Glitch Found in Ohio Counting
By JOHN SCHWARTZ  New York Times

 November 6, 2004 
 An electronic voting machine in hotly contested Ohio added 3,893 votes to President Bush's tally in a suburban Columbus precinct, even though there are just 800 voters there.

The error was discovered in preliminary vote counts from Tuesday night, and local officials say it would have been caught in any case and corrected in the final count now under way. But the glitch fed the rumors that have been flying across the Internet since Election Day that results were tipped by high-tech voting machines.

Preliminary counts show Mr. Bush won Ohio by about 137,000 votes out of roughly 5.5 million cast. Senator John Kerry conceded after determining that recounts and provisional ballots would not give him enough votes to change that result.

Experts in computer security have found flaws in the software and physical design of new touch-screen voting machines from every major maker, and they have sparked a broad campaign demanding that electronic voting systems include a printer to provide voters with a greater sense of security that their votes have been recorded and to allow officials to conduct recounts.

But Ohio uses very few of the new machines. The machines in Franklin County, which includes Columbus, are made by Danaher Controls, Inc., and have been in use since 1992.

Unlike more modern touch-screen machines sold by companies like Diebold Election Systems, the Danaher machines have switches that voters press to make their choices. The decisions are recorded on eight separate memory chips, some of which are on a removable cartridge that is taken to the election center for counting votes.

The mistake on Tuesday occurred when the cartridge reader misread the totals, said Matthew M. Damschroder, director of the Franklin County Board of Elections. He called it "some kind of hiccup" and said that the equipment performed properly yesterday when the cartridge was re-read. An industry official familiar with the Danaher systems said he had never seen the cartridges provide a false reading before.

The problem, first reported by The Columbus Dispatch, is one of many election night glitches that have surfaced in several states. But all problems so far fall into a class described by Doug Chapin, director of the Election Reform Information Project, as "no big and lots of littles," with no discernible effect on the outcome.

But Aviel D. Rubin, a professor of computer science at Johns Hopkins University, said the relatively smooth election should not be taken as a green light for electronic voting.

"It is like asking a surgeon who states that a particular medical procedure is risky whether he might change his opinion because there was a successful operation using that procedure somewhere in the world," Professor Rubin said.

Lack of a paper trail has helped feed rumors of fraud, said David L. Dill, a computer science professor at Stanford University. Without the "safety net" of voter verification, voting technology "has to be near perfect," he said.

"If we get this many reports, we obviously haven't reached this state of near perfection," he said.