TodayBetter ballot boxes? Not so fast
Print
Share
Email to a friend
Add This
Article Rating
Election Day is usually a lousy time to publish an op-ed column.
It's too late to write anything useful or informative about the current election. But since election issues dominate the news, it's tough to find anything non-political to write about that's also current.
Nevertheless, I've found a topic for today that's not only political, but appropriate for the day: counting ballots by computer.
You'll recall the 2000 election debacle in Florida, which put "hanging chads" in the general vocabulary and has spurred a switch to new technology at our polling places.
But however unreliable some of today's vote-counting mechanisms may be, we shouldn't replace them with newer systems that may be less accurate, less secure or more subject to fraud and abuse.
And the gathering consensus is that computerized voting terminals, so far, are all of the above. In fact, the term "black box voting" is starting to come into vogue to describe computerized balloting.
"Black boxes" are devices that the 1970s predecessors of today's hackers built to steal long-distance phone calls — and that analogy isn't accidental.
For starters, the computer ballot terminals going into use now have no paper trail. Mechanical voting machines punch a tape that can be used for recounts, and of course we here in Cumberland County have good, old-fashioned paper ballots. If the ballot scanner numbers are doubted, officials simply can hand-count the ballots.
With the computer ballot terminals, you read the numbers off the screen at the end of the day. If you doubt them, tough. Read them again.
This is a problem when you consider the computer can show you what you chose before you click the "vote" button, then post an entirely different result in its totals. Imagine if the software was programmed to multiply Democratic totals by 1.15 and Republican totals by 0.85, for example.
That supposedly could be prevented by having the software independently tested, verified and certified. That was done in the 2002 elections in Georgia, the first state to roll out a significant number of touch-screen computer balloting terminals.
But Wired magazine reports an employee of a subcontractor to Diebold, the company that made the terminals, claims Diebold modified the software after it was certified and didn't tell anybody.
The employee is quoted as saying the machines were generating an error rate approaching 25 percent, which would put in doubt several statewide elections that were all upsets when compared to late poll results, including races for governor and senator.
Diebold denies the charges.
Meanwhile, a portion of a new book on the Internet says Diebold posted the source files to all its ballot box software on an unsecured FTP site, where anyone with Internet access could download it. In the hands of an experienced programmer, the software could be modified and reposted back on Diebold's site, and the company might not know anything had happened.
Bev Harris, the author of the book-in-progress called "Black Box Voting," demonstrates in Chapter 7 that one of the files exposed to the public that way was a software updater called "rob-georgia.zip." Meaningless? You decide.
Harris also reports that the vote totals in a March 2002 primary in San Luis Obispo, Calif., appeared on a Diebold website while the polls were still open.
Diebold's not the only firm with these problems. Just a couple of weeks ago, the operating system files for machines made by Sequoia Voting Systems were found on the unsecured server of a Sequoia reseller.
Then there are the usual human frailties imposed on top of these questionable scenarios. Imagine, for example, how upset Californians might be if Gov. Gray Davis had won the recall election — and then they found out that California had been buying its voting machines for the past five years from a company partly owned by Davis.
Substitute Nebraska and Sen. Chuck Hagel in the story above, and you'll have an anecdote that actually occurred, according to The Hill, a newspaper that covers Congress.
The Hill noted that in Hagel's first election in 1996 he won despite being an underdog in both the primary and general elections. But he kept his ownership stake in voting machine company American Information Systems, now called ES&S, out of his Senate disclosure forms. The Senate Ethics Committee never has held hearings on the issue.
It's unlikely that Hagel fixed his own election. But it's not impossible, as we have already seen.
Before we go gung-ho for computerized voting, there needs to be a process, open to scrutiny, in which machines are shown to be reliable and secure from unauthorized access. The software used on them must be independently certified to assure that it does nothing but record and count the votes people actually make.
Most importantly, there must be some sort of backup procedure to guard against data loss from unexpected computer crashes and to facilitate verifiable recounts.
Until then, Cumberland County's paper ballots are as close to the cutting edge as elections need to be.
