voting machine solution
Would this work?
When you vote, you get a printout of your vote record with a random key #.
Then, after the voting's complete, the full voting records are posted publically on the internet. Anyone who wants can download the records, count votes, whatever.
*And*, you can find your vote record by ID # and verify it's what you voted.
(do a digital dump from the machine to the web as soon as all precincts are closed and the news will have the votes counted faster than the precincts)
Where's the flaw?
"Where's the flaw?"
Well, it would make it harder for criminal cabals like the Bushmen to steal elections. Sounds like a major deal-killer right there... ;-)
All that aside, it actually sounds like a good idea. I like it.
200 million votes, each for up to 25 people.
A lot of data to put on-line.
Oh, and 200 million hits between when the polls close and midnight.
Other than that, it might be pretty clever.
September 23rd, 2006 10:27pm
They're done by precinct or state, not nationally.
Probably 100,000 voters in my county - I think they could survive that. Plus, let's be honest - sure it'll get hammered the first time or two, but after that most people probably won't check their votes unless there's a controversy.
Not a random number, a secure one way hash of your SSN.
Anyone could lay claim to a random number, and dispute how "their" vote was cast. Anyone who does that either gets two votes or bogs down the system horribly slowly. If you use a hash it eliminates this part.
And oh, by the way - define the format that the votes have to be made available (CSV would probably make the most sense), but other than that leave implementation up to the districts. Let vendors compete district-by-district. None of this "everyone use Diebold" crap.
Buncha Oracle, buncha Microsoft, buncha LAMP, buncha home-grown PHP stuff - all over the map. :)
You need a standardized security audit.
"Anyone could lay claim to a random number"
You missed the part about the ticket?
When you vote, a ticket spits out with your number and your vote record. I'd think there should be a process where you do a final "commit" - you've approved what the paper says.
I don't think this discriminates against non-english speakers or non-readers (you have to have *some* level of reading ability to vote, right?) - you can make the tickets as easy to read as the voting machine.
The only dispute can then be that your ticket stub doesn't match the online record. Coming back later to say the stub doesn't match your vote is a bit of a "uh, sorry - you had your chance to dispute that" situation.
"You need a standardized security audit."
For what? Security to protect the votes? I think the process protects the votes better than any process we've had in history - if a break-in is detected, get people to verify their votes online. If they want to challenge their vote, they bring their ticket back to be re-entered.
In another voting discussion I read, it was pointed out that systems like this enable reliable vote buying/coercion. The seller/victim has proof of their vote to redeem. I'm not sure how serious a problem that would be since such tampering would be more obvious than what may be happening with the current system.
September 23rd, 2006 10:37pm
I didn't miss it. Ticket forgery would become standard play, and poll workers would be bribed to change a vote, saying people disputed stuff.
Enables vote buying.
I'll give you $50 for your vote, you give me
the ticket at the end to prove you voted how I paid you.
Physical reciept checked by voter is good, but maybe you shouldn't be able to take it with you.
September 23rd, 2006 10:40pm
Hmm...which is worse, enabling vote buying, or enabling vote theft?
September 23rd, 2006 11:14pm
If the concern is the electronic voting machine can be compromised, why not the online system?
September 23rd, 2006 11:29pm
Full blue sky.
Local news runs a story that "the other guy" won by a surprising margin. Everyone who voted for the expected winner goes online to check their votes against the posted record. The news pulls the posted record and runs a recount (bulk load into any database and quick query will give the count in a few minutes)
I voted for Webb. The online record shows the number on my ticket voting for Webb. The recounts from the posted record all agree with the original result.
There's a possibility of inserting fake voting records, but you can count the voter sign-in roll count against the record count.
It gets tougher and tougher to pull off...
As opposed to the Diebold system, where there is NO WAY to check *anything*.
Or the old punch-card system, where huge blocks of votes can "go missing" and again - no way to check anything. With punch cards how do you know your vote was counted?
A printout with your vote as a secondary backup to confirm the digital version was an option that was shelved. I believe, though, that rather than keeping the slip, you dropped it in a box which was kept & then could be counted by hand if there was any suspicion of foul play.
September 23rd, 2006 11:52pm
The real solution would have been to allow the EU to oversee voting, as they'd done for their own member nations, as they'd offered to do. But we "invented" democracy and there was no way we were going to let anyone tell us how to do it better.
September 23rd, 2006 11:53pm
I know. We can perfect humanity so that everyone is honest.
son of parnas
September 24th, 2006 12:23am
the solution is to have the system work by:
you make the vote, the paper receipt is printed out, you verify it visually, you put the receipt into a box and the _receipts_ are counted.
that stops employers, mafia or anyone with undue power over another from forcing them to vote a certain way, it ensures that you get whatever the benefits of a good computer interface are and it ensures that cheating an election is a lot tougher than simply gaining access to a computer system and toggling 3000000 bits.
jeez guys, if you read slashdot you would understand this already.
September 24th, 2006 3:00am
Or why not both. One paper for the box, one paper with a hash of location, vote and unique ID.
A page to look up your hash in the database can be put online in the event of suspected tampering...or to make it more safe you can take the ticket to an office that does the verification for you after you've ID:ed your self.
But if the box count and the digital count match exactly, the back tracing wouldn't be made available.
"you put the receipt into a box and the _receipts_ are counted."
We do that now, and there are plenty of cases of boxes of ballots "vanishing" before they're counted.
My way provides a method of verifying that your vote is counted as you voted. I find it interesting how many people immediately jump on the vote-buying idea - never would've occurred to me.
I know one simple solution to the vote-buying problem:
Anyone who turns in a vote-buying scam to the FBI will get a reward of 10% of the recovered proceeds upon conviction.
Voila - any vote-buying syndicate that gets big enough to make a difference will get big enough that someone will consider it worthwhile to turn them in.
We should try that for everything. Bang heists, tax evasion, embezzlement, drug dealing..........
September 24th, 2006 9:26pm
The system is still open to vote buying and vote stuffing.
September 25th, 2006 9:26am
>There's a possibility of inserting fake voting records,
>but you can count the voter sign-in roll count against
>the record count.
Supposing they don't match. Do you call foul play or assume it's a mistake? Who do you blame?
September 25th, 2006 9:31am
"The system is still open to vote buying and vote stuffing."
I suggested a solution for vote buying - buy back snitches. :)
Vote stuffing - absent biometric tagging or a universal ID, I'm not sure anyone has come up with a solution for that...