Slashdot Mirror
Voting Machines Malfunction: 5,000 Votes Not Counted In Kansas County
An anonymous reader writes A malfunction in electronic voting machines in Saline County, Kansas, left over 5,000 votes uncounted. That's roughly one-third of the votes cast. Counting those 5,207 votes didn't change any outcomes in this case however. “That’s a huge difference,” county Chairman Randy Duncan said when notified by the Journal of the error. “That’s scary. That makes me wonder about voting machines. Should we go back to paper ballots?”
Damn, even Kansas can figure this one out.
I think with all the programmers on this site , we should be insisting on paper voting. At least there is a reliable record to go back to (and no chads jokes please) recount.
Otherwise, why bother voting on a machine you don't get to see the source code for. You having a choice will not matter to whomever controls the code.
_ _ _ Go for the eyes Boo! GO FOR THE EYES!
What if the cast votes simply went to the system equivalent of /dev/null? That would be the electronic version of shredding ballots, with no unsightly cleanup or disposal of shredded paper.
Votes? What votes?
Left MS Windows for Linux Mint and never looked back!
Vote for Bernie in 2016!
We switched to permanent absentee voting the moment they introduced electronic ballots in our county.
" ... Should we go back to paper ballots?"
Because we know that there have never been any missing votes or other irregularities with paper ballots.
The old systems work. The electronic system is too prone to failure and abuse. Paper or pottery shards.
Seriously, the rush to electronic voting after the 2000 Presidential election was just a bad idea all the way around -- and, frankly, most IT people with any experience were saying so. It is vastly, vastly harder to change physical media than to change electronics.
Bruce F. Webster (brucefwebster.com)
the more they do not trust dre voting machines. Voting Machines Elect One Of Their Own As President
This is old hat, and honestly the horse has been beaten to powder on slashdot, but systems that are both complex as well as powerful should be open source. Breathalizers and voting machines have no intrinsic monetary value in a society. Certainly it is a need to perform such tasks, but the greater good, the preservation of liberty and the accurate as well as precise regulation of a functional society, are of such an overwhelmingly greater imporance as to render the quite visible hand of the american free market moot. But we're hardly a capitalism here anymore. We're a plutocratic oligarchy.
Paper ballots are pretty damn open-source.
Just because a voting machine is supposedly running open-source software doesn't preclude tampering - hardware or software.
Voting needs a system where lay people can look at the entire process and spot irregularities. That means a simple system, using well-understood technology. Paper and pencil fits the bill. Integrated circuits anything, not so much. And that's before we get into tussles with voting machine vendors that insist their source code is a "trade secret". Secrets in our voting machines? We can't afford to stand for that, no. So yes, yes we should go back to paper, and stay there.
So there's about 15000 to 18000 votes to count?
Paper ballots. Electronic sounds awesome, but it's a lot of hassle for a small amount of votes.
Say you've got 5 polling stations with 4 people at each one, so 20 people. 350 or so ballots per station, each person has to tally up 100 votes at the end of polling.
You could count the entire lot twice in an hour at 4 ballots a minute per person.
So your 5 voting machines cost, what, $5K each? So $25K all up?
You can pay those 20 people $500 for that one day and spend $10K on wages.
You print 30,000 voting forms (at 5 cents each that's $1500) and getting some nice locked boxes ($2000) and storage of ballots for 12 months ($1000) in case of recount.
Oh look, you've got $10.5K left over. Use that to make a park nice and pretty somewhere.
You are in a twisty maze of processor lines, all alike.
There is a lot of hype here.
What an astounding surprise that voting machines malfunction so frequently.
That's totally not what I would expect from the US government.
The article is a little vague on exactly what happened.
A malfunction of electronic voting equipment left 5,207 votes out of the original Nov. 4 Saline County vote total, but no election outcomes were affected, according to the Saline County Clerk’s Office.
Then at the end
Outcome wasn’t affected
Merriman said that had the extra votes resulted in a change in the outcome of the election, everyone would have been notified immediately.
