Virginia has decertified its most hackable voting machines
techcrunch.com
techcrunch.com
There are multiple local party volunteers (I was one in the last election) from all the major parties at all stages of the process watching the flow of paper ballots from the sealed black collection bins to the voting halls where local council officals that volunteered, with paltry compensation, to a 12-18 hour shift to collect and count the vote. Some local councils take pride in being amoung the first to report their results, and there is a side competition to see who reports first.
We usually know the broad outcome of the election by the early hours of the next morning, with a handful of recounts in tight races going into the next couple of days.
In the Japanese writing system, candidate's name can be written in multiple ways. If writing doesn't match one of the exact registered candidates names, the human vote counter decides if it refer to the unambiguous candidate. If such candidate exists, it is counted to the voting for that candidate.
> There are multiple local party volunteers ...
Same.
Japan's literacy rate is very high, but it's not 100%.
That it is archaic, inefficient and a waste of human time?
We don't hold elections that often, nor do we need their results instantly. Voting with a pencil takes about as long as voting on a touchscreen. So why is there this stupid urge to optimise?
Would that more people in the US understood the difference.
And even if you fully count both records, what happens when there is a discrepancy?
Computer and hand count does its own thing, and in the end the numbers should agree.
4% of the time it prints a different paper receipt, it just makes the electronic numbers go in someone's favor by a few percent in a couple choice districts. Or it uses crafty UX to make people vote for people or things they normally wouldn't.
And what if the numbers don't agree? Are you just going to pick one side? Are you going to hold a re-election? Will the re-election have the same turnout? Will people not go because their side won by a landslide before? Will some people change their mind because the media is reporting that side "X" is behind the discrepancy? Or will it just give side "Y" the few extra weeks they needed to campaign and win?
Cast the ballot on paper. Every ballot has unique counterfeit patterns (totally randomized). The ballot is scanned and send to server, with both the scan of the ballot and the eletronic reading. Voter should be asked to verify (but we know half don't do it). The paper ballot is casted the same way as it has always been. We can also read the paper ballot one more time if we have to during the first round of counting. You can't have a fraud unless you cab mainpluate both paper and eletronic. Either both come to same number or they don't. I am just not familiar with how we verify eligibility tbh.
When we count paper ballot, we scan for the counterfeit. If the final number has mismatched, we either have a miscount or an error. Paper ballot will alwsys get recount when there is dispute.
Have we had a re-election in the entire American history, or at least in the past 30 years? Regardless, paper ballot fraud can lead to reelection so regardless of what proposal we will face re-election if we have to, but we normally don't.
This seems like over engineering, but we have to take into account people living abroad almost always now rely on eletronic voting if EV is offered.
I am a novice on the whole voting system, but just off the top of head my proposal would seem to work in ideal situation. The fact we have a record of ubique vote electronically and on paper is cruical to the integrity of each vote. There must have been a few dozen studies already on this topic. I'd suprise if no one has a bullet proof system yet. It CAN'T be that hard. Just emulate cryptography here.
You don't need to mail anything anywhere in a paper vote, count by hand system.
> having a few people watching a locked box also has problem. The whole counting by hand also has problem.
What problem do they have?
> I'd suprise if no one has a bullet proof system yet.
We do have one: no electronics, count by hand in public view.
And requiring a large amount of manpower is a benefit of paper voting, not a problem. In the grand scheme of things it's a miniscule amount of work, and it ensures freedoms and laws for generations to come. This isn't an area where efficiency matters. The more manpower it takes, the harder it will be to "hack" it without being noticed.
How do they know that you actually cast the vote for you? Say my 5 best friends all tell me, "Nah, I'm not gonna vote. My vote doesn't count anyway." I then go to each of their districts, claim to be them by giving their name and address, and vote how I want. What's to stop that from happening?
Also you need to consider:
1) Those kinds of people are probably not registered to vote. Voter registration is not automatic in the UK. It takes 5 minutes with an online form, but some people are just that disengaged.
2) The 'wards' in which you vote are often quite small and in close proximity, so it might be a little hard to figure out which polling place your friends are registered at if you haven't received their polling notification card. You can get this information, but its a small barrier.
3) The 'wards' have an average population of a few thousand, so there is a risk that an official might recognise you as trying to double vote, but I can't imagine how they might confront if they aren't 100% personally certain.
None of those are strong guarantees. If evidence of wide spread fraud were to come to light, party policy ID requirements might change, but to date no party is seriously considering introducing them. Past governments have repeatedly tried to introduce a universal ID that could be used in part for voting ID, but each attempt has ended in failure due to public backlash over civil liberties concerns.
