Opinion: Let's impeach e-voting

Computerworld - Were software patches that didn't fix problems but instead changed results applied to electronic voting machines in two Georgia counties? Were the patches applied at the instruction of a top Diebold executive, without informing local election officials?

This charge has been leveled several times since a rather surprising election in which two Democratic candidates had comfortable leads in polls just before Election Day yet lost by substantial margins.

Of course, there's a strong correlation between your degree of suspicion of those results and which party you support. But we should all be frightened if there's no way to prove that tampering didn't occur. And when voting machines are electronic, paperless and proprietary, it's all but impossible to do a recount or check for errors in a way that can uncover a malicious hack (or honest mistake). Tech professionals across the political spectrum should be unhappy with that kind of system.

Election consultant Chris Hood told Rolling Stone magazine that he was working for Diebold in Georgia in 2002 when the head of the company's election division arrived to distribute a patch to workers. That code was applied to only about 5,000 machines in two counties. Hood says it was an unauthorized patch that was kept hidden from state officials. (Diebold says the state approved the update, although state officials have since asked for more information on the patch's effect.)

The Georgia allegations are disturbing but, sadly, not unique. An attorney and IT security consultant last month cited that incident to renew challenges to 2004 Ohio elections, which had a similar mix of paperless Diebold machines and statistically curious results.

In Alabama, questions linger about a supposed "glitch" that caused officials to change the winner of the governor's race six years ago; a research paper presented to the Alabama Political Science Association later described the new, changed results as statistically "anomalous" and outlined a possible scenario of vote-counting fraud.

Paper isn't perfect, as Florida's "hanging chad" fiasco of 2000 painfully demonstrated. And paper systems aren't necessarily 100% secure; it's certainly possible to destroy or alter paper ballots locally. But unprecedented statewide and even nationwide tampering with elections is theoretically possible when a company controls the counting devices without the independent verification that paper receipts provide.

Without paper ballots, there's no way for candidates -- or voters -- to contest a suspicious result, since each person's action must remain anonymous.

There's a simple answer to all of this: paper ballots that are scanned and counted by machines. Automated ballot counting is much quicker than manual counting, but retallying by hand is still possible should the results be doubted. These systems have the added benefit of being able to easily process large numbers of voters when turnout is unusually high, preventing hours-long waits at polls due to a lack of functioning machines. It's a lot easier to ramp up for an unexpected crowd if all you have to do is hand out more paper ballots.

If election officials insist on an electronic system, they should require a paper printout that voters can verify and deposit in a secure storage area. Those paper ballots can then be counted by hand if need be, and electronic results can be confirmed.

Those of us who vote on paperless machines, whether with touch screens or levers, can never be confident that our votes will be properly submitted, registered and counted. That leaves our elections under perpetual suspicion, whichever candidates prevail.

It's time to outlaw any voting machine that doesn't offer the possibility of a paper-based recount.

Sharon Machlis is the managing editor of Computerworld.com. You can reach her at sharon_machlis@computerworld.com.

Read more about Applications in Computerworld's Applications Topic Center.