Sunday, November 14, 2004

This time around, a third of the voting population cast their votes using electronic voting machines. The majority of which produced no records of the votes cast, seperate from the tally constructed by the machines themselves. This sweeping change in the election procedure was deemed an overwhelming success. Why? Well, the machines are very user friendly. And they produced a total number of votes cast that pretty much matched the total numbers of voters that showed up to vote. And almost none of them broke down. Ergo, a success.

Of course, this has to be put in perspective. We live in the 'Information Age' now. Which basically means that we have had to start trusting in machines. Believing in that they will do shat we want them to do. That our cars are actually going 65 mph when the speedometer says so. That we actually owe 1,234 dollars in income tax if our tax software says we do. There is usually no practical way for us to actually check if the machine is actually running correctly. We generally do not have access to information on how the machines actually do their thing to check if these methods are flawed or not. Nor would we have time, let alone the inclination, to do it, even if someone would give us the data. Sure, we might get pulled over for going 85, when we were sure the speedometer read 65, but who is to say that the cop's radar could not be just as faulty as the speedometer? Besides, it must have been we, reading it wrong, because the machines do not make mistakes.

So to get by, in a world embedded with machines that do their magic in a way that is mostly opaque to us, we assume, and trust, and believe. And we also develop our instincts, to use to indicate when the machines might actually have gone awry after all: The blue screen of death, the 'freeze', the 'service needed'. And we have a set of universal tools to deal with these symptoms: The 'quit', the 'reset', the 'reboot'. Taking it in for service. But generally, if it doesn't complain, or show any symptoms, it must be fine.

That's why the electronic voting excercise was such an unabashed success.

And this is one of those rare occasions where my day-time job intersects with another one of my obsessions, politics.

You see, it really is not enough to note that most of the tens of thousands of voting machines did not crash, and that the systems spewed out a final tally roughly equal to the total number of votes cast. No, the real measure of success is how accurately the systems reported how the votes were cast. And it is there that we run into problems.

Surely, there is an independent way to determine this? To seperately verify the results? Nope, not really. Granted, a few states, e.g. Nevada and New Hampshire, adopted a process which produces a paper ballot, seperate of the vote cast electronically. But even there we really do not know if the electronic method worked correctly, because the paper ballots would only have been used if the electronic system had given a close results. See the problem here? Besides, most of the third of the country, that suddenly went electronic this year, did not marry a seperate system with the electronic one. They just have to believe.

I can almost hear you say 'But these must be heavily scrutinized and audited systems'. Sure. Yes. Well, sort of. They are 'certified', for sure. But only by a handful of companies. Contracted by the very makers of the machines themselves. And there is no way for academics, security experts, the public, to do their own verification of the veracity of the machines and their proces. Why? Because the system and its software are secret! A third of the votes cast in the election was cast using a system that is closed to public scrutiny. I kid you not.

And what happened to the aphorism "justice should not only be done, but should manifestly and undoubtedly be seen to be done", usually attributed to Lord Hewart? Well, to quote The President himself, The People just are "not that concerned" about it.

There is some hope, though. For example, a number of people working in the field of information security have voiced their concern. One of them actually worked as an election judge, both in the Super Tuesday primary last March, and now in the general election. He blogged his thoughts about it. Here is a snippet from the March entry:
My biggest fear is that super Tuesday will be viewed as a big success. By all accounts, everyone at my precinct felt that way. The more e-voting is viewed as successful, the more it will be adopted, and the greater the risk when someone decides to actually exploit the weaknesses of these systems.
Maybe the general population will come to its senses. However, judging from the election results, I'm not holding my breath.