And when did this mail-in vote thing get soooo fucked up? It should be VERY tightly restricted as to who can mail-in their vote with VERY strict requirements why they can't haul their ass to a polling station. Military, etc.
And both/all parties sign off the count at the polling station.
But you still need to tally the counts from the stations and do it in a secure manner and verifiably - this stage is can be done securely and verifiably by computers - working with cryptographically multisign (sign per party) reports from the polling station. This is secure because when the multisigned reports from the stations are public anyone can verify the tallying with a simple program and just as well verify that each report is signed by a delegate of someone they trust.
Just as well the provisional counting results should be officially signed and published to detect irregularities/fraud.
Add a clean voter file, clear audit procedure (for example, if the turnout is suspiciously high - check with random people who are marked as "voted" if they really voted or not) and the elections become trustworthy again.
The problem with human counting is they can 'count' however they feel like counting. My solution is paper ballots scanned by several parties.
Scanners can be made small and inexpensively, so I suggest a machine each voter puts their paper ballot into that moves the ballot past several scanners. An official one from the government, one from each party, and civic organizations can have scanners. The official scanner checks to make sure it's the correct ballot for that district and each scanner sends the voting machine what it just read from the ballot. The voting machine then compares all the scanners output and if they all match, and if the official scanner says it's the correct ballot, the voting machine puts it in the box. If it's the wrong ballot or the scanners don't all agree the voting machine spits the ballot back out and indicates the ballot has an error. Each scanner records what it read and whether the ballot was accepted or rejected.
At the end of the night several parties will have the same raw data and can demand a hand recount if the official results don't match what they recorded.
None of the machines would be allowed to have any wireless capabilities and wouldn't be connected to any network until after the polls close. More could be done, like scan a label for each box as they're changed so everyone has a record of exactly how many boxes there are and how many ballots are in each box. The voting machine could also mark an edge of each ballot it accepts, the machine would be physically incapable of marking anywhere people mark to vote, and the ballot would be rejected if someone tries to run it through a second time.
That's not crucial and is there to stop any party from knowing the results in advance. Exit polls and counting how many people registered to which party have already voted already gives a pretty good idea of that already. If it is necessary, on election day morning, before the polls open, some of the scanners could be randomly replaced and set aside to be opened up and their circuit board inspected. Have mandatory prison time that's enforced for the head of whichever organization is found to have equipment that can transmit wirelessly.
That's a good point. I guess when you put your ballot into the machine and it accepts it, you know your vote was counted. None of this nonsense in Georgia where thousands ballots were found to not have been counted and no one can figure out why.
I disagree. There are some really good ways to use machines and some really bad ones (e.g. Dominion).
For example, we could have paper votes counted by machines from both parties. If the vote result doesn't match, a team of inspectors can verify the suspect batch. If they quickly process verified batches like this it will be much faster and more accurate than human counting.
These problems are easily solved. The team of inspectors will be made from members of both parties of course, overseeing the two machines. The machines can be sitting right next to each other. This totally removes the incentive to rig a machine as it will be obvious from the first batch counted.
I would argue that software based voting is very much viable, if its designed to be fully audit-able using cryptographic signatures, and fully verifiable and fully transparent. With features like duplicate ballot detection, full admin log, and full operator log, connection to upload servers only via a secure VPN, admin access only from console or a secure terminal, all data transfers from the voting machines
one-way upstream, and all data transfer to analytics and external endpoints one-way downstream.
The software provider must be fully liable for security and compromise of the servers.
It wont be too hard to create a fool proof system as long as that is the actual goal of the software. Dominion was created with the express purpose of being able to cheat on voting without getting caught. But luckily they failed in that as well.
Until someone finds a way to remote hack paper and ink I'd stick to it. Security aside it keeps money out of the hands of parasitic busybodies looking for easy government contracts.
I would prefer to use public hence auditable system with cryptographic signatures for tallying data from the polling stations up. But actual voting to be still done in person, with photo id and good old pen and paper.
Afteral,, that's what the most well known auditable and "everything cryptograhical signed" "verifiable online elections" system Helios recommends for general elections.
Absolutely.
And when did this mail-in vote thing get soooo fucked up? It should be VERY tightly restricted as to who can mail-in their vote with VERY strict requirements why they can't haul their ass to a polling station. Military, etc.
