Computer science prof argues for a paper trail with e-voting
By ZACHARY A. GOLDFARB
SPECIAL TO WASHINGTON POST
The back cover of "Brave New Ballot," a new book on the young and controversial subject of electronic voting, promises dramatic revelations: "Aviel Rubin, a computer scientist at Johns Hopkins University and a specialist in systems security, knows something the rest of us don't. Maybe we suspected it, maybe we've thought it, but we didn't have proof. Until now."
Declaring that "democracy has never been more vulnerable," the book sells itself as something more than a summary of the bitter battle that has unfolded in recent years: computer scientists and activists who argue that electronic voting machines are vulnerable versus machine vendors and election officials who say the systems are safe.
"Brave New Ballot" comes as the vast majority of the American public will use electronic voting machines in next month's midterm elections -- many for the first time. And primaries this year in Maryland and other states have shown that electronic voting can cause election chaos.
The book begins with author Aviel D. Rubin's revelatory report three years ago on the computer code that ran a popular voting machine made by Diebold Election Systems. The report, which said the machines were designed sloppily and open to tampering, propelled Rubin and the topic of electronic voting into national news.
From there, Rubin tells how a "reluctant activist" became a "crusader" against electronic voting, weaving his argument about why the machines are so problematic into the story.
But while Rubin, who has a Ph.D. in computer science, delivers an engrossing narrative about the intersection of politics, the media, academia, corporate America and technology, in many ways he falters in the presentation of his views on voting machines, the voting industry and the election process.
His stance on electronic voting consists of arguments you have heard before, whether you are a believer, a naysayer or a curious outsider.
Rubin succeeds at describing how he felt the weight of corporate PR masters, government officials, university administrators and challenges to his reputation as he tried to warn the country its elections are at risk.
At one point, a reporter alleges that Rubin had a link to a Diebold competitor, prompting claims of a conflict of interest. Rubin persuasively says it was all one big misunderstanding, but the pressure he felt as a young dad who had recently uprooted his family to take a job at Johns Hopkins -- without tenure -- is palpable.
"How tenuous it all suddenly felt," he writes. "I lay awake, tormented with anxiety about my family, my home, my career."
One can't expect him to be totally dispassionate, but Rubin finds too much time in the book to gripe about his critics. Rubin suggests that if you can't program computer software, you shouldn't be talking about voting machines. "Despite their total lack of familiarity with the cryptography, program verification and formal risk analysis however," he chides in one instance, "election officials don't hesitate to give their opinions on the security and reliability of their voting systems."
Other times, Rubin uses excessively technical explanations for his views or decides to bring sharp focus to something relatively obscure and tangential.
When he clears away the clutter, though, Rubin presents a lucid argument that the machines are flawed. He also offers a solution.
In electronic voting, Rubin explains, voters select their candidates on a touch screen. That information is stored on a disk in the machine, and at the end of the day the data are sent to a central computer that tallies the votes.
How would they know?
Look, he says: Voters have no way of knowing that their votes were recorded the way they intended, and if elections are disputed, as they inevitably are, you can't have a meaningful recount without a paper record. (His fight for paper records -- now required by 27 states -- is covered amply in the book.) Sure, Rubin says, the vendors say the machines are reliable, but when lots of experts say that's not true, can you believe the companies when they're so secretive about how their machines operate?
"Machines must be completely trusted not to fail, not to have been programmed maliciously, and not to have been tampered with," he writes.
Rubin's solution is to add a step: After voters enter their choices on the touch screen, the electronic voting machine should print out a ballot showing voters' choices. Voters could make sure the machine had registered their choices correctly, and if there were a dispute officials could count the paper ballots.
Rubin acknowledges that he has undergone an evolution in thinking about how the real-world use of the technology differs from tests in labs -- not least of which because, as a poll worker, he saw that many voters like the machines, and that they are often easier to use by people with disabilities.
Rubin ends the book with a lecture on the virtue of dissent in democracy, sounding more like a professor of political science than of computer science. This soliloquy and other references to the "totalitarian" and "Orwellian" acts of election officials distract from his core argument that the country's election systems are broken and need fixing.