Hacker News Reader

Comment by tmaly | Hacker News Reader

I was having a discussion about security of the voting systems with my neighbor from Germany. He was of the belief that the mechanical machines without computers were more secure.

I like the type of machines we have in Connecticut. There is a paper ballot that is filled out and then scanned optically. We get a paper ballot that acts as an audit trail. There is always the possibility that the optical scanner contains a bug or something, but the 2000 election showed that even mechanical machines have flaws if not maintained.

I would not prefer the touch screen voting machines as there are always reports of the machines switching the vote. It just seems like more can go wrong with those machines.

View on HN

The voting machine's viability is all about how easily a layman can verify its correctness. This is why any mechanical voting machines are theoretically "better" because the layman can actually SEE it working correctly or not.

To make things worse - you can't really verify your vote unless there is a complete paper trail. Basically you can have one of two things - computerased voting machines without denialbility of vote (impossiblity to sold the vote or impossiblity of getting punished for voting wrong) OR paper ballot voting somewhat augmented by computers.

Reason - we haven't solved basic issue of endpoint security yet!

A friend of mine served as the registrar of voters for a small town in Connecticut. He told me there is a whole procedure they have in place to even handle the ballots once they are placed in the machine. There has to be a chain of custody.

There is always that aspect of it, it someone can get a hold of a box of ballots and somehow slip those into the process, it would be easy to circumvent any machine that relies on a paper audit trail.

A double entry system where you get your paper ballot optically scanned and you also place your vote into a independent system maybe even an electronic one could provide a better check against having just one system as a point of failure?

In MN optical voting processors are OK because the paper ballot is always there to back up the electronic version. Not perfect but a decent compromise.

Have you ever tried to inpelement a modified d'hont method manually? There are always computers involved. Counting hundreds thousands of votes manually will always yield an error that can be exploited. Voting as a concept has changed beyond what mr. Halderman is willing to admit. Let's say you have widespread electronic id and electronic voting. In isolation, this sounds mad. In conjunction with physical voting and common use of eid, not so much.

Seems harder to exploit manual counting of paper ballots than the firmware of some closed source system that's poorly understood by those deploying it.

The larger the election, the more manual counters you will have. Which means the more people you would have to influence in order to rig an election.

Ironically, Russia is particularly good at it.

Manipulating a manual ballot process is easy. Just stuff some ballot boxes. Doing so without leaving evidence all over the place is really hard.

Widespread vote manipulation of electronic systems much, much easier to keep a secret.

Russia tends to leave a massive trail of evidence when it rigs its own elections. It's just that Putin has enough of an iron grip on the Russian press and political system that it doesn't matter.

> Have you ever tried to implement a modified d'hont method manually?

What do you mean? Doesn't it just requiere to perform a few divisions?

> Voting as a concept has changed beyond what mr. Halderman is willing to admit.

I don't understand what you mean. Why specifically can't we use the technology he describes?

> Have you ever tried to inpelement a modified d'hont method manually?

Where in the U.S. is that method used?

How would one exploit a seemingly random error to favor a specific candidate over the other?

Can't touch screens give you a "receipt"? Particularly if it's validated by the voter, this removes the machine-failure hole where the optical scan fails for non-compromised reasons.

Touch screens in Maryland give a receipt that you voted. But the votes themselves are not included on the receipt. The purpose behind this is to mitigate people paying for votes in a very close district. Having a reliable method of determining someone's vote compromises this mitigation.

Is one threat greater than the other? I actually think the threat of vote payment is lower due to the the high risk of the illegal act being exposed. But this is strictly intuition and I have no actual knowledge of the matter.

Edit: in the article, they describe a receipt method that matches your description. But the voter doesn't retain it.

Reply on news.ycombinator.com