Centerfield

San Diego County uses Diebold voting machines. And as in the primary, I will be using absentee voting. Why? Because at least that system uses a Scantron-style ballot where you fill in a circle. If a hand recount is somehow necessary, there will be no question who I voted for.

I don't trust companies like Diebold with my vote. They've been caught and are being sued by the state of California for putting uncertified software on the machines, which have no paper trail.

I'm afraid that e-voting is too new for government officials to have had a chance to think through the process of how it should be managed, from software design and testing, to setup and installation throughout the country. (I could be wrong here, as I have not thoroughly followed this issue.) My gut feeling is that the government, working through third-party companies or agencies, should be very active in certifying the underlying source code of e-voting software and collaborating with the software vendors to thoroughly test it. They already to this to some extent via FDA certification of vendors with whom they contract.

The bugs, I think, should be managable. Although there might be 1 bug for every 100 lines of code, a prioritization process (one that rates bugs that skew the vote count as "severe") might reduce that rate to 1 per 1000 or even 10000 lines. The real concern should be hackers, who have proven themselves to be very industrious. The only thing working against them is time, as they'll not have access to the software before actually sitting at the computer on election day.

If you work in the IT industry, working for an e-voting company would be among the most exciting places to work these days!

Wow, that's scary. It's like old Chicago on a national scale. I heard a story on NPR the other day about this and the consensus seems to be that there is no perfect system of voting and that any one has potential flaws. It seems to me that, at least for federal elections, the federal government needs to be involved in setting some sort of standards for voting machines. At least standardize the system so that people in different parts of the country aren't voting in different ways.

I'm sure computer voting will eliminate fraud every bit as effectively as computers have led us to the age of the paperless office. Oh wait, my bad.

This is like watching a train wreck in slow motion. Two freight trains traveling 5 miles an hour, still miles apart, braking furously, and yet the collision just can't be stopped.

Voting for the President should be done on a uniform and separate simple paper ballot (An opinion expressed by others all across the blogosphere, BTW.) simply because it's a process every American can understand. With any form of computer system, we're left with 99+% of the people really having no reasonable certainty whatsoever how secure the voting is. It will be entirely based on trust of technology.

Conspiracy theorists and numerologists will have a field day. It will become a commonly expressed article of faith and excuse for not voting that "it's all fixed" anyway. Multiple movies over the next few years will include plot threads in which savvy techies game the system with the click of a mouse and the public is none the wiser. Manchurian candidate as the ghost in the machine.

My opinion is that for something so important, you just can't overrate the importance of trust. Paper ballots.

It's true that there's no perfect system. In fact, e-vote can be more accurate and safer than other voting systems, if there is serious testing to make sure it works, including verified printed-out paper ballots as a backup, and spot checks to make sure it's working in 2-5% of precincts.

It's just that right now there are NO checks. E-vote companies do have their software tested, but with completely secret results, limiting its effectiveness. In fact, e-vote vendors work hard to discourage even the kinds of system tests that are routinely done for other important applications. No business will go live with a new application without a long field test cycle , and a robust testing regimen. E-vote is, though.

You can't manage bugs if you never find them. To find bugs, you need, well, robust testing.

I'm sure e-vote COULD be more accurate as well as more efficient, I just think trust is more important.

As long as we use paper ballots, cheating can only be marginal, a few ballots here and there. Since e-voting can't generate a verifiable chain of evidence paper trail, trust will always be lower unless you can get people to eschew the conspiracy mouse-click vote-change theory.

in essence, my argument rests on parsimony. Use paper ballots and then count them seems entirely adequate to the task. The alternative of using buttons, machines, wires, electricity, software, hardware, engineers, bug testing, diagnostics, and so on just doesn't seem like an improvement. i realize that this is the Luddite view, but technophobia is real, and it's not guaranteed wrong.

bk is correct.

I just wanted a chance to say that...cuz it doesn't happen all that much. :)

hehehe

It scary and depressing how certain politicians are ignoring the computer experts AND common sense because they want a voting system that cannot be verified.