17 ms·

Am I the only one still against network-based, and to a lesser extent electronic-based, voting? It's near impossible to rig or suppress a physical election wit

by aoowii 6y ago

Am I the only one still against network-based, and to a lesser extent electronic-based, voting? It's near impossible to rig or suppress a physical election without a lot of effort, but one person can DDoS an entire network and no one can vote and the whole election needs to be scrapped. Not even the strongest cryptographic or software systems are free from exploits (especially over time) and there's no way to be sure the open source code for the system is the same code actually being served on the system. A lot of software has died by its own hubris by assuming their systems are secure and then a single 17 year old on 4chan finds a bug and ruins it all. You can't afford for that to happen in an election. Forget hackers, some skilled social engineering gets you the votes of thousands, but you cannot do that in person so easily. I'm sure the problems have been discussed extensively but other niche problems include lack of availability for rural areas (which has been a huge problem even with paper voting). I think the only reliable voting system at scale is in person.

plankers 6y ago
You're not the only one. https://xkcd.com/2030/ https://xkcd.com/2030/
- doublesCs 6y ago
  > Our entire field is bad at what we do. Our entire field is bad at what we do, because it's full of coders instead of engineers. Think about it. Building airplanes reliably has been figured out by aircraft engineers, building elevators safely has been figured out by building engineers, and building software reliably has been figured out by software _engineers_ (the NASA kind, who landed a spacecraft on the moon). However, this isn't the experience you have on your regular job. "Oh, oops, everything failed because I forgot to do X". Can you imagine if the moon lander crashed because some guy forgot to do that manual thing he's always supposed to do? Of course you can't imagine, because they put actual engineers in charge of that, not coders.
  - ForHackernews 6y ago
    > building software reliably has been figured out by software _engineers_ This is patently untrue when it comes to security. OpenBSD is the closest thing I'm aware of to a "secure" widely used system, and I'm sure that OpenBSD machines are compromised every day. The day that a network exploit is treated with the same seriousness as a commercial jetliner crash is the day I'll believe software security has grown up.
    - doubleCs 6y ago
      Well when I said reliably I wasn't really thinking about something as difficult as security in a system as complex as an OS...
  - shireboy 6y ago
    You mean like the 125m Mars orbiter lost due to engineers forgetting imperial/metric conversion? https://www.simscale.com/blog/2017/12/nasa-mars-climate-orbiter-metric/ https://www.simscale.com/blog/2017/12/nasa-mars-climate-orbi...
    - doubleCs 6y ago
      I could answer, but I'm not going to because I don't like your sarcasm.
  - the-dude 6y ago
    Flying is only safe because people actually die.
- zelly 6y ago
  Aviation and elevators both run on software written by professioanl programmers so I'm not sure that comic makes a good point. The problem with voting is that it's a distributed systems problem that requires coordination between not just many computers but people that cannot be controlled by code. An elevator can operate as a self-contained system without coordination with other systems. That's why it can be reliable while websites and apps cannot.
sebmellen 6y ago
You can read a some of my comments in this thread addressing various concerns. That said, I would still agree. At least for the next 5-10 years (until we have Starlink-level internet and advances in cryptography/UX), electronic voting systems are not feasible. The nice thing about blockchains is you can avoid DDoS by only allowing people who are "authorized" to "talk to" the blockchain. This can be done by ensuring that "Right to Vote" tokens are only sent to those who would otherwise be participating in the election, and ensuring they can only submit one vote, and one transaction, by sending that token to a specific burn account. This way, with 300 million voters, you would have a cap of 300 million votes. No one else could "submit" a vote, because they wouldn't have permission to on the blockchain network. This is how blockchains avoid DDoS attacks already, but open and public blockchains have the problem that anyone can buy their native currencies, and with enough money can spam the network. With a "permissioned" system for elections, this risk would be mitigated. EDIT: I would appreciate if the downvoters engaged with me or explained their reason for downvoting.
- aoowii 6y ago
  It's worth working on to be sure. Another thing I think is worth working on is educating people on what blockchain actually is (which it seems you are doing). I think paper voting works primarily because people know and trust how the votes are counted and how they get to the counters, for the most part. But personally I barely understand transistors and flip flops let alone blockchain and that makes me slightly worried about how it might be possible to exploit them.
  - sebmellen 6y ago
    I fully share your concerns. There is a massive amount of blockchain infrastructure which 0.0001% of the world population comprehends. Beyond further technical development and the creation of UX and UI libraries around blockchain internals (like https://blockstack.org https://blockstack.org is trying to do), we need more education. I'd love to have more debates like this, but I'm trying to help researchers with simple to use blockchain-based tools (like https://assembl.app/chronos https://assembl.app/chronos, our timestamping service for research outputs). Paper voting is, in my opinion, still the most "secure" way to vote. This is mainly because any sort of voter fraud requires a lot of people and a lot of time, whereas flawed technology can be hacked by very few in a very short amount of time. I'm interested to see how this discussion develops.
atoav 6y ago
You are not the only one. However I don't think security is even a relevant concern, because there is one much more fundamental flaw in all electronic voting systems: The goal of voting is to produce agreeable consent — so it doesn't really matter how you organise voting, as long as everybody afterwards can agree who won, who lost and by how much. So phrased differently, one of the most important properties of any voting systems is that people afterwards can't just call the result into question. This can only be true if most people voting understand the mechanism, can verify it themselves and manipulation on big scales is hard. This is why I think any electronic voting system is problematic. Even a total expert would have a hard time proving that one machine worked as it should on election day. If you then have some Autocrat who raises doubt about the election result whom do you trust? That expert who tells you that manipulation was mathematically impossible or that autocrat whose party you vote?
- chrismeller 6y ago
  And don't forget that the US operates on the Electoral College, which a shocking number of people don't even understand. Since each state has different rules about their EC delegates, popular vote does not necessarily equal winner... Now try telling that to people when there's a big screen with "official" numbers that were all provided electronically an hour after polls closed...
- enaaem 6y ago
  Taiwan has the best way of counting votes. Ballot is hold in the air, the name is called out and someone in the back keeps the score. Everyone can see what is going on. https://www.youtube.com/watch?v=cqKt-lPfJuw https://www.youtube.com/watch?v=cqKt-lPfJuw
