's rebuttal to Geroge Gilbert's claims about DREs
Vote-tallying woes lie ahead
Devices may not pay off
Now is the time to press your County BOE members for the best voting equipment. The Board of Elections is leaning towards the...
Tracked: Dec 30, 21:31
The County Commissioners will meet at
Tracked: Jan 16, 22:36
Attention all Voting Activists Weblog: NC Voter
Tracked: Jan 16, 22:44