Comment by kristopolous | Hacker News Reader

Comment by kristopolous | Hacker News Reader

so the real question is where is the startup with the solution of $35 tablet "voting machines" on $10 stands and a $30 SBC "voting server" in a $15 lockbox that can deliver 4 "deluxe voting terminals" with a "server" for $2,000? (89% profit)

The references of the current players in the industry I see talk upwards of $3,000 per terminal! The industry standard pricing for this would be $12,000. $12,000 for buggy, insecure, closed-source, non-updateable, unverifiable, crappity crap. Awesomes.

How about this ... a voter walks into the polling station and can either use a terminal or their own smart-phone - connect to the open wireless "voting hotspot" where they can enter in an anonymous token at something that we use as the "hot spot sign-in page" that they are given by the people who volunteer the polling station that they can use to cast their ballot without waiting in line.

It's not like secure voting is an open problem. There just needs to be some crotchety capitalist company with people in suits selling it. Open source, anonymous, verifiable voting machines on open hardware with an optional paper trail (cups + network printer). We can do this making comfortable profits and at atleast a 60% savings to the tax payer.

The best part is that if the company goes out of business, everything is open source and documented, so we don't put communities at risk of the security issues with code rot.

So ... who wants to make some money? email me: kristopolous (at) gmail

How about this: A voter takes a paper ballot. The voter marks the ballot with a felt pen a deposits it in a bin. Polling place staff mark the voter's thumb with indelible ink.

Electronic voting is a problem looking for a problem. You need voter verifiable ballots with a high degree of tamper resistance. As far as trustworthiness goes it's hard to complete with ballot boxes and adversarial observers.

I don't see how your proposal can prevent double voting. Voting happens over a period of several days not to mention absentee voting.

Plus, it seems to me the process is not particularly tamper resistant during the ballot tallying phase, when compared with what is possible using cryptographic techniques.

How would one go about tampering when, as is traditional, representatives from all parties are in the room when the votes are counted?

Paper voting works in Australia. You don't even need your ID (edit: and by "don't need", I mean "they aren't allowed to ask for it") - you rock up at a polling station in your electorate, give your name, they mark it off on a sheet. You give verbal confirmation that you haven't already voted. They give you a voting paper, you fill it in at a booth, then drop it in the box on the way out.

(If you're so inclined, you can also just get your name marked off and walk out if you want to donkey vote, which is the equivalent to a protest/blank vote. Or draw penises on the voting paper).

If you don't vote, you get a nice letter in a few months from the Australian Electoral Commission with a fine of about $100 or so for not voting. Voting is also on a Saturday so most have the day off, and there are early polling stations open for a while prior to the day so if you can't make it on the day you can do it then.

There is also online voting. At least in the most recent NSW state election there was. I was overseas during the election and attempted to vote via this method. Unfortunately my hotel's WIFI kept timing out and I gave up in frustration. I returned to Australia and there was a $57 fine waiting for me for failing to vote.

I think our voting system is good but it has flaws. I recall reading an article a few years ago about somebody who showed up to vote (i.e. Got his name marked off) at multiple polling stations on election day just because he could. He received a fine for doing it but all of his votes were counted.

Edit - googled it apparently you can be charged with fraud for multiple voting or for voting under a false name (up to six months prison - so slightly more than a slap on the wrist and a fine).

Hmm I wasn't aware that anywhere in Aus was doing online voting! My only experiences have been either Federal elections or Victoria State Elections, which both follow pretty much the same process.

Yes Australia has quite a good system - we just need to get preferential voting above the line for the upper houses. At the last state election I had to number all the candidates from 1 to 410 if I didn't want to give my vote to the faceless back room political operators.

I should mention that a donkey vote is not a equivalent to a blank vote. A donkey vote is when you number the candidates from 1 to x from top to bottom. You actually give your vote to whoever happens to be at the top.

Qld state election required Id. Each voter was sent a registration slip and they had to provide it at the ballot box.

The loss of the senate ballots and subsequent post-election re-run of the entire state - which had a different outcome - was a massive failure of Australian electoral process.

I appreciate the consumer advocacy, but if the machines are going to exist (and they are), they should be done well.

The attack on paper voting is not flipping votes, it is invalidating votes. In Chicago, election officers counting the votes used to hide a small piece of pencil lead under their finger nail. If there were too many Republic votes, they would start putting an extra mark on some of them, thus invalidating the vote because of stray marks.

