More problems arise with 'black-box voting'
Occasionally I'm going to revisit some of my earlier columns to them with new information. Now is one of those times.

Back on December 5, 2003, I wrote about the problems with electronic voting machines. Questionable software, questionable vendors, and tons of problems were surfacing, and these machines were going to count millions of Americans' ballots in this year's elections.

How accurately they would count those ballots is a matter of debate.

Since then, things have gotten worse, which actually makes them better. Huh? you say. That is, so many problems have surfaced that so-called "black-box voting" has been getting pummeled by voters and governments across the country.

Of course, that doesn't help the folks who had the problems.

Florida is the poster child of the voting fiasco, thanks to the poorly designed paper ballots that caused all the ruckus in 2000. So now the state is deploying 'carefully tested' electronic voting machines.

What happened? The testing may not have been careful enough. In a special election for House District 91, which includes parts of Broward and Palm Beach counties, the winner — Ellyn Bogdanoff — won by 12 votes. But electronic voting machines in Broward County recorded 134 "undervotes" (a person comes to vote but doesn't actually cast one — not a likely scenario, since there was only the one contest on the ballot).

Well, Florida knows what to do: recount. In fact, that was required by law. Trouble was, the machines didn't create any kind of paper trail. It was impossible to recount. Oops. Further, secretary of state Glenda Hood had the nerve to say, in a news release published the day after, that, "Both counties utilize touch screen voting equipment which operated properly and provided accurate and timely results."

But it wasn't just Florida, The March 2 primaries were a festival of e-voting errors, especially in California.

San Diego County had all sorts of problems. A reader from the area who experienced them first-hand wrote to me. "Not all of the polls had trouble but many did," she told me. "People were turned away or told to vote elsewhere. Some polls were gotten to work in a couple of hours and others half the day."

Precinct workers who turned on the machines were surprised by unfamiliar screens. There weren't enough IT folks around, so some unofficial volunteers stepped in. "Meanwhile people arrived to vote and there was no backup plan," my reader told me. "They were turned away and told to come back later."

Machines in Georgia had problems also, and some people were told to use paper ballots because the e-voting machines didn't work.

You might argue that neither of these scenarios means the machines are bad — it's just a matter of new technology having a bumpy start. There's no concern about, say, tampering.

Tell that to the Georgia Tech student who found and photographed some Diebold voting machines just sitting, unlocked, in the college's student center the day before the election. Considering that they run Windows, and we've all seen how many security patches are released each month… well, how does that make you feel?

Back to Florida. Miami-Dade and Broward counties discovered their iVotronic machines have a defect — they call it a "glitch," naturally — that prevents them from being properly audited in the event of a recount. Well, as long as no one calls for one in November, no problem. Heck, what are the chances of that happening? (The manufacturer has developed a "workaround" that involves hooking up laptops to the machines. For some reason the idea of a "workaround" for the voting process doesn't leave me with the warm fuzzies.)

Further north, in Maryland, some voters found their voting machines didn't display all the day's races on Super Tuesday. Senator Barbara Mikulski, running for her fourth term, found that at least three counties' machines wouldn't let people vote for Senator. Nice.

These are problems I found with about an hour's worth of searching the Web. There are undoubtedly a lot more. Some were human error — techs who didn't set the new machines up properly. But too many are the result of faulty machines and bad programming (in one case, the software accidentally switched the votes for two candidates; only a routine audit discovered the flaw). Hopefully — hopefully — none is the result of deliberate tampering.

Good news

After the various debacles in the state, California has begun to get its act together. An advisory panel told secretary of state Kevin Shelley to ban the use of 16,000 Diebold machines in four counties. (Shelley might order a statewide ban until all the machines' problems are solved.) He's also considering civil and criminal charges against the company for installing uncertified systems.

More importantly, the state senate passed a bipartisan bill that requires all e-voting machines to create a paper trail — a printed copy of each voter's choices that's stored in the event a recount is needed. (The bill now goes to the Assembly, then to the Governor.)

In my home state of Ohio, voters made enough of a fuss — and enough security flaws were found — to stop the deployment of Diebold machines here until (get this) all 57 problems found are fixed. (Those include voter cards that could be counterfeited, easy-to-guess supervisor passwords — "1111" is the default for Diebold machines nationwide, and unencrypted transmission of results. Great.) Never mind that secretary of state Ken Blackwell had already certified these faulty machines.

Also in Ohio, Diebold's president, Walden O'Dell, had send a letter last year to Ohio Republicans saying he was "committed to helping Ohio deliver its electoral votes to the president next year" — a scary promise from a voting machine-company executive. Now O'Dell admitted saying that was a "huge mistake." (That is, being committed to delivering the votes is all right, but saying so was a mistake.)

And around the country, as more people experience problems with these black-box voting machines, legislatures and task forces are pulling the plug on deploying them until they can be made workable and secure.

Next steps

There's nothing wrong with the basic idea of electronic voting machines. In a lot of ways they can be better than punch cards, slips of paper, and other systems. But until they're done right they present far too much risk.

Here are two steps that need to be taken before these things are used:

1. Require a paper trail. Every electronic voting machine should make a printout of the voters' choices. These would be visible so you can verify your choices, then would be stored in the machine for a recount. (You wouldn't be able to take a copy with you for fear of vote-selling schemes.) Without a paper trail, if the machines screw up there's no way to check them.

2. Make the machines' software available to anyone to see. Far from being a security risk, by allowing thousands of people to view the code, any potential problems could be found and the systems made more secure. This is the same way open-source software works, and the same way encryption technology does — the way it works should not be a secret. In fact, the more people who look at it, the more secure it becomes; 10 thousand people looking for holes are more likely to find them than a small group of programmers.

Face facts: Today's electronic voting machines are unreliable and unsecure. Until the multitude of problems they present are worked out, these systems need to be removed from service, period. This is the foundation of democracy we're talking about — there's a lot more at stake than the egos of politicians, vendors, and election officials.