Printouts latest plan in voting security debate

by Steven T. Dennis
Staff Writer

Nov. 19, 2003

With computer experts still charging that Maryland's $73 million electronic voting system remains highly vulnerable to fraud, a state delegate is proposing adding paper printouts as a check against the machines.

Del. Karen S. Montgomery (D-Dist. 14) of Brookeville has drafted legislation mandating voter-verified paper records, acting on concerns that the Diebold Election Systems machines could be compromised without anyone knowing.

Montgomery's bill would allow voters to correct errors they find on a paper printout of their vote. It also would require random checks of paper records in 2 percent of election districts against the computer records to ensure that there has been no tampering with the computers. The paper records would be used as the final arbiter in the event of a recount.

Montgomery's proposal, which state election officials warned could cost $20 million, came as a Johns Hopkins University computer scientist reiterated his claims that the state's voting system is riddled with holes that could be exploited by hackers or malicious programmers with virtually no chance of being detected.

Aviel D. Rubin's report earlier this year prompted Gov. Robert L. Ehrlich Jr. (R) to ask for an independent investigation by SAIC Corp., which affirmed that the system was "at high risk of compromise."

"Our work is not just a bunch of lunatics at Johns Hopkins making outrageous statements," Rubin told the House Ways and Means Committee last week. He said he would prefer optical scanning machines to the Diebold system.

But Del. LeRoy E. Myers Jr. (R-Dist. 1C) of Clear Spring was not buying Rubin's argument. Myers defeated former House Speaker Casper R. Taylor Jr. by 76 votes last year; the Allegany County portion of that district had Diebold machines, while Washington County used optical scanners.

"I'll take the Diebold-type machines any day of the week over the optical scan," Myers said, noting that paper ballots can be stolen and scanners can fail.

Rubin said that while ballots can be stolen, someone would have to steal an awful lot of ballots to have a huge effect on most elections. But a computer programmer could shift 10 percent of the votes from one candidate to another and no one would be the wiser, he said, especially without voter-verified paper audits of the type Montgomery is seeking.

A recount with the Diebold machines is meaningless, Rubin told the committee, because the computer will simply spit out the same number over and over again, even if it is wrong or fraudulent.

Myers asked if the software code could be protected. It is impossible to guarantee that the code has not been compromised, Rubin replied.

Rubin said the Diebold computer code a version of which he had found on an Internet site was amateurish and incompetent. "I felt like I was an English professor looking at an essay by a foreigner who didn't know English," he said.

Board of Elections Administrator Linda H. Lamone defended the Diebold system, though she professed ignorance as to the technical particulars and acknowledged that the state is trying to hire a computer security expert with a salary in the $60,000 range.

All of the systems are tested extensively before being put into use, Lamone said.

Lamone blasted Rubin and other computer scientists who are critical of electronic voting systems and are calling for paper printouts.

"I think they are doing a great disservice to democracy, frankly," she said, by undermining trust in the system.

Lamone said paper printouts would be too expensive. Diebold has told the state it would cost $1,000 to $1,200 per machine, she said; with more than 16,000 machines statewide, the tab would run to nearly $20 million. Montgomery contends that the cost could be much less.

Frank Schugar, project manager for SAIC, praised Rubin's work and said he is "extremely well-versed and well-qualified, and probably more so than I am personally."

Schugar agreed that someone could tamper with the program and that it would be "extremely difficult to detect," but not impossible.

Schugar refused to answer when asked if the Diebold system passed muster. SAIC's job was to let the state know the risks it is taking. "Whether or not those risks are acceptable is a political decision," he said.

Bob Urosevich, president of Diebold Election Systems, defended the Ohio company's system, saying Diebold takes security very seriously.

He acknowledged that Diebold can add printers to the system, but added that no customers have asked for it and declined to estimate the cost.

Urosevich said the Diebold system has passed extensive independent testing at both the state and federal levels, and said his company had already fixed the security issues found by SAIC. He likened the computer scientists' worries to the hysteria over the Y2K bug.

Rubin "found a key to the house. It's a pretty simple procedure to change the locks," Urosevich said.

Rubin, meanwhile, criticized the state for keeping two-thirds of the SAIC report a secret, arguing that if the system is truly secure, there would be no need for secrecy. State officials said they did not want to release the full report for fear that it would provide a road map for fraud.