  - Beldin 6y ago
    That is a great way to elect the one who buys the most votes. If you want to avoid such shenanigans, you'll need to enforce voter privacy.
    - jadyoyster 6y ago
doublesCs 6y ago
> Am I the only one still against network-based, and to a lesser extent electronic-based, voting? You're not the only one, most people on HN have the same opinion, and all these ideas have been extensively discussed here.
Slartie 6y ago
No, you are in good company. There are those that believe technology would eventually be the solution for any problem, and those that have lived through this phase already and sobered up, eventually coming to the conclusion that technology is not some magic dust able to solve any problem you throw it on, but just a tool like other tools, each of which has its purposes. All of them, not just modern technology, are fascinating, but none of them is magic.
mrweasel 6y ago
Add to that, that the average person have a right to understand how voting works and how their vote is counted. Trying to make voting electronic, regardless of it being block chain or something else, ensures that only a few people are able to understand the system. Right now it easy to verify that count was done correctly, in theory you could go do it your self. Computers will back voting a black box system where verification is going to be almost impossible for all but a few. Pen and paper works, it simple, easy to understand and the process is transparent in all steps except for the few seconds where the voter is in the booth. There’s a talk by PHK, where among other things, he goes into why you shouldn’t mess with elections : https://www.infoq.com/presentations/Predictable-Failures/ https://www.infoq.com/presentations/Predictable-Failures/
- the_snooze 6y ago
  >Trying to make voting electronic, regardless of it being block chain or something else, ensures that only a few people are able to understand the system. This. Too many Internet/blockchain voting proponents tout their systems without appreciating the fact that people need to trust their elections. Trust comes from being open and scrutable. It's amazingly tonedeaf and elitist to insist that election systems need to be cloaked in cryptographic mumbo-jumbo to be trustworthy. It's also dangerous because the poll workers actually carrying out the election won't be able to tell when there's a failure.
djaque 6y ago
No, at least in my circles nobody thinks electronic voting is a good idea. I think Tom Scott has the best argument: you can't build a system that requires you to have a PhD in computer science to fully understand why you should trust it. Every day people need to have faith in voting and not just because a group of experts say they should.
- NestedLoopGoBrr 6y ago
  We put our faith into systems built by PhDs every day. People don’t have to understand why airplanes stay in the air in order for them to be the only practical means of rapid international transport.
  - zdkl 6y ago
    Because the fact that the maths add up to allowing for planes to fly doesn't need to be understood to be accepted. On the other hand, the process of electing a representative from a group must be understood to not be contested or misrepresented.
  - Beldin 6y ago
    But people do have to understand at least at a high level why the announced result matches what the voters put in. The whole point of a voting system is not to pick a winner, but convincingly show that the loser(s) lost. Undermining trust in the voting system is an old trick. Unless the voting system can show such allegations to be wrong, its results will remain disputed.
  - djaque 6y ago
    It's also inconsequential when people say that planes are a lie from the fake news media. Science works whether or not you believe in it. Elections, not so much. If people just decide not to believe the results of an election, then democracy falls apart.
  - mrob 6y ago
    I can watch an airplane fly and be confident that it's actually flying. I can watch election workers count and be confident that they're actually counting. I don't need to know why something works when I can personally observe it working. But I can't observe the operation of an electronic voting system. All I can see is the result, and a fraudulent result looks very much like a legitimate result. I'm forced to understand it (which is likely impossible considering the complexity of modern hardware) or to trust people who have incentives to cheat me.
TheSpiceIsLife 6y ago
I think what you're missing is: Moving forward, large scale election rigging will be seen as a feature by those in power.
- hpcjoe 6y ago
  I hate to say it, but it looks like this has been the case for a while. Look at the history of gerrymandering. What I see is something that could potentially change the basis of trust for elections in a positive way. It won't be perfect, it won't be without its flaws. But it could significantly reduce multiple forms of fraud, and enfranchise more citizens to vote. Again, I've always viewed blockchain as a solution in search of a problem. And this might ... might ... be a problem potentially solvable by an appropriately designed/implemented system.
hoorayimhelping 6y ago
>Am I the only one still against network-based, and to a lesser extent electronic-based, voting? Yes. You are literally the only person who holds an extremely popular, prevalent, and often repeated opinion on the internet and in tech circles. Thank you for your bravery and insight.
slim 6y ago
voting != elections. everyone agrees that voting on elections is not wise. but direct vote for deliberation (for example on laws) works great
segfaultbuserr 6y ago
No, you're not. This is the consensus in tech. Well, at least in the infosec community, this opinion is held by the majority. For example, Bruce Schneier is an advocate of paper ballot for 20 years. * Voting Security (2004) https://www.schneier.com/essays/archives/2004/07/voting_security.html https://www.schneier.com/essays/archives/2004/07/voting_secu... * Securing Elections (2018) https://www.schneier.com/blog/archives/2018/04/securing_electi_1.html https://www.schneier.com/blog/archives/2018/04/securing_elec... * See also: xkcd #2030 - Voting Software > "They say they've fixed it with something called "blockchain." > "AAAAA!!! Whatever they sold you, don't touch it, bury it in the desert, wear gloves." https://xkcd.com/2030/ https://xkcd.com/2030/ > There are lots of very smart people doing fascinating work on cryptographic voting protocols. We should be funding and encouraging them, and doing all our elections with paper ballots until everyone currently working in that field has retired.
- koheripbal 6y ago
  Neither of your two sources talk about crypto here at all. Crypto would allow voters to follow their vote through tabulation, in private, and see exactly how it's being counted. It has only two downsides... 1. People could prove who they voted for after the fact. 2. People who don't understand technology may not understand how it's so secure. ...but the reality is that these issues exist in the current system as well.
  - Steuard 6y ago
    You can't prove who you voted for today: that's a crucial part of the electoral system, not a minor feature. How are you thinking that you could you prove who you voted for? (Keep in mind that you can always request a new ballot if you've "spoiled" your original one, so a photo showing your completed ballot could easily be faked. And my memory is that it's officially illegal to take photos in a polling place anyway, so if you tried to take a picture as your ballot was being fed into the machine, the poll workers would stop you.)
  - segfaultbuserr 6y ago
    > Neither of your two sources talk about crypto here at all. I was quoting the xkcd title text.
snarf21 6y ago