With fully certified voter registration I.D cards without which you cannot vote
Honestly, id be happy with voter ID. One vote per one live person, and secure machines NOT connected to the internet. Best of both worlds.
And in person with photo id.
And both/all parties sign off the count at the polling station.
But you still need to tally the counts from the stations and do it in a secure manner and verifiably - this stage is can be done securely and verifiably by computers - working with cryptographically multisign (sign per party) reports from the polling station. This is secure because when the multisigned reports from the stations are public anyone can verify the tallying with a simple program and just as well verify that each report is signed by a delegate of someone they trust.
Just as well the provisional counting results should be officially signed and published to detect irregularities/fraud.
Add a clean voter file, clear audit procedure (for example, if the turnout is suspiciously high - check with random people who are marked as "voted" if they really voted or not) and the elections become trustworthy again.
The problem with human counting is they can 'count' however they feel like counting. My solution is paper ballots scanned by several parties.
Scanners can be made small and inexpensively, so I suggest a machine each voter puts their paper ballot into that moves the ballot past several scanners. An official one from the government, one from each party, and civic organizations can have scanners. The official scanner checks to make sure it's the correct ballot for that district and each scanner sends the voting machine what it just read from the ballot. The voting machine then compares all the scanners output and if they all match, and if the official scanner says it's the correct ballot, the voting machine puts it in the box. If it's the wrong ballot or the scanners don't all agree the voting machine spits the ballot back out and indicates the ballot has an error. Each scanner records what it read and whether the ballot was accepted or rejected.
At the end of the night several parties will have the same raw data and can demand a hand recount if the official results don't match what they recorded.
None of the machines would be allowed to have any wireless capabilities and wouldn't be connected to any network until after the polls close. More could be done, like scan a label for each box as they're changed so everyone has a record of exactly how many boxes there are and how many ballots are in each box. The voting machine could also mark an edge of each ballot it accepts, the machine would be physically incapable of marking anywhere people mark to vote, and the ballot would be rejected if someone tries to run it through a second time.
That's not crucial and is there to stop any party from knowing the results in advance. Exit polls and counting how many people registered to which party have already voted already gives a pretty good idea of that already. If it is necessary, on election day morning, before the polls open, some of the scanners could be randomly replaced and set aside to be opened up and their circuit board inspected. Have mandatory prison time that's enforced for the head of whichever organization is found to have equipment that can transmit wirelessly.
That's a good point. I guess when you put your ballot into the machine and it accepts it, you know your vote was counted. None of this nonsense in Georgia where thousands ballots were found to not have been counted and no one can figure out why.
I disagree. There are some really good ways to use machines and some really bad ones (e.g. Dominion).
For example, we could have paper votes counted by machines from both parties. If the vote result doesn't match, a team of inspectors can verify the suspect batch. If they quickly process verified batches like this it will be much faster and more accurate than human counting.
These problems are easily solved. The team of inspectors will be made from members of both parties of course, overseeing the two machines. The machines can be sitting right next to each other. This totally removes the incentive to rig a machine as it will be obvious from the first batch counted.
And should be counted by active military/national guard.
I would argue that software based voting is very much viable, if its designed to be fully audit-able using cryptographic signatures, and fully verifiable and fully transparent. With features like duplicate ballot detection, full admin log, and full operator log, connection to upload servers only via a secure VPN, admin access only from console or a secure terminal, all data transfers from the voting machines one-way upstream, and all data transfer to analytics and external endpoints one-way downstream.
The software provider must be fully liable for security and compromise of the servers.
It wont be too hard to create a fool proof system as long as that is the actual goal of the software. Dominion was created with the express purpose of being able to cheat on voting without getting caught. But luckily they failed in that as well.
Until someone finds a way to remote hack paper and ink I'd stick to it. Security aside it keeps money out of the hands of parasitic busybodies looking for easy government contracts.
I would prefer to use public hence auditable system with cryptographic signatures for tallying data from the polling stations up. But actual voting to be still done in person, with photo id and good old pen and paper.
Afteral,, that's what the most well known auditable and "everything cryptograhical signed" "verifiable online elections" system Helios recommends for general elections.