One of my three sons wrote a paper on Voting Machines that briefly discusses David Chaum's logical approach to this challenge.
Below are three paragraphs from my son's paper. I am posting this not because I am just trying to fill up my blog :-) , but the three paragraphs below do clearly and concisely state a logical approach to voting machines.
"A new more reliable voting machine has been developed by David Chaum, in which you physically type in the name of the person that you are voting for. When your choice is confirmed 2 receipts that look like a random scatter of squares print out, although you only take one. The other receipt drops down into the machine, and it is stored, in case your vote needs to be recounted. Your receipt is specific to the card you didn't choose, and the candidate you voted for is saved from being lost. I felt this was an ingenious idea to keep people's votes from being left out and made it fairly easy to recount them. You can even check online by typing in the serial code on the receipt into a web site to find out if the person you wanted to vote for got your vote. I like the idea of a receipt that is merely an encrypted card; it doesn't tell people who you voted for, but it is simply used to verify that the vote that is cast belongs to that specific card.
These voting machines can affect our system of democracy in both expected and unexpected ways. The obvious way is the technical errors; despite what we would like to believe, machines are not perfect. They do break down, and they can make mistakes. While it does reduce the human error of physically losing a paper vote, it creates a whole new set of possible errors, such as casting a vote twice, not casting it at all, or even casting it for the wrong candidate. This could cause the wrong person to win an election that maybe should have gone to the other candidate. This would certainly affect our democratic election. Such problems on national scale would not go unnoticed and the proper actions would be taken to correct the problem, but the few who do experience these problems probably never know it. The lack of a paper trail means we are putting our vote into something that we cannot physically see or touch, and this can frighten most people.
This leads to another unseen impact the machines have on democracy: people's willingness to vote. In the states where only machines are allowed, a person with very little confidence in the credibility of these machines, may feel so inclined as to not even show up to vote on Election Day. This does not apply to everyone, but I'm sure it is very possible. This is why I feel the best solution is to get the most reliable machines to cast our votes in election, but always provide paper to the people who want it."
Below are three paragraphs from my son's paper. I am posting this not because I am just trying to fill up my blog :-) , but the three paragraphs below do clearly and concisely state a logical approach to voting machines.
"A new more reliable voting machine has been developed by David Chaum, in which you physically type in the name of the person that you are voting for. When your choice is confirmed 2 receipts that look like a random scatter of squares print out, although you only take one. The other receipt drops down into the machine, and it is stored, in case your vote needs to be recounted. Your receipt is specific to the card you didn't choose, and the candidate you voted for is saved from being lost. I felt this was an ingenious idea to keep people's votes from being left out and made it fairly easy to recount them. You can even check online by typing in the serial code on the receipt into a web site to find out if the person you wanted to vote for got your vote. I like the idea of a receipt that is merely an encrypted card; it doesn't tell people who you voted for, but it is simply used to verify that the vote that is cast belongs to that specific card.
These voting machines can affect our system of democracy in both expected and unexpected ways. The obvious way is the technical errors; despite what we would like to believe, machines are not perfect. They do break down, and they can make mistakes. While it does reduce the human error of physically losing a paper vote, it creates a whole new set of possible errors, such as casting a vote twice, not casting it at all, or even casting it for the wrong candidate. This could cause the wrong person to win an election that maybe should have gone to the other candidate. This would certainly affect our democratic election. Such problems on national scale would not go unnoticed and the proper actions would be taken to correct the problem, but the few who do experience these problems probably never know it. The lack of a paper trail means we are putting our vote into something that we cannot physically see or touch, and this can frighten most people.
This leads to another unseen impact the machines have on democracy: people's willingness to vote. In the states where only machines are allowed, a person with very little confidence in the credibility of these machines, may feel so inclined as to not even show up to vote on Election Day. This does not apply to everyone, but I'm sure it is very possible. This is why I feel the best solution is to get the most reliable machines to cast our votes in election, but always provide paper to the people who want it."
This scheme misses one important property that is a requirement. It must not be possible to prove to a third party which candidate you voted for. This property prevents vote sales and coercion.
Posted by Brian Utterback on November 24, 2008 at 11:57 AM EST #
I amazes me that people get a receipt from the gas station pump yet don't demand a verifiable receipt from voting machines.
Posted by Neil on November 25, 2008 at 07:08 PM EST #
Brian,
You're right, but you can not have it both ways. You can not provide a mechanism for the voter to verify their own vote, but then somehow limit the voter (who has sold their vote) with the ability to prove it to the payer. The selling of votes is a corner case compared to the importance of individual voters to verify their votes have been counted.
Posted by David Edstrom on November 26, 2008 at 12:06 AM EST #
You can have it both ways. Ron Rivest's "Three Ballot Voting System" provided a way to prevent miscounts and allowed vote verification and prevented 3rd party verification. The scheme is not proof against both simultaneously and so not a good real world scheme, but it does act as a good proof of concept that such a scheme may be implementable in the real world.
Posted by Brian Utterback on November 26, 2008 at 08:38 AM EST #
Brian,
Thanks for pointing me at Ron Rivest's: The ThreeBallot Voting System" which I did read.
First, even Ron Rivest admits there are ways to foil his system in Section 4.4 The “Three-Pattern Attack" in his white paper.
But, while his paper is very interesting, it simply fails the "can it be implemented" test. Picture your average senior citizen asking questions on this system works :-)
Regards,
--Dave
Posted by David Edstrom on November 30, 2008 at 09:20 PM EST #