Exactly. The other thing the current system does is that it involves LOTS of people. The people counting a specific polling place may only count 1000 votes. They know their count. They can talk to the other counters at other locations. So if your state or area all of a sudden miraculously goes blue/red, you can know it instantly and have the paper to prove it. This is the beauty of the system. It can't be hacked at scale and you have a hard time inserting fake ballots because each location can triple check their reported district counts.
- elcritch 6y ago
  That’s a good point. Based on that idea, do you think large scale mail in ballots would have similar issues to an electronic system? Normally I don’t follow the news cycle, but this makes it seem large scale mail in ballots would be more readily corrupted at scale?
  - mhh__ 6y ago
    Maybe, but it still involves doing various difficult tasks at huge scale. The paper trail for fraud of that scale is enormous - think of the Watergate slush fund only way bigger and in the age of twitter. This is why paper voting works - you can maybe fraudulently gain a few votes but good luck doing it at a national scale (also an argument against FPTP). Look up granny farming for a more realistic election fraud strategy
627467 6y ago
I agree with you. 1. Traditional paper ballots, manual counting are easy for voters to understand and increases trust in the system. 2. I also agree that traditional voting seems to increase the cost of substantial rigging (with multiple points of failure) compared to all electronic systems we have seen. And yet: we all know the election rigging has always existed and will always exist and many times voters don't always understand how that is possible. I'm in a country where elections are considered by international standards to be fair: and yet, every election common people see reports and talk about rigging through various physical means (ie. Disappearing ballot boxes, stuffing ballot boxes, dead people voting). I agree that as a principle electronic voting seems to expose to rigging at lower cost and higher scale but I don't think it would be necessarily less trusted by the people.
thatcat 6y ago
The Spanish government tried to suppress the Catalan succession vote, but they used IPFS to host the voting system to prevent DDOS and DNS blocking to successfully cast votes anyway. http://la3.org/~kilburn/blog/catalan-government-bypass-ipfs/ http://la3.org/~kilburn/blog/catalan-government-bypass-ipfs/ https://news.ycombinator.com/item?id=15367531 https://news.ycombinator.com/item?id=15367531
diebeforei485 6y ago
It's a generational thing. Younger software engineers who grew up doing everything on their phones see this as something worth solving, because it would vastly improve access, end voter suppression, and people wouldn't have to take time off work to vote anymore. Universal mail-in ballots do address many of those points, but also have downsides (signature verification is tricky). Older engineers on the other hand don't seem to want this to be solved.
- Ericson2314 6y ago
  I don't think that's fair. I'm young enough? And thing that given our current electoral systems and shoddy educations, paper voting is definitely better. If everyone knowingly did some basic applied asymmetric key cryptography every day, and it was thought in grade school, that could change things. If we voted far more often, so that the benefit of a single compromised vote was far less, that could change things. I don't think first electrifying the vote, and then trying to bring about the above reforms, is a good strategy at all.
  - diebeforei485 6y ago
    Meanwhile people have difficulty filling out paper ballots, because it's terrible UX. If you want to have a nice voting system like STV or even ranked-choice, shading ovals on paper ballots is not ideal. Sometimes it feels as though existing problems get a free pass because of tradition, but new issues (even if addressed) are scary and so the entire thing should be stopped. I think the right strategy is to pick a state, implement a voting system there for people to vote on their phones, and then see if it works out or not. Had this been done a couple of years ago, we would have been more prepared for covid-19 this election. And we're only going to have more pandemics and other disasters going forward.
    - stansbury 6y ago
      Iowa Democrats tried an app this year, and it was a colossal failure: https://www.npr.org/2020/02/04/802583844/what-we-know-about-the-app-that-delayed-iowas-caucus-results https://www.npr.org/2020/02/04/802583844/what-we-know-about-...
hpcjoe 6y ago
> It's near impossible to rig or suppress a physical election without a lot of effort No, sadly, it's not [1][2][3]. As some of the texts note, the premise that fraud happens or not is highly politicized. People on one side believe it happens/happened. On the other side, they believe it to be impossible. Though, curiously, those who believe it to be impossible also strongly believe in voter suppression, which would be a form of election fraud. See the Georgia US governor race from 2018 as an example of this dichotomy. Can't have suppression without fraud[4]. Can have fraud without suppression, as this would be one potential form of fraud. Suppression requires fraud. Arguments to the contrary aren't logical. Right now the one I am most concerned with is the vote harvesting operations, which appear to be completely legal in California, despite the rather significant perception of impropriety that handing ones ballot to a non-election person has. I'd be much happier if all these people were non-partisan deputized election workers/officials, oath bound under pain of criminal code to serve the same function. Instead, we have a single party set of "volunteers" to collect votes, with a huge chain of control issue, that is apparently acceptable in one state. Here in MI, I voted by "mail". I got my ballot, marked it, sealed and signed it, then dropped it off in the specific election drop box on the side of city hall. I would not ever hand it to a third party for any reason. For blockchain, I think its been a solution in search of a problem. And I think that this might actually be a reasonable problem to solve, and this one might work. For those worried about DDoS, this would simply require releasing ballots early, returning up to and including the official due date. By mail, electronic, in person via a phone app or similar. You can change your vote up to the end of election day. Only the final vote counts. No conditional ballots counted. Send official election people around to addresses that don't have votes on file starting a month before election day. And it would (if designed/built correctly) enable chain of control, identity/citizen confirmation, etc. That is, it could provide better, more accurate, more complete, more engagement. Or we can retain the mess we have now, with completely inconsistent voting across the states and territories, which is overtly suspect in a number of states, and in some ways, rigged. This is a step in the right direction. It need not be perfect, and we need to take it. [edit: fixed a mis-wording, and deleted a poorly worded sentence] [1] https://www.salon.com/2016/02/14/election_fraud_chicago_style_illinois_decades_old_notoriety_for_election_corruption_is_legendary/ https://www.salon.com/2016/02/14/election_fraud_chicago_styl... [2] https://www.realclearpolitics.com/articles/2020/04/24/28_million_mail-in_ballots_went_missing_in_last_four_elections_143033.html https://www.realclearpolitics.com/articles/2020/04/24/28_mil... [3] https://www.kuow.org/stories/it-s-easy-to-commit-election-fraud-in-washington-state https://www.kuow.org/stories/it-s-easy-to-commit-election-fr... [4] https://en.wikipedia.org/wiki/Electoral_fraud https://en.wikipedia.org/wiki/Electoral_fraud
miki123211 6y ago
