Printer fails to satisfy e-vote activists

By Rachel Konrad, The Associated Press, February 3, 2005



Excerpt:



The prototype that Diebold Inc. is now touting is exactly what some critics of the ATM-like machines have been demanding for several years.



Even so, paper records alone are not enough to satisfy computer scientists who say transparency in the electronic machines' design and software must complement paper backups.



The Diebold prototype seeks to reassure voters by displaying their selections under a piece of glass or plastic alongside the touch-screen machine. If they spot a problem, they can cancel the ballot and start over. And while voters can't touch the paper records, elections officials will be able to use them to verify close elections.



"Results in the last election reflected the accuracy and security of the (paperless) system," said Diebold spokesman David Bear. "But the fact of the matter is, there are some states that are demanding printers."



After months of criticism by computer scientists that electronic voting systems are unreliable, California and Illinois recently passed laws requiring a paper trail for electronic ballots, and at least 20 other states have considered similar legislation.



Critics of North Canton, Ohio-based Diebold say the AccuView Printer Module is a step in the right direction but doesn't address the potential for buggy software or malfunctioning hardware that could misrecord votes or expose voting systems to hackers, deletions or other disasters.



The printers are only valuable to the extent that counties use them, and critics worry that county elections officials with tight budgets may not opt for them.



Computer scientists also are concerned that the handful of private laboratories licensed to certify voting equipment, including the printer module, still operate in secret and without any federal guidelines.



"It's a very, very small step forward in terms of security of elections," said Avi Rubin, technical director of the Information Security Institute of Johns Hopkins University and co-author of a scathing report on Diebold machines.



Like many computer scientists, he thinks paperless voting systems should be banned.



"I'd say a Diebold machine with a paper trail is better than a Diebold machine without a paper trail, but that's as positive I can be about it," Rubin said.



Diebold stock price rose sharply in the months after the presidential election, when the machines fared far better than critics had predicted. But executives warned investors last week not to expect more dramatic improvements from its voting equipment division. The company's stock hit a 52-week high of $57.75 in mid-January, and closed at $54.91 on Thursday.



About 40 million Americans cast electronic ballots during the Nov. 2 election, but only 2,600 touch screens in Nevada - made by Oakland, Calif.-based Sequoia Voting Systems Inc. - and a few other prototypes around the country produced paper records.



Some of the paperless systems were blamed for high-profile failures in November that included these:



_ In Carteret County, N.C., where paperless machines failed to retain 4,438 votes during early voting, one Democratic incumbent lost by 2,287 votes out of about 3 million cast, and election administrators had to conduct a special election in early January to determine the next agriculture commissioner.



_ About three dozen voters in six states complained that they tried to select Democrat John Kerry, but the touch screens showed them as having voted for President Bush until they revised their ballot. Equipment manufacturers blamed miscalibration.



_ And in a Franklin County, Ohio, a precinct where 638 voters cast ballots in the presidential election, a computer recorded 3,893 extra votes for President Bush. The error was corrected in the certified vote total.



Even with the printer, breakdowns and paper jams are possible, said Stanford University computer scientist David Dill, a leading touch-screen critic.