The problems occurred in machines at four voting locations in the following precincts: 12-13-14; 17-18-19; 20-22; and 15-16.
Votes for Sen. Pat Roberts, Rep. Tim Huelskamp, Gov. Sam Brownback and Kansas Secretary of State Kobach all slightly increased.
Opposition to the jail/justice complex increased from a 953-vote difference to 1,748 votes, or from 53.95 percent to 55.08 percent.
So they evidently found the missing votes. But I'm not sure how.
Saline County Clerk Don Merriman said after the meeting that four of the 34 PEBs, or Personal Electronic Ballots, were not reading correctly on election night, which left the votes out of the original count. The problem has been fixed, he said.
He said the missing votes weren’t discovered until after votes were canvassed on Nov. 10. Merriman said he learned of the error during a “triple check” with flash cards from the PEBs.
...
The error was found the afternoon after votes were canvassed when flash card totals were compared to the printed totals.
“We always pull those flash cards and check those final totals to make sure we are OK,” he said. This is the first time we’ve had the PEBs act up like that. I’m pretty sure it is the programing in the PEBs.”
So which was missing votes? The flash cards or the printed totals? What are the printed totals? Just a summary or actual printed ballots?
If the printed totals were actual printed ballots that voters checked then I don't think there's anything to worry about.
But if there's no actual per vote record and people are just relying on the machines to correctly record the votes then I have to wonder how monumentally stupid people are to use or even create a system that insecure.
I stole this Sig
Yes, yes, yes.
I live in the Seattle area of Washington State. We used to have a nearly perfect system and I would like to see it adopted everywhere. (We now have mail-in voting only, which is convenient but I worry about fraud.)
Here's the perfect system:
Ballots are stiff paper/very light cardboard, printed with oval "bubbles" next to the things for which you can vote. You vote by filling in a bubble with an ink pen.
Once you are done voting, you feed the ballot into an optical reader over a collection bin. If you have made any obvious mistakes (such as voting for both candidates for a single position) the machine kicks the ballot back out to you; you then get a fresh ballot and vote again.
Once my wife had a little ink smudge on her ballot, and the machine kicked it back. It was designed to err on the side of absolute clarity; if it accepted a ballot, that ballot was unambiguous.
The optical reader keeps an unofficial tally, and at the close of voting the tally is forwarded to the state elections department. An unofficial but very accurate result is available within an hour or so of the end of voting.[1]
There are physical paper ballots so there is a literal paper trail. Recounts are easy.
There is no "hanging chad" problem; the optical scanner at the polling place makes sure that each ballot is unambiguous.
Then all you need is a good "chain of custody", making sure that all ballots are delivered (and no fake ballots are introduced into the counting).
[1] Of course, absentee ballots will also be counted and the absentee results will not be available that fast; but non-close election results will be known that fast.
lf(1): it's like ls(1) but sorts filenames by extension, tersely
it might be state-wide by now.
I've been an elections inspector (head worker at a polling place) in Alameda county a few times, and AFAIK almost every county in California is using this type of optical-scan system. There are also some special systems for disabled voters using touchscreens or audio devices.
My thinking is that it could be programmed to reject valid votes to give an edge to one political candidate
If a single machine were reprogrammed, it would be detected when the scanner cartridge was audited. There are also uniquely numbered pull-tie-like seals protecting the sensitive parts of the machine, so any unauthorized access is likely to be noticed.
or to give poll workers knowledge of who is voting democrat or republican, etc.
We already have this knowledge, as voter registration is listed in the street index used to identify voters at the polling place. Not only that, but another copy must be posted outside the polling place. As voters are crossed off the list, the "outside" street index can be used by party campaigns to see who hasn't voted. In some places, they'll actually use this information to go to your house and ask you to go vote (if you're registered with their party).
.: Semper Absurda
I think you replied to the wrong post, but thank you anyway. FWIW, I too have worked the polls, back when we still used the punched cards. I won't say that it would have been impossible to stuff the ballot box back then, but I will say that the procedures required to run the precinct and close it up properly would have made it very hard, especially as members of the public were always allowed to watch it being done. My feeling is that as long as there's a verifiable paper trail that can be re-checked, the exact method isn't that important. I'd only be worried if voting were by touch-screen or something similar, with no other record kept.