There's a pretty easy way to solve this. Give everyone a random ID/password combination (in an unmarked envelope, so no one knows who got what). To vote, you connect to the system, either through a website or even an old, DTMF-based phone system for rular areas, and you cast your vote. After voting ends, a list of all IDs and their votes is published. If an ID didn't vote, it's marked appropriately. There should be lots of separate lists, perhaps one per street/district/village, depending on population density. That way, it would be hard to add fake IDs, while still maintaining voter anonymity. The system is easy for anyone to understand, works on paper as well as electronically, doesn't require complicated cryptographic implementations which might contain bugs, and is much easier to reason about, write and audit. The only drawback is the ability to prove who you voted for, perhaps with a photo, but all current mail-in-voting systems and some in-person-voting system suffer from that problem too, and bribing voters isn't really that much of a problem.
- M2Ys4U 6y ago
  >The only drawback is the ability to prove who you voted for, That's pretty much the biggest problem to have in any election...
- macNchz 6y ago
  This is still vulnerable to a DoS attack as the parent suggested. Not terribly difficult for an individual to tie up all the phone lines or overload the web voting portal. I also agree with the parent comment re hubris about this stuff: the incentives are immensely high for someone to compromise a system like this, and it only takes a small oversight to cause a huge disaster. Everyone gets a random password? Better hope whoever is generating them used a good RNG and a proper source of entropy...
M2Ys4U 6y ago
> Am I the only one still against network-based, and to a lesser extent electronic-based, voting? No, you're not. Electronic voting of any kind is a horrendously bad idea and it should be rejected at every possible opportunity because it cannot be secure.
ddingus 6y ago
No, and I just dropped a comment. E-voting of any kind, unless it results in a public record of the vote, has basic trust issues. Voters cannot know their intent was reflected in the final tally. Blockchain does not help with this.
rudolph9 6y ago
Keep in mind, network packets could be delivered by carrier pigeons. I’m all for paper ballots and conventional signature based system based system but combine that with cryptographically secure techniques for distributed systems (on paper).
KKKKkkkk1 6y ago
Why can we have online banking and Tesla autopilot and online medical prescriptions but not online voting?
- the_snooze 6y ago
  The requirements aren't as strict for money as they are for civil elections. Financial transactions are frequent, so an error isn't that big of a deal in the grand scheme of things. All parties to a transaction know what everyone did; there's no requirement for strong privacy, so being able to roll things back is an achievable (and desirable) failsafe. Money also solves money problems, so you can insure transactions and mitigate losses. Elections don't have those properties.
Spooky23 6y ago
Political people play all sorts of games with elections physically. Don’t underestimate the lack of good faith that people will act with. But the system itself works. The scary thing about electronic voting to me is the centralized nature. It needs to be decentralized to protect the overall system. The current scenario where a trusted entity (the USPS) is being de-legitimized needs to be factored in to protect democracy.
throwaway936482 6y ago
Nope, it's an idiotic idea. In the UK we have an ENTIRELY manual process, no ridiculous voting machines, butterfly ballots or hanging chads and no OCR machines. Just marking a box with a cross, sticking it in a box and then a load of volunteers physically counting the votes while representatives of the parties and anyone else who wants to wander in watches. Postal votes are treated the same and counted at the same time in the same place. The system literally couldn't be simpler or less corruptible. There is no need for anything more complicated because the system works and can be understood by a five year old.
- the_snooze 6y ago
  How many contests do you have on your ballot? Here's a sample ballot from Alameda County, California in 2018 (starting on page 12 going to page 25). That wasn't a presidential year, so this is a bit on the low side. https://www.acgov.org/rovapps/vig/236/38.pdf https://www.acgov.org/rovapps/vig/236/38.pdf Manual counts aren't practical in the US because our ballots can be crazy long with tons of contests and candidates. It would be very error-prone to do this manually.
  - throwaway936482 6y ago
    1. If there is more than 1 contest going on at a time you have more than 1 ballot. Again this is a simple solution to a dumb problem, you have poorly designed ballots.
    - the_snooze 6y ago
      The ballots are designed to accomodate the governments we have in place. We can't easily change the structure of national, state, and local governments. In California, one-ballot-per-contest would mean a voter would get a stack of up to 75 ballots on a presidential year.
      - throwaway936482 6y ago
        Whereas currently they get a multi page document. Seriously, as logistical problems go, print separate ballots is really not a hard one and would cut down on the number if people who accidentally vote for the wrong candidate for the wrong thing. Counting may take longer. It may stretch over a few days. Well, fine. Count the presidential ballots first, announce the results and move on down the list until you get to the dog catcher.
hrktb 6y ago
> Am I the only one still against network-based, and to a lesser extent electronic-based, voting? I think that’s an extreme position that would be hard to argue keeping, in particular in the current situation. I see it the same way we deal with money: what is fundamentally different that makes us accept electronic money transfers as core part of our life but argue there’s no way electronic voting could be possible ? Saying current proposed solutions are not good is different from being against it on principle.
- the_snooze 6y ago
  >what is fundamentally different that makes us accept electronic money transfers as core part of our life but argue there’s no way electronic voting could be possible ? The requirements aren't as strict for money as they are for civil elections. Financial transactions are frequent, so an error isn't that big of a deal in the grand scheme of things. All parties to a transaction know what everyone did; there's no requirement for strong privacy, so being able to roll things back is an achievable (and desirable) failsafe. Money also solves money problems, so you can insure transactions and mitigate losses. Elections don't have those properties.
  - hrktb 6y ago
    Anonymity/privacy is only on the vote content, everything else needs to be traceable (who voted, where, when etc.) And no, losing transactions is a big deal. Not receiving your house sale transfer when the sending bank emited the operation is a big deal, and you won’t solve it with a phone call. It can be repaired with money by the bank, but the system is heavily built to prevent it from happening. Actually if the last mile issue is to have a paper in a box, you could connect a physical printer to a central secured voting machine. To compare, election votes error margin is not small either, and those errors are mostly ignored as of now. I don’t see the status quo as some perfect state that should be preserve ad vitam eternam, and it is more and more preventing some people to vote.
undersuit 6y ago
I'm against it for all your points and many more. We're stupid, electronic voting puts the onus on you to figure out so many things from why your operating system broke voting to why the apostrophe in your last name still breaks things.

