L.A. County gets state approval of new open-source vote-counting system
latimes.com
latimes.com
> The ballot-counting equipment is part of a broader redesign of Los Angeles County’s voting system, which will include new equipment while relying on a traditional paper ballot
Article is light on details, and it's difficult to tell exactly what this sentence means, but if it means the new software will produce paper ballots that can be verified and observed by humans, that's a good thing at least.
The machine and ballots are tied together right there. I really like the system as if all else fails you can get a paper trail and I belive even ID machines.
Also, they could have granted long term contracts to existing suppliers, don’t have a budget to replace existing machines, don’t want to make the person who bought them look bad, or even possibly don’t want to say the old machines are bad because it calls into question the integrity of past elections.
And even then, you need to trust the machine. In many cases, these machines are using terribly old processors (sometimes even using the venerable 68000), and building a corrupted clone could probably be something doable.
The corrupted processor may have a different behavior, depending on the situation, such as the exact time of operation (ie. it would not be detected by random tests before or after an actual vote), or the analysis of people voting (ballot might be modified only during peak hours, to limit detection)
I'm baffled to see that democracies still engage in electronic voting systems.
At the end of the day, it even goes further than that -- you have to trust the people who build the machines.
Ken Thompson said it best in his Turing Award talk: https://www.archive.ece.cmu.edu/~ganger/712.fall02/papers/p7...
Because it's open sourced, we can be virtually certain that it's not built into the software, and if it is you need to prove it. Which would fix it, or obsolete it.
Someone will have to ask the security detail handling custody, or one of the other thousands of people who could potentially be involved in trying to rig an election.
It's finally cheap enough to have the transformation of it, from raw compounds into an engineered SoC for tallying votes, livestreamed 24/7 and operated by openly auditable robots.
Convince me that would cost more than one fighter jet.
https://abc7.com/politics/new-voting-system-approved-by-la-c...
That seems to be the case:
>"Separately, Logan signed a contract on June 13 with Smartmatic USA, making it the VSAP prime contractor and systems integrator. Smartmatic USA will help Logan's office in managing the manufacture and implemention of components scheduled for introduction in the March 2020 California presidential primary election." https://www.techwire.net/news/la-county-oks-open-source-elec...
Why are all these "voting machine" companies so shady?
>"The Venezuelan-owned Smartmatic Corporation is a riddle both in ownership and operation, complicated by the fact that its machines have overseen several landslide (and contested) victories by President Hugo Chavez and his supporters. The electronic voting company went from a small technology startup to a market player in just a few years, catapulted by its participation in the August 2004 recall referendum. Smartmatic has claimed to be of U.S. origin, but its true owners -- probably elite Venezuelans of several political strains -- remain hidden behind a web of holding companies in the Netherlands and Barbados. The Smartmatic machines used in Venezuela are widely suspected of, though never proven conclusively to be, susceptible to fraud. The company is thought to be backing out of Venezuelan electoral events, focusing now on other parts of world, including the United States via its subsidiary, Sequoia." https://wikileaks.org/plusd/cables/06CARACAS2063_a.html
The ballot-counting
equipment is part of a
broader redesign of Los
Angeles County’s voting
system, which will include
new equipment while relying
on a traditional paper ballot.
They have been asking for suggested areas for "Voting Centers" recently where LA County citizens can vote at the location of their choosing. Many people here have long commutes from traffic or public transport. For this reason (and maybe others) the County wants to make it so people can vote at any of the available "vote centers", rather than one assigned polling location by their home.
You can watch the video that they were sending around here: https://www.youtube.com/watch?v=fC_-8Nl-O3U
Also, you get to examine the actual ballot before casting your vote, even if you don’t fill it out with a pencil.
For example, in a 2017 election for a seat of the Virginia House which would have switched party control, an election recount came up one vote in favor of the Democrat, but then GOP officials declared that a (clearly spoiled under any reasonable interpretation of the official rules for counting) ballot should count for the GOP candidate, making the vote a tie. Another GOP official then decided the contest in the GOP candidate’s favor by drawing his name from a hat. https://en.wikipedia.org/wiki/Virginia_House_of_Delegates_el...
I wish election results came with a margin of error calculation, because too many people assume that a vote result is 100% accurate. Especially to justify the result of a winner-take-all election.
To me it seemed actually stupid to believe that the only thing preventing algorithms from yielding the sympathetic magic required for a peaceful transfer of sovereign power, was their lack of complexity. Arguably, without the ritual element, democracy reduces to a lottery with a biased mechanism, and for it to work, it must necessarily be more than that.
It's a larger philosophical question, to be sure, but it's like comparing a Turing test to an Indifference curve. The first is to determine whether something can convince people it is another person, the second is to predict the point at which you will cease to care enough to choose something else. These are analogous in that, like an AI, we can design an e-voting system that can act convincingly as though it facilitates democracy, but mainly it is just an acceptable substitute for people who no longer care whether the democracy they are dealing with is real or not.
In the case of voting, it's not just a thought experiment, or a product dev question, as by real, I mean sufficiently legitimate for people not to reject the results and cause civil disturbances.
It might sound a bit extreme, but we should really be asking when we institute electronic voting (or counting) whether we are willing to accept a simulation of the ritual we use to grant sovereign powers to people.
Paper ballots cannot be hacked at scale.
Properly designed paper ballots and a national design and education rolled out about them would propel us forwards as a nation.
https://www.techwire.net/news/la-county-oks-open-source-elec...
The trick is just to have people scan a QR code when they register, to make sure it’s one vote one person.