I think the solution is to crowd-source verifying the vote. If every ballot were given a unique number or code, and when the ballot was counted each ballot with who it was counted for as available for download on the web, third parties could verify that the numbers were counted correctly. And each voter now has an anonymous number that he can use to look at how the vote was counted and verify that it was counted correctly. If it wasn't, he complains. The political parties would be happy to urge their loyal constituents to check that their vote was correctly counted and to sue should they have a sufficient number of members claim theirs was wrong.

Wouldn't this allow a bad actor to coerce a voter into disclosing their anonymous number and revealing how they voted?

That is the weakness.

It wouldn't just be that their vote would be revealed, though, it would be that then the bad actor could contest the vote, claiming that they were the voter and their vote was counted wrong.

But if people were in the habit of destroying their unique code immediately after the results are posted, then they at least have plausible deniability. "No, I burned that piece of paper yesterday. Can't remember what the numbers were."

If you can trace a vote to a person, through force or otherwise, then you can coerce said person to vote however you like.

"Vote for this guy or you're fired" type stuff

You are right. I failed to thoroughly think it through.

But I think I would rather have a system where you need to intimidate and coerce people en masse in order to change the vote than a system where you just need to coerce the handful of people who are responsible for counting the vote.

I think the best way would be to augment paper ballots with an on site machine vision system that permits votes to be reviewed by the voter and then prints a matrix code on the ballot for future auditing.

In the UK the count is done under the supervision of:

* The candidates themselves

* Journalists

* TV cameras

* Party representatives

* Electoral officials

* Anybody else who has a good reason to be there

The volunteers doing the counting are not going to risk a long stretch in prison with such a high chance of being caught.

Exactly that. The argument you get against this is that you could say sell your vote and verify to a 3rd party you voted as asked. But frankly I think that is not nearly as bad as hidden vote tampering. Not even close.

Worse than selling votes is coercing votes. Being fired because you didn't vote the way the boss likes? Terrible. Verifying votes this way turns a secret ballot into an open ballot, with all the attendant problems.

Hidden vote tampering of the pencil-under-the-fingernail variety is a solved problem here in Australia. Each vote count is done by an employee of the electoral commission. They're watched by volunteer scrutineers (plural), provided by the major parties, with each scrutineer looking to maximise their own party's response and distrusting all other parties. You effectively have two opposing meatspace inspectors willing to challenge anything that looks shady.

Paper voting isn't perfect, but it's far, far better than any electronic voting system that the general public can use.

EDIT: to clarify, with any electronic medium, the only way to confirm that your vote was counted as cast is to cryptographically sign it in some way with a key that's under your control. Even us techies have problems doing that, and the general public could never be expected to do it, especially the socially disadvantaged members. Not even security conference attendees keep up to speed with their cryptographic signing, so how can we expect mere mortals to do it?

With paper votes, you confirm that your vote goes into the box as you've marked it. You can then opt to be a scrutineer and watch that box travel to the counting area, and confirm the votes are counted correctly. And has the benefit of having the record independently verified if required.

Any form of voting which has a "just trust us on this" stage is not good enough.

Modern paper voting involves a thick black markers, OCR computers, and election monitors from both parties. The 1880s are long ago history.

Not in the UK it doesn't. All pencil and human counting here.

Not in Sweden. Here it involves paper, pencils and manual counting. Nothing which could not have been done in the 1880s and we still get a preliminary result a few hours after the elections close.

All paper and pencil in The Netherlands after some techies realized black box voting is a dangerous thing. http://wijvertrouwenstemcomputersniet.nl/English

Although some people really want to get the electronics back so it's an ongoing battle of educating people (even most hackers and others in the security industry that did not think a lot about voting in particular).

In NZ we use GIANT orange felt pens to mark our ballots, good luck hiding one anywhere.

I've heard that pencils are sometimes used to prevent an attacker from substituting pens at the polling booth, with modified pens containing disappearing ink, but that seems like a problem better controlled by using new, easily identifiable pens each election.

As for serialising the ballot-sheets, that enables a third party to check that a person they coerced into voting a certain way actually did as instructed, which undermines the value of a secret ballot.

In the couple jurisdictions where I've done most of my voting, the ballot box for the paper ballots has been an electronic tabulator.

I don't see much downside to this sort of system, there is a paper ballot for any checks that are needed or desired, and the vote tally is readily available when the election ends.

Cost of customer acqusition. States have to be bagged one-by-one, each with an insanely long sales cycle, your profits have to cover those nice salespeople's salaries, while leaving you with sufficient profit to cashflow expansion. This is not a dev problem; the sort of orgs which are prepared, and can cashflow these sort of sales simply don't have (nor are they incentivized to have) any dev culture whatsoever.

first, that margins are pretty standard on accounting for sales and overhead

but even if they aren't enough here then you sell the machines to those people to flip them to the states at a higher cost. Don't be in THAT sales game. Otherwise, sure, you can add 30% to the cost ... it's not going to be another 600% just to sell the thing.