by aoowii 6y ago

plankers 6y ago

You're not the only one. https://xkcd.com/2030/ https://xkcd.com/2030/

doublesCs 6y ago
> Our entire field is bad at what we do. Our entire field is bad at what we do, because it's full of coders instead of engineers. Think about it. Building airplanes reliably has been figured out by aircraft engineers, building elevators safely has been figured out by building engineers, and building software reliably has been figured out by software _engineers_ (the NASA kind, who landed a spacecraft on the moon). However, this isn't the experience you have on your regular job. "Oh, oops, everything failed because I forgot to do X". Can you imagine if the moon lander crashed because some guy forgot to do that manual thing he's always supposed to do? Of course you can't imagine, because they put actual engineers in charge of that, not coders.
- ForHackernews 6y ago
  > building software reliably has been figured out by software _engineers_ This is patently untrue when it comes to security. OpenBSD is the closest thing I'm aware of to a "secure" widely used system, and I'm sure that OpenBSD machines are compromised every day. The day that a network exploit is treated with the same seriousness as a commercial jetliner crash is the day I'll believe software security has grown up.
  - doubleCs 6y ago
    Well when I said reliably I wasn't really thinking about something as difficult as security in a system as complex as an OS...
- shireboy 6y ago
  You mean like the 125m Mars orbiter lost due to engineers forgetting imperial/metric conversion? https://www.simscale.com/blog/2017/12/nasa-mars-climate-orbiter-metric/ https://www.simscale.com/blog/2017/12/nasa-mars-climate-orbi...
  - doubleCs 6y ago
    I could answer, but I'm not going to because I don't like your sarcasm.
- the-dude 6y ago
  Flying is only safe because people actually die.
zelly 6y ago
Aviation and elevators both run on software written by professioanl programmers so I'm not sure that comic makes a good point. The problem with voting is that it's a distributed systems problem that requires coordination between not just many computers but people that cannot be controlled by code. An elevator can operate as a self-contained system without coordination with other systems. That's why it can be reliable while websites and apps cannot.

sebmellen 6y ago

You can read a some of my comments in this thread addressing various concerns. That said, I would still agree. At least for the next 5-10 years (until we have Starlink-level internet and advances in cryptography/UX), electronic voting systems are not feasible. The nice thing about blockchains is you can avoid DDoS by only allowing people who are "authorized" to "talk to" the blockchain. This can be done by ensuring that "Right to Vote" tokens are only sent to those who would otherwise be participating in the election, and ensuring they can only submit one vote, and one transaction, by sending that token to a specific burn account. This way, with 300 million voters, you would have a cap of 300 million votes. No one else could "submit" a vote, because they wouldn't have permission to on the blockchain network. This is how blockchains avoid DDoS attacks already, but open and public blockchains have the problem that anyone can buy their native currencies, and with enough money can spam the network. With a "permissioned" system for elections, this risk would be mitigated. EDIT: I would appreciate if the downvoters engaged with me or explained their reason for downvoting.

aoowii 6y ago
It's worth working on to be sure. Another thing I think is worth working on is educating people on what blockchain actually is (which it seems you are doing). I think paper voting works primarily because people know and trust how the votes are counted and how they get to the counters, for the most part. But personally I barely understand transistors and flip flops let alone blockchain and that makes me slightly worried about how it might be possible to exploit them.
- sebmellen 6y ago
  I fully share your concerns. There is a massive amount of blockchain infrastructure which 0.0001% of the world population comprehends. Beyond further technical development and the creation of UX and UI libraries around blockchain internals (like https://blockstack.org https://blockstack.org is trying to do), we need more education. I'd love to have more debates like this, but I'm trying to help researchers with simple to use blockchain-based tools (like https://assembl.app/chronos https://assembl.app/chronos, our timestamping service for research outputs). Paper voting is, in my opinion, still the most "secure" way to vote. This is mainly because any sort of voter fraud requires a lot of people and a lot of time, whereas flawed technology can be hacked by very few in a very short amount of time. I'm interested to see how this discussion develops.

atoav 6y ago

You are not the only one. However I don't think security is even a relevant concern, because there is one much more fundamental flaw in all electronic voting systems: The goal of voting is to produce agreeable consent — so it doesn't really matter how you organise voting, as long as everybody afterwards can agree who won, who lost and by how much. So phrased differently, one of the most important properties of any voting systems is that people afterwards can't just call the result into question. This can only be true if most people voting understand the mechanism, can verify it themselves and manipulation on big scales is hard. This is why I think any electronic voting system is problematic. Even a total expert would have a hard time proving that one machine worked as it should on election day. If you then have some Autocrat who raises doubt about the election result whom do you trust? That expert who tells you that manipulation was mathematically impossible or that autocrat whose party you vote?

chrismeller 6y ago
And don't forget that the US operates on the Electoral College, which a shocking number of people don't even understand. Since each state has different rules about their EC delegates, popular vote does not necessarily equal winner... Now try telling that to people when there's a big screen with "official" numbers that were all provided electronically an hour after polls closed...
enaaem 6y ago
Taiwan has the best way of counting votes. Ballot is hold in the air, the name is called out and someone in the back keeps the score. Everyone can see what is going on. https://www.youtube.com/watch?v=cqKt-lPfJuw https://www.youtube.com/watch?v=cqKt-lPfJuw
- Beldin 6y ago
  That is a great way to elect the one who buys the most votes. If you want to avoid such shenanigans, you'll need to enforce voter privacy.
  - jadyoyster 6y ago

doublesCs 6y ago

> Am I the only one still against network-based, and to a lesser extent electronic-based, voting? You're not the only one, most people on HN have the same opinion, and all these ideas have been extensively discussed here.

