All-Digital Voting

[Z]

http://www.kitv.com/politics/19573770/detail.html?treets=hon&tml=hon_9am&ts=T&tmi=hon_9am_1_02000105272009

Voting Drops 83 Percent In All-Digital Election
People Could Vote Online, On Phone For Neighborhood Board

HONOLULU -- Officials saw an 83 percent drop in the number of voters participating in the Honolulu Neighborhood Board's recent election that is the nation's first all-digital election, where people could vote over the Internet or by phone.

For the first time, Oahu voters had to use computers or the telephone to vote for their neighborhood board candidates and many people did not bother.

About 7,300 people voted this year, compared to 44,000 people who voted in the last neighborhood board race in 2007.

"That is of great concern to me. It is disappointing, compared to two years ago," said Joan Manke of the city Neighborhood Commission.

Manke heads the commission that oversaw Internet voting. She said voters obviously did not know about or did not embrace the change to high-tech voting.

"This is the first time there is no paper ballot to speak of. So again, this is a huge change and I know that, and given the budget, this is a best that we could do," Manke said.

The city cut its expenses in half by using computers and phone technology by Everyone Counts. It cost about $95,000.

The question is whether the state and the counties will use the new voting technology to reach out to overseas voters, people who are not able to walk into voting booths like these to vote.

"This is the future for presidential elections, general elections, primary elections, all the way," Everyone Counts consultant Bob Watada said.

Watada is the former Campaign Spending Commission director.

"(It) gives access to a lot of people who haven't had the access, and you don't have the hanging chads, you don't have the miscounted absentee ballots, you don't have the ballots lost," he said.

More public education is needed to assure voters, who are used to going into voting booths that online voting is safe and easy, Watada said. However, that costs money, which is difficult to get in these tight budget times.How do you feel about all-digital voting? Would you prefer it over paper ballots? Why or why not?

I haven't formed a solid opinion on it yet. On the one hand, I trust the integrity of paper ballots a bit more, and everyone can cast a paper ballot (whereas some people don't have the internet or phone access to do it digitally).

On the other hand, the internet and phone voting would be a lot more convenient, but I would it has a greater possibility of being corrupted. I also don't know if the older generations would be able to figure it out or bother trying (as seen in the above story).

[Gio Takahashi]

i think it's a good start, and should be an optional alternative, not a mandatory option. (Not everyone is going to have access to computer and internet, nor will they want to do it online.) I do feel that it has a long way to go, there's still a lot of trouble with using digital voting. (corruption, hacking, etc)

[Arainach]

There needs to be a paper trail. Period. Voting systems NEED to be recountable; this is not an optional thing, it is a core reqirement of the system. Also, internet/phone can't work because another requirement of a voting system is that it has to be anonymous - that is, there can't be any receipts or way to tell who someone voted for to prevent buying/selling votes.

[Killer_Man_]

I have to agree with Ary, paper trail only please.

[deathofcheese]

Couldn't there be a way to digitally register votes while registering who already has voted without identifying who those votes came from? Couldn't one system be responsible for checking someone off a list of people who are eligible to vote while allowing them (after checking them off the list) access to another system that merely tallies their vote?

I like the idea of digital voting, but definitely only as an option. We're not anywhere near the point where everyone can vote remotely or digitally, so nowhere should plans be made to replace any preexisting voting system entirely with internet or digital voting. However, for those of us that are technologically-minded, it's nice to have the option to make it more convenient. In the case of people "forgetting" to vote, a simple census-linked email system which spams GO VOTE messages in the week before the election would cover all reasonable efforts to remind people to vote.

[Arainach]

Couldn't there be a way to digitally register votes while registering who already has voted without identifying who those votes came from? Couldn't one system be responsible for checking someone off a list of people who are eligible to vote while allowing them (after checking them off the list) access to another system that merely tallies their vote?There needs to be some sort of authentication token to the second system. If it's unique, they can trace back who voted. If it's not unique, people can vote multiple times.

[Seegtease]

Bad idea right now. The main concern is that, as this example shows, a drastic decrease in voter turnout. Many people just don't get it, or refuse to change. Once computers are more engraved in our lifestyles (I know they are now, but not enough for this), then a switch may be feasible.

[Killer_Man_]

Well perhaps they'll make you sign your vote.(I know for a fact that there are sign pads for laptops/computers.). They could just ask you to finger print or sign off on your vote.

But i'd still like paper votes.

[Arainach]

Well perhaps they'll make you sign your vote.(I know for a fact that there are sign pads for laptops/computers.). They could just ask you to finger print or sign off on your vote.

But i'd still like paper votes.That can't work. Votes MUST not be tracable back to individual people, or democracy fails. Retaliation against people who didn't vote for the winner, elections bought by the person with the most money/power.......it's not a pretty picture.

[Bloodcinder]

Consider the electronic voting as a zero-knowledge proof (zkp). Build three machines: Kiosk, Register, and Tally. Allow Register and Tally to communicate with Kiosk over a secure channel. Get a voter named Ann to participate and a staffer named Bob to ensure proper conduct at the polling place.

1. Ann goes to Kiosk and attempts to vote using some credentials authenticated by Bob.

2. Register validates to Kiosk that Ann is allowed to vote. If not, stop here.

3. Ann fills out her ballot.

4. Kiosk uses a zkp to prove to Register that Ann voted.

5. Kiosk uses a zkp to prove to Tally that a specific ballot was cast but not by whom.

6. Tally confirms receipt of the ballot from Kiosk.

7. Kiosk informs Ann that her vote has been placed.

8. Bob ushers Ann out of the polling place.

The implementations of the zkp's and encrypted credentials are not trivial, but I don't see any (irreparable) problems with this model.

