Electronic Voting and the Death of Democracy

If you thought the ballot counting in Florida in 2000 was bad, you ain't seen nothin' yet. Imagine if Gore had called for a recount and... there was nothing to count. No chads to examine and fret over. Just a piece of paper with some totals on them.

That's what we're probably going to get in the rush to replace all these "obsolete" mechanical voting systems with electronic ones. It won't exactly be the death of democracy, but it'll be another straw on the camel's back, or another nail in the coffin (depending on whether you consider democracy dead already or not).

Back in the Jurassic Period when I was a teenager, I wrote a school paper in which I asserted the reliability of computers. "They don't make mistakes. They can't! I've since learned enough about the physics of computer hardware to know that's not strictly true. But it seems to be the mentality behind this headlong rush into e-voting.

My next statement in that paper was, "If they make a mistake it's because the programmer told it to." That's the more important point. Computers do make mistakes. In the time it takes me to type this, countless computers will do something wrong. Most of the time it's because of programmer error. The folks at Microsoft, for example, have churned out megabytes of code that just doesn't work properly. Oh, it works most of the time, but under the right (or wrong) circumstances, you get a blue screen telling you that a program has executed an illegal operation.

With the proper processes in place, and a dedication to Getting It Right, some remarkably secure software has been created. OpenBSD, for example, is an operating system in which only a couple of "remote holes" have been found in the 7 years it's been around. (Compared to at least a couple a month for Windows.) In theory, voting software could meet this standard. But in practice, they don't.

But that's not the most serious problem with all-digital voting. That lies in the other way for a programmer to tell the computer to make a mistake: on purpose.

From time to time, people have tried to sneak these kinds of deliberate mistakes into operating systems like OpenBSD. It's never worked. That's because the "open" in its name refers to an "open source" policy: all of the code contributed to the project is open and available for others to look over. If there are mistakes programmed into it (accidental or intentional), they get fixed.

But most of the developers of electronic voting systems refuse to make their code available for public review. Granted, that would make it easier for people to figure out how to crack it, but A) they'll figure it out anyway (go to the above link for more about that), and B) it also makes it harder to figure out how to fix it. With enough eyes looking over the code, it could become nearly perfect.

To make matters worse, in keeping with the current political fashion of out-sourcing everything to private corporations, the development of voting software is being contracted out to businesses... businesses with their own agendas. Diebold is one of the major players, not just in the e-voting industry, but in supporting the Republican party. Which is how they're getting a lot of the contracts. Kind of a viscious circle going on there.

It would be fairly easy for Diebold to program their voting machines to generate bogus votes for a particular candidate only when/if an important election is coming down to the wire, and then replace the ballot-stuffing code with "correct" code in case the result was disputed. Hell, they could even program the computer to break down in the event it records too many votes for the "wrong" candidate. Or a disgruntled minority-party voter could crack into a system and render every single vote recorded in it (most of which were for the local majority party) suspect and invalid.

With no paper trail, no means of actually verifying the vote count after the fact (yes, that time-consuming, anxiety-producing, but fundamentally-necessary ordeal we went through in Florida 3+ years ago) there's no way of holding these systems accountable.

I'm not actually accusing Diebold of election tampering (at least not yet). But the fact that they have both motive and opportunity to do it is a problem right there, in and of itself. It's that whole "appearance of impropriety" thing, that forbids poll workers from wearing campaign buttons, requires judges to recuse themselves when a case involves their third-cousin-in-law, and so on. The distrust and skepticism that a partisan software developer inspires is just one more reason for people to give up on the democratic system. Add "it's all rigged" on top of "there's no real choices", "they all lie", and "you can't change the system", and you have little reason left for Joe Q. Public to vote.

On the other end of the spectrum is the other half of the problem: the people who aren't skeptical enough. These are people who - like me back in high school - believe that computers won't make mistakes. Sure, they see their Windows system crash periodically, but they figure it's their own fault for not knowing how to use it. (This is underscored by the number of times I get Help Desk calls from people who say "the computer says I've performed an illegal operation" even though the message says it's the application program that's done something wrong.)

The folks at Diebold and elsewhere are taking great pains to make their software easy to use. So easy that voters will assume that everything is going just fine, because the computer didn't crash and it gave them a smiley face and a "Thank you for voting!" at the end. This phenomenon was reported by Avi Rubin, a critic of these systems who volunteered to work the polls in a test of one of them in the Super Tuesday democratic primaries this year.

Once upon a time those in power used literacy tests as a way to screen out uneducated (mostly black) voters, and they justified this as a way of making sure that only "qualified" people voted. The great irony is that this combination of discouraging those smart enough to be skeptical of electronic voting, so that only the suckers bother to show up, may be having the opposite effect.