Slartie 6y ago

No, you are in good company. There are those that believe technology would eventually be the solution for any problem, and those that have lived through this phase already and sobered up, eventually coming to the conclusion that technology is not some magic dust able to solve any problem you throw it on, but just a tool like other tools, each of which has its purposes. All of them, not just modern technology, are fascinating, but none of them is magic.

mrweasel 6y ago

Add to that, that the average person have a right to understand how voting works and how their vote is counted. Trying to make voting electronic, regardless of it being block chain or something else, ensures that only a few people are able to understand the system. Right now it easy to verify that count was done correctly, in theory you could go do it your self. Computers will back voting a black box system where verification is going to be almost impossible for all but a few. Pen and paper works, it simple, easy to understand and the process is transparent in all steps except for the few seconds where the voter is in the booth. There’s a talk by PHK, where among other things, he goes into why you shouldn’t mess with elections : https://www.infoq.com/presentations/Predictable-Failures/ https://www.infoq.com/presentations/Predictable-Failures/

the_snooze 6y ago
>Trying to make voting electronic, regardless of it being block chain or something else, ensures that only a few people are able to understand the system. This. Too many Internet/blockchain voting proponents tout their systems without appreciating the fact that people need to trust their elections. Trust comes from being open and scrutable. It's amazingly tonedeaf and elitist to insist that election systems need to be cloaked in cryptographic mumbo-jumbo to be trustworthy. It's also dangerous because the poll workers actually carrying out the election won't be able to tell when there's a failure.

djaque 6y ago

No, at least in my circles nobody thinks electronic voting is a good idea. I think Tom Scott has the best argument: you can't build a system that requires you to have a PhD in computer science to fully understand why you should trust it. Every day people need to have faith in voting and not just because a group of experts say they should.

NestedLoopGoBrr 6y ago
We put our faith into systems built by PhDs every day. People don’t have to understand why airplanes stay in the air in order for them to be the only practical means of rapid international transport.
- zdkl 6y ago
  Because the fact that the maths add up to allowing for planes to fly doesn't need to be understood to be accepted. On the other hand, the process of electing a representative from a group must be understood to not be contested or misrepresented.
- Beldin 6y ago
  But people do have to understand at least at a high level why the announced result matches what the voters put in. The whole point of a voting system is not to pick a winner, but convincingly show that the loser(s) lost. Undermining trust in the voting system is an old trick. Unless the voting system can show such allegations to be wrong, its results will remain disputed.
- djaque 6y ago
  It's also inconsequential when people say that planes are a lie from the fake news media. Science works whether or not you believe in it. Elections, not so much. If people just decide not to believe the results of an election, then democracy falls apart.
- mrob 6y ago
  I can watch an airplane fly and be confident that it's actually flying. I can watch election workers count and be confident that they're actually counting. I don't need to know why something works when I can personally observe it working. But I can't observe the operation of an electronic voting system. All I can see is the result, and a fraudulent result looks very much like a legitimate result. I'm forced to understand it (which is likely impossible considering the complexity of modern hardware) or to trust people who have incentives to cheat me.

TheSpiceIsLife 6y ago

I think what you're missing is: Moving forward, large scale election rigging will be seen as a feature by those in power.

hpcjoe 6y ago
I hate to say it, but it looks like this has been the case for a while. Look at the history of gerrymandering. What I see is something that could potentially change the basis of trust for elections in a positive way. It won't be perfect, it won't be without its flaws. But it could significantly reduce multiple forms of fraud, and enfranchise more citizens to vote. Again, I've always viewed blockchain as a solution in search of a problem. And this might ... might ... be a problem potentially solvable by an appropriately designed/implemented system.

hoorayimhelping 6y ago

>Am I the only one still against network-based, and to a lesser extent electronic-based, voting? Yes. You are literally the only person who holds an extremely popular, prevalent, and often repeated opinion on the internet and in tech circles. Thank you for your bravery and insight.

slim 6y ago

voting != elections. everyone agrees that voting on elections is not wise. but direct vote for deliberation (for example on laws) works great

segfaultbuserr 6y ago

No, you're not. This is the consensus in tech. Well, at least in the infosec community, this opinion is held by the majority. For example, Bruce Schneier is an advocate of paper ballot for 20 years. * Voting Security (2004) https://www.schneier.com/essays/archives/2004/07/voting_security.html https://www.schneier.com/essays/archives/2004/07/voting_secu... * Securing Elections (2018) https://www.schneier.com/blog/archives/2018/04/securing_electi_1.html https://www.schneier.com/blog/archives/2018/04/securing_elec... * See also: xkcd #2030 - Voting Software > "They say they've fixed it with something called "blockchain." > "AAAAA!!! Whatever they sold you, don't touch it, bury it in the desert, wear gloves." https://xkcd.com/2030/ https://xkcd.com/2030/ > There are lots of very smart people doing fascinating work on cryptographic voting protocols. We should be funding and encouraging them, and doing all our elections with paper ballots until everyone currently working in that field has retired.

koheripbal 6y ago
Neither of your two sources talk about crypto here at all. Crypto would allow voters to follow their vote through tabulation, in private, and see exactly how it's being counted. It has only two downsides... 1. People could prove who they voted for after the fact. 2. People who don't understand technology may not understand how it's so secure. ...but the reality is that these issues exist in the current system as well.
- Steuard 6y ago
  You can't prove who you voted for today: that's a crucial part of the electoral system, not a minor feature. How are you thinking that you could you prove who you voted for? (Keep in mind that you can always request a new ballot if you've "spoiled" your original one, so a photo showing your completed ballot could easily be faked. And my memory is that it's officially illegal to take photos in a polling place anyway, so if you tried to take a picture as your ballot was being fed into the machine, the poll workers would stop you.)
- segfaultbuserr 6y ago
  > Neither of your two sources talk about crypto here at all. I was quoting the xkcd title text.

snarf21 6y ago

Exactly. The other thing the current system does is that it involves LOTS of people. The people counting a specific polling place may only count 1000 votes. They know their count. They can talk to the other counters at other locations. So if your state or area all of a sudden miraculously goes blue/red, you can know it instantly and have the paper to prove it. This is the beauty of the system. It can't be hacked at scale and you have a hard time inserting fake ballots because each location can triple check their reported district counts.

