> The ballots themselves are gpg signed by the voters.
Looks like the voters combine their vote with a personal secret key and find the md5sum,
making it difficult to reverse to break anonymity.
That's fine for a relatively small set of technically competent voters. How would you design
a system that implements this for, say, 100 million voters, half of whom are of below-average
intelligence?
Posted Mar 18, 2008 19:34 UTC (Tue) by tialaramex (subscriber, #21167)
[Link]
Right, the most important thing about any voting system is public confidence. Most people in a
democracy have only a vague and probably erroneous idea of how it actually works, but so long
as they have confidence that it /does/ work it remains viable.
It's much like the situation with runs on the bank. Irrational behavior by investors who've
lost confidence can bring a well run financial institution to its knees in hours, while over
the road a confidence trickster is running a pyramid scam doomed to fail but they can't take
money from investors fast enough.
It's not enough to build a system that math wizards and other experts judge to be sound, it
must be so simple that you can take a dozen people off the street and expect every one of them
to explain correctly how and why it works. That's why so many countries have stuck with marks
written on pieces of paper, stuffing the pieces of paper into sealed metal boxes, and lots of
people standing around putting the papers into piles and counting them. My grandmother can
understand that.
It even aids public understanding when things go wrong. Some guy with a nice suit who
"modified the binary code of a voting machine" and thus changed the outcome of an election is
difficult to understand - a jury may never grasp what he's really accused of doing. But the
equivalent operation on real ballot papers involves a group of people deliberately miscounting
or putting papers in the wrong piles. A child can see that's naughty, and a jury will convict.