Sunday, April 29, 2007

Voting, Software and (In)competence

There's an interesting article in Wired regarding voting issues with electronic voting machines in Ohio. It's been very interesting to me to watch over the past few years the number of people who have spoken of rigged votes, stolen elections and conspiracies with our elections system. What started in Florida with hanging chads was supposed to culminate in mistake-free elections with the introduction of "modern" technology.

I don't believe that any votes were stolen, any elections went the "wrong" way or that there's a widespread conspiracy in the elections system. I believe there's a much simpler explanation - sheer incompetence. I also believe there's a much simpler solution - open-source.

One of the most troubling points in the article was the claim that Diebold, the voting machine manufacturer, was claiming proprietary trade-secrets regarding the voting machine software, database design and data. I can't quite tell from the article if this was an actual claim or an assumption on the part of the auditor. Either way, this is extraordinarily troubling to me.

Voting is something that we should all take seriously in an almost sacrosanct way. Whatever your political beliefs, or lack thereof, voting is your most fundamental way of expressing your political beliefs.

That anyone, a manufacturer or a consumer of voting machines could believe that there's proprietary secret-sauce within the machine is shameful. The way that machine works must be transparent. Otherwise, there will be no faith left in the elections system at all. Transparency breeds trust, opaqueness breeds fear and distrust.

No elections commission in this country should give away their rights to fully understand the inner workings of any machinery related to the voting process; including the software.

My biggest fear is not that a conspiracy will steal elections - that takes too much coordination and too much secrecy. My fear is rooted in the many software systems I've seen and the possibility, or even the probability, that someone will make a moronic/crappy/idiotic decision in either the design or implentation of the system.

If Cuyahoga County in Ohio wants a true audit of their elections process then they'll demand to have a look at the source code.

Open-source is not just a fight against closed-source vendors. It's not just a new business model. Open-source is something that gives us back the transparency we demand in our government. Leaving our destiny to the magic of mysterious tabulating machines is not how I want our votes counted and you shouldn't either.


Nick said...

I may be way off base, but it was my understanding that the ballot was counted electronically, and machine readable versions were printed at the same time to be fed through certain machines that will read it. Would a business have to reveal its money making strategy if we have the ability to count votes quickly using two different methods?

Dave Dargo said...

This may be how some systems are designed - my understanding is that this is not the case.

Either way, though, I don't think that voting machines or machines that provide criminal evidence should have any "trust me" mystery to them.


Daryl Monge said...

The local system here uses a computer to print the same ballot you would normally fill in by hand. I believe I have read that the Diebold system is fully electronic with no paper trail. Still, even in the paper trail, I bet most voters would not visually audit the printed ballot from the machine before submitting it and therefore the Open Source argument still stands even for those systems.