elcritch 6y ago
That’s a good point. Based on that idea, do you think large scale mail in ballots would have similar issues to an electronic system? Normally I don’t follow the news cycle, but this makes it seem large scale mail in ballots would be more readily corrupted at scale?
- mhh__ 6y ago
  Maybe, but it still involves doing various difficult tasks at huge scale. The paper trail for fraud of that scale is enormous - think of the Watergate slush fund only way bigger and in the age of twitter. This is why paper voting works - you can maybe fraudulently gain a few votes but good luck doing it at a national scale (also an argument against FPTP). Look up granny farming for a more realistic election fraud strategy

627467 6y ago

I agree with you. 1. Traditional paper ballots, manual counting are easy for voters to understand and increases trust in the system. 2. I also agree that traditional voting seems to increase the cost of substantial rigging (with multiple points of failure) compared to all electronic systems we have seen. And yet: we all know the election rigging has always existed and will always exist and many times voters don't always understand how that is possible. I'm in a country where elections are considered by international standards to be fair: and yet, every election common people see reports and talk about rigging through various physical means (ie. Disappearing ballot boxes, stuffing ballot boxes, dead people voting). I agree that as a principle electronic voting seems to expose to rigging at lower cost and higher scale but I don't think it would be necessarily less trusted by the people.

thatcat 6y ago

The Spanish government tried to suppress the Catalan succession vote, but they used IPFS to host the voting system to prevent DDOS and DNS blocking to successfully cast votes anyway. http://la3.org/~kilburn/blog/catalan-government-bypass-ipfs/ http://la3.org/~kilburn/blog/catalan-government-bypass-ipfs/ https://news.ycombinator.com/item?id=15367531 https://news.ycombinator.com/item?id=15367531

diebeforei485 6y ago

It's a generational thing. Younger software engineers who grew up doing everything on their phones see this as something worth solving, because it would vastly improve access, end voter suppression, and people wouldn't have to take time off work to vote anymore. Universal mail-in ballots do address many of those points, but also have downsides (signature verification is tricky). Older engineers on the other hand don't seem to want this to be solved.

Ericson2314 6y ago
I don't think that's fair. I'm young enough? And thing that given our current electoral systems and shoddy educations, paper voting is definitely better. If everyone knowingly did some basic applied asymmetric key cryptography every day, and it was thought in grade school, that could change things. If we voted far more often, so that the benefit of a single compromised vote was far less, that could change things. I don't think first electrifying the vote, and then trying to bring about the above reforms, is a good strategy at all.
- diebeforei485 6y ago
  Meanwhile people have difficulty filling out paper ballots, because it's terrible UX. If you want to have a nice voting system like STV or even ranked-choice, shading ovals on paper ballots is not ideal. Sometimes it feels as though existing problems get a free pass because of tradition, but new issues (even if addressed) are scary and so the entire thing should be stopped. I think the right strategy is to pick a state, implement a voting system there for people to vote on their phones, and then see if it works out or not. Had this been done a couple of years ago, we would have been more prepared for covid-19 this election. And we're only going to have more pandemics and other disasters going forward.
  - stansbury 6y ago
    Iowa Democrats tried an app this year, and it was a colossal failure: https://www.npr.org/2020/02/04/802583844/what-we-know-about-the-app-that-delayed-iowas-caucus-results https://www.npr.org/2020/02/04/802583844/what-we-know-about-...

hpcjoe 6y ago

> It's near impossible to rig or suppress a physical election without a lot of effort No, sadly, it's not [1][2][3]. As some of the texts note, the premise that fraud happens or not is highly politicized. People on one side believe it happens/happened. On the other side, they believe it to be impossible. Though, curiously, those who believe it to be impossible also strongly believe in voter suppression, which would be a form of election fraud. See the Georgia US governor race from 2018 as an example of this dichotomy. Can't have suppression without fraud[4]. Can have fraud without suppression, as this would be one potential form of fraud. Suppression requires fraud. Arguments to the contrary aren't logical. Right now the one I am most concerned with is the vote harvesting operations, which appear to be completely legal in California, despite the rather significant perception of impropriety that handing ones ballot to a non-election person has. I'd be much happier if all these people were non-partisan deputized election workers/officials, oath bound under pain of criminal code to serve the same function. Instead, we have a single party set of "volunteers" to collect votes, with a huge chain of control issue, that is apparently acceptable in one state. Here in MI, I voted by "mail". I got my ballot, marked it, sealed and signed it, then dropped it off in the specific election drop box on the side of city hall. I would not ever hand it to a third party for any reason. For blockchain, I think its been a solution in search of a problem. And I think that this might actually be a reasonable problem to solve, and this one might work. For those worried about DDoS, this would simply require releasing ballots early, returning up to and including the official due date. By mail, electronic, in person via a phone app or similar. You can change your vote up to the end of election day. Only the final vote counts. No conditional ballots counted. Send official election people around to addresses that don't have votes on file starting a month before election day. And it would (if designed/built correctly) enable chain of control, identity/citizen confirmation, etc. That is, it could provide better, more accurate, more complete, more engagement. Or we can retain the mess we have now, with completely inconsistent voting across the states and territories, which is overtly suspect in a number of states, and in some ways, rigged. This is a step in the right direction. It need not be perfect, and we need to take it. [edit: fixed a mis-wording, and deleted a poorly worded sentence] [1] https://www.salon.com/2016/02/14/election_fraud_chicago_style_illinois_decades_old_notoriety_for_election_corruption_is_legendary/ https://www.salon.com/2016/02/14/election_fraud_chicago_styl... [2] https://www.realclearpolitics.com/articles/2020/04/24/28_million_mail-in_ballots_went_missing_in_last_four_elections_143033.html https://www.realclearpolitics.com/articles/2020/04/24/28_mil... [3] https://www.kuow.org/stories/it-s-easy-to-commit-election-fraud-in-washington-state https://www.kuow.org/stories/it-s-easy-to-commit-election-fr... [4] https://en.wikipedia.org/wiki/Electoral_fraud https://en.wikipedia.org/wiki/Electoral_fraud

miki123211 6y ago

There's a pretty easy way to solve this. Give everyone a random ID/password combination (in an unmarked envelope, so no one knows who got what). To vote, you connect to the system, either through a website or even an old, DTMF-based phone system for rular areas, and you cast your vote. After voting ends, a list of all IDs and their votes is published. If an ID didn't vote, it's marked appropriately. There should be lots of separate lists, perhaps one per street/district/village, depending on population density. That way, it would be hard to add fake IDs, while still maintaining voter anonymity. The system is easy for anyone to understand, works on paper as well as electronically, doesn't require complicated cryptographic implementations which might contain bugs, and is much easier to reason about, write and audit. The only drawback is the ability to prove who you voted for, perhaps with a photo, but all current mail-in-voting systems and some in-person-voting system suffer from that problem too, and bribing voters isn't really that much of a problem.