And even if it was, I'm sure I can personally, in my spare time, produce a far better product than the one outlined in the article.

>where they can enter in an anonymous token that they are given by the people who work the station

That doesn't seem very anonymous. In fact, it seems like a possible way of deanonymizing votes, since in many places you have to sign a ledger or provide identification before voting.

you don't have to couple the two.

You don't have to, but unless it's illegal, they'll tend to be coupled just for the sake of convenience.

Well, it could happen like this:

- Walk in, present some form of ID

- You sign the slip or give a thumbprint so they record that you've voted

- Random token is generated and given to you

- You enter the token in order to vote

- You keep the token as one way to verify your vote was cast

It'd be hard to match the voter to the token if they don't write down the token and associate it with your name.

That might work, as long as no one (including the voting machine) was recording the time you arrived or the time you voted.

Well, couldn't _only_ the voting machine record the time the token was used as long as the token is random? (They could just generate N tokens -- one for each registered voter -- and pull one from the pool each time.)

If they recorded the time you took your token and the time the token was used then they'd have a small enough pool of people to work with. I know in my state if a precinct has a small enough pool of registered voters it doesn't show the vote totals from the precinct because of privacy issues.

Storing the timestamps is a design decision which can be intentionally omitted.

There is also no law of computing which states that the ordinal of the voting record is determined by the chronology of the votes. That is to say that you don't have to keep things ordered by time - or at all; it's not like an 8AM vote should be treated different than a 5PM one.

Please do better than these folks who just won a bid for elections in Los Angeles, and previously held elections in South Dakota, New Jersey, Chicago, IL, Cook County, IL, Denver, CO, Oregon, New South Wales, El Paso, CO, Utah, West Virginia, Honolulu, HI, Franklin County, WA, Swindon Borough, UK: http://www.everyonecounts.com

  so the real question is where is the startup with the
  solution of $35 tablet "voting machines" on $10 stands and 
  a $30 SBC "voting server" in a $15 lockbox that can deliver 
  4 "deluxe voting terminals" with a "server" for $2,000?

They appear to be using an iPad with their "SaaS voting app": http://www.everyonecounts.com/voting-unit/

I don't understand this race to the iPad ... the LAUSD (Los Angeles School District) had this bizarre iPad for every student program (http://www.scpr.org/news/2015/04/24/51241/lausd-developing-n...).

I say it's bizarre because I know you can get tablets in bulk for $30-$35 each while the lowest price iPad is $400.

After getting past "why does every student need a tablet computer" we get to "why from a single vendor with prices that are 1,333% higher than the cheapest competition?"

I'm not saying the Apple product isn't nice - it's just I don't see why so many tax dollars should be spent when there's other options.

Going back to the voting machine ... we'd not really care about some wonderful durability of the cheap tablet because the owner would only be using it a few days a year at the most; in ideal conditions, without moving it; the thing will survive fine.

> After getting past "why does every student need a tablet computer"

Speak for yourself. I'd rather people didn't get past that quite so easily.

the idea is to give less affluent students the same access to the "digital divide" as the more affluent ones. I think it's a valid concern ... but this is not a valid solution to the problem.

There's a school here in Dublin where the kids were starting to get back trouble from the weight of all the books they had to lug around, so they gave them all iPads and books in digital format instead, which solved the problem; that's certainly a legitimate need.

The cost of a single paper textbook averages anywhere from $150-$400.

Will a $30 voting machine be accessible by those with vision/motor disabilities?

Nope. You can have a special machine for them - just like you have a special porta-potty --- no need to increase the cost of every unit for that.

An accessible system would not increase costs. In fact, it could be more expensive because you'd have to spend a good amount of time testing both the regular voting system, and the accessible voting system. If your voting system requires a separate machine for simple disabilities like vision impairment, you've already given more of a potential of discounting an already marginalized portion of the population.

Thankfully, the guidelines to ensure things are accessible are fairly easy to find, and easy to implement as long as you keep accessibility as a requirement, and not a suggested feature. For example, the National Federation of the Blind has a list of features that cover the spectrum of visual impairment on their website for free [1]. Groups for people with other disabilities, like motor impairment, also have guidelines which aren't that difficult to follow. Smart phones with touch screens and the like are used by people with visual and motor disabilities too, so this isn't uncharted territory, it's just a matter of asking.

[1] https://nfb.org/purchasing-voting-machine

Reply on news.ycombinator.com