I think we should return back to paper ballots. When there are witnesses to elections, I don't think any of them are qualified to judge if an election has been done in a honest fashion. This would require experts on the voting equipment where they can guarantee that it has not been tampered with and that is too high a burden and something that complex can not be trusted.
I don't think you should leave without (at least seeing) proof that your votes were properly and accurately recorded.
If that means that you have to be given a "receipt" with the names of those you voted for on it, well, so be it. It isn't like your name and/or any other personally identifiable information would be on it -- just the minimal details needed to achieve the singlemost important purpose: verification.
The proof would be a paper ballot that you turn in. Your actual vote.
A printed receipt, from an electronic polling place could just as easily be placed in a secure box prior to leaving.
It's amusing watching "ballot selfies" issue too -- it seems that folks just don't understand how important it is not allowing any coercive force influencing or verifying your vote.
Now imagine cults and other groups that pressure their members into producing those receipts with the correct candidate on them... or else.
One way it's done is that after entering your choices on touchscreen, you see the printed receipt through a glass window, and then it's stored so a recount can be done if necessary.
However, an electrically scanned paper ballot seems a lot simpler.
I actually believe this is actually the best possible situation.
The voting machine should print a clear, unambiguous, ballot and on-screen tell you to verify it before you officially "cast" your vote. "If the ballot below does not represent your choices, please click <HERE> to request an attendant."
I'm thinking of this more as having printers that print out an unambiguous completed ballot and less as voting machines that "also print out a copy".
Many other countries use paper ballots, including highly developed ones (e.g., France, UK...)
I do think we need paper -- a "hard copy" -- at some level, whether for the ballot itself or just a verifiable audit trail. I think 100% electronic voting would be fine so long as there's an audit trail that matches the (electronically) talkied results perfectly. I don't think dealing with millions of paper ballots would be any more advantageous than that -- you'd have to distribute and collect them (although that would happen at the centralized polling places), a few people would have to hand tally them (to ensure everyone's numbers match up), and then securely store them (via very strict "physical security" procedures/protocols) until, at some later point, they can be properly destroyed.
Logistically, electronic voting with a paper audit trail would be much easier, would require fewer people (i.e., election workers -- who are already in short supply), would give us results much quicker, and would be easier to verify/authenticate later (in the event of disputes).
But... why? What is the point of adding all this complexity? What problem is electronic voting solving?
Putting a piece of paper in a box is something everyone understands how to do, even elderly people who have never touched a computer -- how is electronic voting easier?
> would require fewer people (i.e., election workers -- who are already in short supply)
They are? I can't believe it would be more expensive to hire a few more election workers for one day than we're paying for these machines.
> would give us results much quicker,
France uses paper ballots and results are known within hours.
> and would be easier to verify/authenticate later (in the event of disputes).
This is not realistic. We are never going to go back and change an election result whether or not records of the votes were saved. Cf. Florida 2000.
Part of the witness duties should be making sure the votes for that polling location are counted accurately and there is no way to do that with electric voting systems. It's just a black box you have to trust.
Here's a better system: you go into the voting booth. They have some envelopes and stacks of preprinted ballots, each with one candidate's name. You take a ballot for the candidate you want, and put it in the envelope. Then you come out of the booth and drop the sealed envelope in a box.
After polls close, they open the box and all the envelopes, and count the ballots, in the open. Anyone can go watch. If an envelope is empty or has more than one ballot inside, the vote doesn't count.
Works well for elections with one race; those with a dozen offices being elected and an equal number of ballot measures make that... impractical.
In extreme cases you get huge ballots with this system, see e.g. (http://www.faz.net/aktuell/rhein-main/frankfurter-wahlzettel...) for Frankfurt with a ~0.9 square meter ballot.
Counting by hand still works even for those over-complicated ballots.
With 5 offices with 6 candidates each and 10 yes/no measures (a fairly modest ballot for a California election) this would require each voter be handed a packet of nearly eight million ballots to choose from.
This is clearly not practical.
One thing I will say is that the US often has a lot more offices on the ballot than some other counties. In Washington, DC, a ballot can include Advisory Neighborhood Commissioner (somebody representing an area of a maybe twenty blocks), city council member, school board member, mayor, non-voting delegate to congress, and depending on the year "shadow senator" and president. In other places one may have county commissioners, judges, sheriffs, etc. That does make for a lot of tabulation.
Voting needs plausible deniability so you can't be coerced or bribed into voting for a particular candidate as was alleged to happen in the Balkans many times: https://en.wikipedia.org/wiki/Bulgarian_train
The network can be run by the public. Ofcourse there is risk associated with the 51% attack, but it's still a more transparent process than what we have today.
Secret ballot is a feature, not a bug. It protects you from coercion and retaliation.
Secret paper ballots really deserve more credit. The tech might be old but it gets the essential features right.