M2Ys4U 6y ago
>The only drawback is the ability to prove who you voted for, That's pretty much the biggest problem to have in any election...
macNchz 6y ago
This is still vulnerable to a DoS attack as the parent suggested. Not terribly difficult for an individual to tie up all the phone lines or overload the web voting portal. I also agree with the parent comment re hubris about this stuff: the incentives are immensely high for someone to compromise a system like this, and it only takes a small oversight to cause a huge disaster. Everyone gets a random password? Better hope whoever is generating them used a good RNG and a proper source of entropy...

M2Ys4U 6y ago

> Am I the only one still against network-based, and to a lesser extent electronic-based, voting? No, you're not. Electronic voting of any kind is a horrendously bad idea and it should be rejected at every possible opportunity because it cannot be secure.

ddingus 6y ago

No, and I just dropped a comment. E-voting of any kind, unless it results in a public record of the vote, has basic trust issues. Voters cannot know their intent was reflected in the final tally. Blockchain does not help with this.

rudolph9 6y ago

Keep in mind, network packets could be delivered by carrier pigeons. I’m all for paper ballots and conventional signature based system based system but combine that with cryptographically secure techniques for distributed systems (on paper).

KKKKkkkk1 6y ago

Why can we have online banking and Tesla autopilot and online medical prescriptions but not online voting?

the_snooze 6y ago
The requirements aren't as strict for money as they are for civil elections. Financial transactions are frequent, so an error isn't that big of a deal in the grand scheme of things. All parties to a transaction know what everyone did; there's no requirement for strong privacy, so being able to roll things back is an achievable (and desirable) failsafe. Money also solves money problems, so you can insure transactions and mitigate losses. Elections don't have those properties.

Spooky23 6y ago

Political people play all sorts of games with elections physically. Don’t underestimate the lack of good faith that people will act with. But the system itself works. The scary thing about electronic voting to me is the centralized nature. It needs to be decentralized to protect the overall system. The current scenario where a trusted entity (the USPS) is being de-legitimized needs to be factored in to protect democracy.

throwaway936482 6y ago

Nope, it's an idiotic idea. In the UK we have an ENTIRELY manual process, no ridiculous voting machines, butterfly ballots or hanging chads and no OCR machines. Just marking a box with a cross, sticking it in a box and then a load of volunteers physically counting the votes while representatives of the parties and anyone else who wants to wander in watches. Postal votes are treated the same and counted at the same time in the same place. The system literally couldn't be simpler or less corruptible. There is no need for anything more complicated because the system works and can be understood by a five year old.

the_snooze 6y ago
How many contests do you have on your ballot? Here's a sample ballot from Alameda County, California in 2018 (starting on page 12 going to page 25). That wasn't a presidential year, so this is a bit on the low side. https://www.acgov.org/rovapps/vig/236/38.pdf https://www.acgov.org/rovapps/vig/236/38.pdf Manual counts aren't practical in the US because our ballots can be crazy long with tons of contests and candidates. It would be very error-prone to do this manually.
- throwaway936482 6y ago
  1. If there is more than 1 contest going on at a time you have more than 1 ballot. Again this is a simple solution to a dumb problem, you have poorly designed ballots.
  - the_snooze 6y ago
    The ballots are designed to accomodate the governments we have in place. We can't easily change the structure of national, state, and local governments. In California, one-ballot-per-contest would mean a voter would get a stack of up to 75 ballots on a presidential year.
    - throwaway936482 6y ago
      Whereas currently they get a multi page document. Seriously, as logistical problems go, print separate ballots is really not a hard one and would cut down on the number if people who accidentally vote for the wrong candidate for the wrong thing. Counting may take longer. It may stretch over a few days. Well, fine. Count the presidential ballots first, announce the results and move on down the list until you get to the dog catcher.

hrktb 6y ago

> Am I the only one still against network-based, and to a lesser extent electronic-based, voting? I think that’s an extreme position that would be hard to argue keeping, in particular in the current situation. I see it the same way we deal with money: what is fundamentally different that makes us accept electronic money transfers as core part of our life but argue there’s no way electronic voting could be possible ? Saying current proposed solutions are not good is different from being against it on principle.

the_snooze 6y ago
>what is fundamentally different that makes us accept electronic money transfers as core part of our life but argue there’s no way electronic voting could be possible ? The requirements aren't as strict for money as they are for civil elections. Financial transactions are frequent, so an error isn't that big of a deal in the grand scheme of things. All parties to a transaction know what everyone did; there's no requirement for strong privacy, so being able to roll things back is an achievable (and desirable) failsafe. Money also solves money problems, so you can insure transactions and mitigate losses. Elections don't have those properties.
- hrktb 6y ago
  Anonymity/privacy is only on the vote content, everything else needs to be traceable (who voted, where, when etc.) And no, losing transactions is a big deal. Not receiving your house sale transfer when the sending bank emited the operation is a big deal, and you won’t solve it with a phone call. It can be repaired with money by the bank, but the system is heavily built to prevent it from happening. Actually if the last mile issue is to have a paper in a box, you could connect a physical printer to a central secured voting machine. To compare, election votes error margin is not small either, and those errors are mostly ignored as of now. I don’t see the status quo as some perfect state that should be preserve ad vitam eternam, and it is more and more preventing some people to vote.

undersuit 6y ago

I'm against it for all your points and many more. We're stupid, electronic voting puts the onus on you to figure out so many things from why your operating system broke voting to why the apostrophe in your last name still breaks things.