---------- Post added at 20:37 ---------- Previous post was at 20:36 ----------

Bad idea right now. The main concern is that, as this example shows, a drastic decrease in voter turnout. Many people just don't get it, or refuse to change. Once computers are more engraved in our lifestyles (I know they are now, but not enough for this), then a switch may be feasible.
Not really a valid point. When people go to the polling place, they use what's at the polling place. They may have the option to do it from home somehow, but people who are stuck in their ways will just go to the polling place. (EDIT: My point is that all-digital is not the problem: it's forcing people to do it from their homes that is.)

[Arainach]

The only way to trust such a model is with a completely open model, however. It's amazing the levels of transparency we demand from gambling machines (they can be and are inspected at any time without advance notice down to the circuit level, and all source code, etc. is analyzed by unbiased boards) and yet when it comes to voting machines, arguably the most important computers of all, we allow the companies to sue anyone who even attempts to test them. Not to mention that all current electronic voting machines suck miserably for a wide variety of well-documented reasons.

[Bloodcinder]

By definition zkp's don't work unless the details are open.

It makes no ethical sense, the situation you stated: that private companies produce the machines but aren't generally allowed to be tested. That's definitely something that's wrong with the current system, and, yes, something that would have to be fixed prior to implementation of a model like mine.

[deathofcheese]

There needs to be a paper trail. Period. Voting systems NEED to be recountable; this is not an optional thing, it is a core reqirement of the system. Also, internet/phone can't work because another requirement of a voting system is that it has to be anonymous - that is, there can't be any receipts or way to tell who someone voted for to prevent buying/selling votes.

There needs to be some sort of authentication token to the second system. If it's unique, they can trace back who voted. If it's not unique, people can vote multiple times.

That can't work. Votes MUST not be tracable back to individual people, or democracy fails. Retaliation against people who didn't vote for the winner, elections bought by the person with the most money/power.......it's not a pretty picture.There seems to be a paradox in here (votes must not be traceable back to the people! votes must be traceable back to the people!) but I think I've had too many beers to find it for realz.

[Gio Takahashi]

The biggest problem with digitally voting is that it would be easy to identify the person who's doing the voting. That and possibly forgery, and multiple voting. I like the idea of digitally voting, but like I said, we're definitely not ready for it yet. It's just asking for too much trouble and mess that could be even worse than year 2000 election.

[Bloodcinder]

There seems to be a paradox in here (votes must not be traceable back to the people! votes must be traceable back to the people!) but I think I've had too many beers to find it for realz.
Ary's "paper trail" refers to completed ballots and registered participation... but not ballots that record participation. That is, the paper trail has to show that somebody voted and has to show that a certain vote was cast but not the combination.

This "paper trail" is preserved in my zkp model.

The biggest problem with digitally voting is that it would be easy to identify the person who's doing the voting.
Only if you use naive methods, like in cookie-based internet polls.

That and possibly forgery, and multiple voting.
This is not unique to digital voting.

[Seegtease]

My point is that all-digital is not the problem: it's forcing people to do it from their homes that is.)

I thought that was the discussion.

[deathofcheese]

Like the reason there isn't a virus threat to Macs or Linux boxes, paper voting is so "secure" because we've been using that method for so many years we've had time to come up with ways to make it "secure". Since widespread digital voting has only become a feasible option in recent years, it makes sense that it's "not as secure" as paper voting.

[Arainach]

Which would make total sense if there were a lot of attacks against paper voting and none against electronic voting, except that the exact opposite is the case.

---------- Post added at 23:48 ---------- Previous post was at 23:41 ----------

There are several requirements for a voting system:
Only those who are eligible should be able to vote No one should be able to vote more than once Voters should be able to tell that their vote was properly recorded The votes should be able to be recounted No one should be able to tell who any individual voted for

Paper systems handle all of this nicely. Simple human people at the door ensure that everyone who gets in is eligible and that they are not able to get in more than once. A person either fills out their paper ballot or uses a machine that shows them a receipt, so they can know a record of their correct vote exists. Any such receipts are not available to the voter, so no one can tell who someone voted for. And the paper trail allows for recounts.

Now, if someone other than Diebold could write a halfway-decent voting machine that included decent tamper resistance and wasn't hacked together by a 6 year old, electronic voting machines at physical voting locations would generally work just fine, so long as they also printed out a paper receipt (visible to the voter) to allow for recounts. Anything on read/write digital media is susceptible to tampering and not sufficient for the "recount" provision.

The real pickle, as BC has alluded to, is trying to allow voting from home. That's a pain in the ass. Strong Crypto protocols DO exist, but they're tough to get right and probably more trouble than they're worth.

[Bloodcinder]

I thought [the voting-at-home issue and not digital voting itself] was the discussion.
The topic is named "all-digital voting," and Ary's concerns were with the paper trail.

The real pickle, as BC has alluded to, is trying to allow voting from home. That's a pain in the ass. Strong Crypto protocols DO exist, but they're tough to get right and probably more trouble than they're worth.
I don't think there's a technological impediment to voting at home. Even the military couldn't break 256-bit encryption within weeks of election day. By then the votes have already been counted, and even if the encryption were broken there'd still be no record of who voted what in my model since that information was never stored in any form.

You'd just have to ensure that Kiosk queues up messages to Tally and sends them in a random order over time to ensure that the order of the records in Register and Tally aren't necessarily the same.

[Charlie]

Paper or bust. I hate the whole idea of digital voting. Unfortunately, it seems to be the way of the future, though...