County betting on 'close to fool-proof' electronic voting
By Mark Wheeler / Hi-Desert Star

SAN BERNARDINO - Scott Konopasek is pretty sure of himself, and he's sure of his machines, too, which is probably a good thing. Konopasek is the county registrar of voters, and his machines are the ones San Bernardino County residents will be using to vote in November.

"These machines are as close to fool-proof as any machine can be," the registrar insisted.

He asserted, "All countermeasures have been reduced to zero," meaning something like: All programming operations have been completed to secure the vote against tampering Š at least under real-life conditions.

The "real life" provision is important - crucial, in fact, for one of the accusations against computer voting systems is that they are an election swindle waiting to happen.

Computer scientist Peter Neumann, quoted in a May 2004 "Discover" magazine article entitled "Hacking the Ballot Box," warned that with electronic systems: "You can rig it in an invisible way on a massive scale."
 

A Maryland test carried out by a team of eight computer security experts at the request of the Maryland Department of Legislative Services last January, cited also in the article, found the Diebold system used in that state was vulnerable to sabotage.

"Smart cards," used by the individual voter to log onto the system, were easily forged.

Even more unsettling, during the test, which had been set up to simulate voting day conditions, one team member managed to gain access to the system's memory in 10 seconds. This gave him further access to the keyboard jack which, Mackenzie's article stressed, created a condition where he "could have overwritten the machine's vote tally by plugging a PDA into the jack." (The PDA is a small, handheld data storage device with a keyboard on it.)

Konopasek is thoroughly familiar with this particular study, and though he admits it revealed "some potential, theoretical weaknesses," with at least the Diebold product, he is adamant the test doesn't prove anything about computer voting systems "under real-life conditions."

He criticized the Maryland work for its "clinical setting," accusing it of being too contrived. It's difficult to imagine a non-clinical or contrived test setting, however, since hacking an actual election, the only logical test-setting alternative, would be a hard sell to the voting public.

The voting system technology currently in use has been around for 20 years, according to Konopasek, and he's happy to report, to his knowledge "there has been no instance of voting ballot hacking."

He's not willing to say there isn't the possibility of problems in the use of computer systems, but believes the majority of them will always be attributed to human error.

For this reason, he has 2,400 trained volunteers scheduled for election day in November. In addition, there will be 100 technicians in the field, and a phone bank of 30 support personnel on duty to handle emergencies and to quickly put polling places in touch with technical support.

San Bernardino County is using the Sequoia AVC Edge voting machine, a different product from the Diebold tested and found vulnerable in Maryland.

The Edge was first used in the county on a limited basis during the March primary. November's election will be the first time it is used on a county-wide basis.

Paper ballots will still be available for absentee voters who apply for them when receiving the sample ballot, and for those who ask for them at the polling place.

Although paper ballots will continue to be available into the near future, in Konopasek's estimation, 98 percent of all voting will be electronic within the next 10 years.