1/15/2008: Oregon VRC's Follow-up to: "Oregon's Election Systems Decertified by Colorado"

Emailed on 1/15/2008

Dear John,

Thank you for your response. After three years of research and working with the State and the County officials in Oregon, we have some feedback for your points and a few follow up questions.

You first state that according to John Gardner of the Colorado Secretary of State's office, the ES&S decertification was based on "a low level of support provided by ES&S staff" in two Colorado counties. You feel that because Oregon county clerks report being "satisfied and positive" about ES&S support and because ES&S has partnered with a local subcontractor, that the serious deficiencies leading to decertification of ES&S in Colorado are not an issue here.

Recall that the following were given as reasons, among others, to decertify all ES&S equipment and software:

"Failure to prevent and detect normal operator changes within system; Inability to determine if device works correctly; Failure to operate at documented performance levels."

How is it that the relative "satisfaction" toward the service of a vendor or a subcontractor, as reported by county clerks, bears on whether criteria such as the above are passed or failed? As you know, election system vendors and their subcontractors enjoy trade secrecy protection over their equipment and software. Why would "quality of support service" translate into accuracy of vote counting processes that are nevertheless secret? Incidentally, who is the local partner of ES&S? Please let us know so we can learn more about this company.

You might also wish to review John Gardner's background (http://tinyurl.com/2czqgp) before taking too much assurance in his assessment of the Colorado decertification.

The Rocky Mountain News reported on his lack of computer security testing expertise (http://tinyurl.com/og5py) as follows: "Democratic candidate for Sec. of State, Sen. [Ken] Gordon and the Democratic Party were alarmed by a deposition in the case released this week, in which the secretary of state's staffer in charge of testing the machines says he did only 15 minutes of security checks. The staffer, John Gardner Jr., also said he had no college training in computer science, causing Gordon and others to question whether he was qualified for the job. Gardner also had been information technology chief for the El Paso County clerk, which runs elections there."

The plaintiff's attorneys say Gardner's security checks on the four systems did not include attempts at hacking. Instead, Gardner merely checked whether the manufacturers included security documentation. "Of course" Gardner should have tried hacking, [plaintiff attorney Paul] Hultin said. "Isn't that the idea of a test?"

You also noted that Paul Craft, a current Oregon consultant and a member of Colorado's review team, expects that ES&S will be recertified in Colorado prior to this year's elections. You seem reassured by this assertion that the Colorado problems are minor or do not extend to Oregon.

We are dismayed to hear that Paul Craft, a former member of the Florida Division of Elections, is advising Oregon on election equipment. Craft has a conflict of interest in that he is both a private consultant who advocates for ES&S equipment and is also a member of the NASED Voting Systems Board that makes recommendations for certification of such equipment.

According to VoteTrustUSA (http://tinyurl.com/2u7fxd), "Craft is in the position of approving testing by the so-called Independent Testing Authorities (ITAs) which are paid to do the testing of voting systems by the voting system companies themselves. The ITAs have shown on a number of occasions that their first loyalty is to the voting machine vendors.

Craft has been a vocal and ardent supporter of certain voting machine companies, particularly Diebold and ES&S. He appeared to speak on their behalf during the creation of the 2002 Voluntary Voting System Standards. When he was Chief of the Bureau of Voting Systems Certification in Florida, he frequently appeared to act for their benefit, rather than for the public good."

You next report that all Oregon ES&S equipment has met federal NASED voluntary system standards as well as Oregon standards under ORS 246.550 and ORS 246.560. You express satisfaction with the testing conducted by the consultants, in June of 2006. You state that you monitor these tests and are impressed with the vigor of the testing and how the machines are challenged. This "impression" together with your assessment of the ES&S 650's past performance leads you to conclude
that ES&S equipment does not need any further testing for certification in Oregon.