Anyway, such a system as you want has been designed, and is inexpensive and works well with existing systems, and makes it easy for third parties to audit the results: https://en.wikipedia.org/wiki/Scantegrity
This system obviously requires centralized trust (which I’m not against in practice) — but is there any provably secure way to make such a system decentralized? (Yes, blockchain..)
In the Scantegrity sysem you can personally verify who your vote was counted for.
However, a Merkle tree may actually be extremely helpful when tallying the results of various districts. Why not have each person simply receive a random off-the-record perfect forward secrecy token and vote once and have all the votes stored in a Merkle tree?
[1] http://file.lacounty.gov/SDSInter/bos/supdocs/123460.pdf
How do you safely pass votes from the machine to the counter?
I suspect above might be unpopular idea. Especially since obvious downside is that if poorly implemented (or exploited) it could allow establishing who voted what and in any case allow establishing who did and didn't vote. And no, I don't think voter fraud is widespread, or has had any impact of consequence to overall results. But in this age and day, I think we should have a system where a registered voter is first identified by the system before presenting them with the ballot. This way, assuming systems were interlinked, one could vote anywhere in the country, and be given the right ballot for where you live.
And then the obvious next step would be allowing voting via browser. One vote for one registered ID. If system thinks your ID has already voted, have an escalation mechanism to re-cast the vote and investigate.
what's that supposed to mean? it's either mandatory or it isn't. also, all of the options you listed are disproportionately less used by the poor, so it doesn't really address the main issues with mandatory voter id
>And then the obvious next step would be allowing voting via browser.
electronic voting brings up a bunch of other issues. how do you prevent vote selling? how do you secure a (poorly secured) personal computer from being hacked?
I was just thinking that some might want to use this software without identification enforcement, or do it some other way (manually via voter list,...).
I think it would not be unreasonable to say that in order to vote, you need to furnish some form of approved identification. That some people don't have a valid government issued Id is in my mind absolutely a problem and we should find a way to make sure people do have some Id regardless of their background.
>>...allowing voting via browser. >and how does this prevent vote selling?
I don't know. But it appears other countries[0] have managed to make this work. I don't see how it could not work here too.
yet documented instances of voter fraud that could be stopped by mandatory id checks (mainly impersonation) is surprisingly sparse[1]. this either means that the fraudsters are really good at not getting caught, or that this isn't that big of a problem. if you think about it, it really doesn't make sense to commit this type of voter fraud: the penalties are stiff if you get caught, you're putting yourself at risk every time you commit it (as opposed to sitting behind a computer in an extradition unfriendly country), and it scales poorly (each person you recruit can maybe cast 30 votes on voting day if you factor in lineups and driving time, and the chances that one of them snitches goes exponentially as you recruit more people).
[1] https://www.nytimes.com/2018/01/03/us/politics/trump-voter-f...
>I don't know. But it appears other countries[0] have managed to make this work. I don't see how it could not work here too.
The wikipedia page doesn't seem to put it in a good light...
https://en.wikipedia.org/wiki/Electronic_voting_in_Estonia#C...
If you want to increase the bar for people to vote, you HAVE to demonstrate that the current system is problematic.
There have been multiple attempts to find ANY significant voter fraud that would have been at all affected by forcing voter-ID requirements. They have universally found nothing.
As such, there's no real reason to support voter-ID laws, aside from deliberately trying to disenfranchise low-income voters.
This is a common point, but I think it assumes too much about low-income voters. You need an ID to access essentially all of the government services that low-income folks would need. I'm really skeptical that mandating voter ID would have any real effect on these groups (understand that we do sound a bit Elitist talking about what's good for the proles ... I like to err on the side of them being competent enough to use their ID to both get access to welfare programs and, also, vote).
Universal, automatic voter registration moots this entire sideshow.
Just like every other mature democracy.
That's not generally the case, as I understand it. Imagine a low-income senior, with a paid off home (or an inexpensive apartment or public housing unit) and a family that help take care of them. What do they need ID for? They don't drive. They may need medical care, but that only requires a medicare card, which is not valid ID (no picture, and just printed on plain paper). Social Security comes as a direct deposit to their bank account, which they opened years ago, before ID was required, or when they did have ID, since lost (or as a US Treasury check that they deposit or cash). Checks don't require ID. State SNAP/EBT/Food Stamp cards don't have a picture, and aren't valid ID.
How would a counting system for traditional paper ballots enforce that? It's not even a component used in the part of the process where that makes sense
Voting via the browser is such a terrible idea for manifold reasons that ought to be obvious.
Most every US voter registration system I'm aware of requires eligible voters to present some kind of identification. Errors do occur, because we're stubborn and won't implement Real ID. The one exception I know of is (was?) North Dakota.
Further, voters must present some kind of identification and sign-in to receive a ballot at a poll site. For postal ballots, your receiving address is assumed to be proxy for your identity.
Voters can only cast ballots where they're registered.
Voting electronically, by any medium, can neither ensure voter privacy or the public vote count. The gold standard is the Australian Ballot, if you'd like to learn more.
All voting systems I'm aware of tracks ballots. Rules vary by jurisdiction. But generally this prevents double voting. Some places allow newer ballots to replace a prior ballot (IIRC).
The voter ID debate is a red herring.
All jurisdictions require ID to receive a ballot.
The argument is over what forms of ID are accepted.
As the token liberal here, and former poll inspector and election integrity activist, I insist that any required ID (eg drivers license) is provided free on demand to all eligible voters.
Otherwise it's a poll tax, which is unconstitutional.