One For The Money, Two For The Show, Three For The Ballot
Ping over at Useable Security [http://usablesecurity.com/] has a great analysis of Rivest’s ThreeBallot voting system [link to http://usablesecurity.com/2006/09/29/rivests-threeballot-voting-system/ no longer works]. The delightful thing about ThreeBallot is that it should be incredibly easy to implement on a small scale and not much harder on a large scale and has in built in provisions to prevent voter error, counter fraud and vote buying. Ping summarizes how ThreeBallot works far better than I can also discusses some potential issues to consider implementing it in electronic form.
Most people are innumerate, even if they have advanced degrees, so I doubt Rivest’s ThreeBallot system would be accepted as described. This might change if the voter had immediate visual feedback that the vote is properly recorded on the cards. Suppose marked and unmarked spots on the cards pass different amounts of light; when they are stacked, four distinct states can be seen – marks on 0, 1, 2, or 3 of the three cards.
The visually impaired would hate the system, but that seems to be true with current methods as well.
Can anyone make this idea so simple and transparent (pun intended) that the population would accept it?
It’s an extremely clever idea. The problem is it’s too clever. Getting voters to understand and follow this scheme would be difficult. The recent primary election in Washington proved that voters can’t reliably follow instructions to pick a party during primary voting (http://seattlepi.nwsource.com/local/285927_ballot21.html), and Florida in 2000 proved that much simpler forms of balloting can still cause mass confusion. Most people vote once a year at best, and sometimes only once every four years, so they don’t get much practice. The voting method needs to be simple, intuitive, and easy to get right. I’m convinced that ThreeVote would leave a lot of people confused about how they voted.