We have already noted our concerns about Paul Craft, a vocal supporter of ES&S whose conflict should be the subject of great concern for your office. Can you please let us know who else acts as a consultant in these tests and whether they are associated with ES&S? Moreover, are these tests public? Even if they are, how would you or any of the county clerks or even members of the public have a basis for being "impressed" or "satisfied" with these tests? Are updates ever inserted into any machines after the vigorous testing but before an election? Are the same vigorous tests run after any and every update? Are the vendors ever left alone with the election equipment?

A growing legion of computer experts such as Dr. Avi Rubin (Johns Hopkins), Dr. Rebecca Mercuri (Harvard), Dr. David Dill (Stanford), and Chuck Herrin have been some of the leading opponents of these private electronic election systems. The results of several independent studies, such as those done by the Brennan Center for Justice at NYU School of Law (http://tinyurl.com/2px4x4), have echoed their concerns by finding numerous, serious flaws and security
problems with the machines and software, including optical scan systems.

You next argue that Oregon has many differences from Colorado in our vote tally process due to our unique VBM system. You say our centralized practice, together with our Centralized Voter Registration System, allows for much greater "control of the tally process." The registration and tally systems are separate, therefore we have a cross reference to insure the "number of ballots received and accepted for counting equals the number of ballots counted by the tally systems."We agree Oregon's VBM, ballot collection, signature verification and processing methods are superior to many states, and these procedures do provide for a centralized process. Our concern is not merely that more or less ballots are counted than are cast, though that is certainly important. (For example, in the 2004 election, Multnomah County's data shows that 501 more Republican ballots were counted than were received.) Our primary concern is that secret software actually counts the ballots. It matters to know that 2 million ballots were cast and 2 million ballots were counted. But it matters even more to know, with confidence, that the outcome produced by the secret software actually matched the outcome dictated by the paper ballots.

Returning to the testing issue, you then assure us that because election officials are required to perform logic and accuracy testing prior to each election and after the counting is done, we should be assured of the accuracy of election outcomes generated by the machines.

The Logic and Accuracy Tests are done on sets of only 1,000 dummy ballots. Instructions within software can take effect after a certain number of ballots have been processed or on specific dates. These instructions can be inserted by direct access to the hardware, by software updates, or through wireless access. This type of coding is similar to that which automatically changes our computer clocks for Daylight Savings Time, where time or an event can trigger the change in logic. We have demonstrated to you and to Secretary of State Bradbury in person how easy it is to access vote counting software on PC
platforms such as those used in Oregon county election offices.

You also mentioned that the Legislature adopted a new law that takes effect this year that requires for each general election a
post-election handcount of select precincts. You mention these will be in addition to "the normal routine recounts" you conduct because every election "inevitably includes a few close races." Finally, you point out that no recount conducted in Oregon has ever turned up evidence that a tally machine failed to correctly count votes. A full recount is the ultimate test, you explain, and with each election there seems to always be at least one or two.

As early supporters of HB 3270 prior to its total amendment by you , we are well aware of the new legislation. It serves to randomly test a handful of machines; it does nothing to verify the machine-tallied outcomes of the elections.

Statistically, the new legislation gives us, at best, a 60% level of confidence that an outcome would not change if a hand count were conducted.

Full recounts are rare, random and optional not "inevitable". Further, in the 2006 Jackson County general election, a citizen noticed that the vote tallies of some of the races were suspect. It prompted the abandonment of the machine count in favor of a full hand count. Kathy Beckett, the County Clerk, said of the incident where candidates' votes were switched, "there are things going on with these machines that shouldn't happen,"[1] and "it's in the program."[2]

You told us that the State turned the investigation over to the vendor who found "nothing wrong," and that you were considering hiring a consultant to do some further investigations. Can you tell us who the consultant was and the findings of that nvestigation? What if the citizen hadn't noticed the suspect tallies? We cannot afford to have the oversight and verification of our elections be random and accidental. We look forward to hearing from you and continuing this important dialog.

Respectfully,

Virginia Ross and Nancy Matela
Oregon Voter Rights Coalition