Good, inexpensive web hosting
I recall seeing some video's online of electronic voting machines performing vote switching on the users within a day of this last election. Not to be trusted, any malicious algorithm can be slipped in to mess with the vote. I also saw them 2 years ago after that election. When it's happening enough, that people can get a cell phone video of it happening, it's happening too much. How many people don't realize that the machine mis-recorded or switched their vote between the time they selected the candidate, and the time they press the submit button?
Here is one: https://www.youtube.com/watch?... and another: https://www.youtube.com/watch?...
Whether voter fraud, calibration issues, or electronic malfunction, it doesn't matter, as in all cases, there is no way to go back and re-check and re-count the ballots.
No of course not! Electronic systems makes it much easier to rig the system for our political 'elite'.
Paper ballots are pretty damn open-source.
Just because a voting machine is supposedly running open-source software doesn't preclude tampering - hardware or software.
Feels like I've said this 100 times now:
Electronic voting: bad.
Computer-assisted voting: good.
Sure, fine, have a touch-screen and pretty pictures and good usability in general, all of that is great. Then have the voting machine print a paper ballot, which is then cast normally. You can check the paper, or just use the paper yourself, if you don't trust the computer, or if it breaks, or has been hacked. And since almost all ballots will be printed cleanly, there will be little room for 2000-style "dimpled chad" and "interpreting the voter's intentions".
Socialism: a lie told by totalitarians and believed by fools.
I'd like to see good electronic FOSS options, but the I just don't trust total electronic voting. My state uses the touch screens to print paper ballots that are human readable and machine scan-able. The paper trail provides the good backup evidence you need. I just don't trust an entirely electronic system.
The people running the polls where I voted all seemed like nice people, but I doubt there was even one of them who can program a VCR.
quiquid id est, timeo puellas et oscula dantes.
I concur. A development methodology ("open source") will not address any of the deficiencies (when viewed from the voter's perspective, the perspective that should matter most) of voting. No matter how much one trusts a voting program, there's no way to be sure that the computer used for voting is running only software one trusts. No electronic system can compete with the simplicity and recount-friendly approach of what is called for here: voter-verified paper ballots.
So address to the question in the /. summary: You never should have stopped using voter-verified paper ballots.
There are computers one can purchase that do as the parent post specified—the voter feeds in a blank ballot (one which they could have filled out manually if desired) and the computer (which has a scanner and printer attached) will scan the ballot, help the voter by showing the choices on a screen, reading the ballot aloud, or reading the ballot text to headphones, and then collect votes from the voter. Then the computer's printer will print the voter's votes on the paper ballot, and eject the printed paper ballot to let the user inspect that printed ballot. At this point the voter can choose to carry the voter-verified paper ballot to be counted or spoil that ballot and start again. The voter can also feed in a marked up ballot (marked by hand or by computer) and let the computer summarize the votes which that ballot specifies. These features let the blind and/or illiterate vote without losing their privacy by forcing them to find & bring in someone else to mark up their ballot for them. This is as close to computers used in voting as one should want to get.
Digital Citizen
I was kinda replying to both posts, so I just replied to the non-AC. I did use the pure-touch screen machines when I was a student poll worker in 2004. Those were the Diebold machines which connected to a phone line, and whose only paper trail was receipt printed by the machine on heat-sensitive paper. I remember advising many people to submit paper ballots instead...
.: Semper Absurda
The state of voting machines is pathetic. Old, slow technology maintained by a company who bought the company who bought the company who contracted with the dead guy who wrote the system. Running Windows XP machines just to be able to count the damned cartridges with unsupported ATA Flash card hardware. Running Windows XP to count paper ballots at about 100 cards per minute when we used to pull 1,000 cards per minute on our old decertified system (thanks Florida).
Over $20 million+ down the toilet for 2 years of use.
SO YES PLEASE! Move it all to paper... Oh wait, we can't.. we have to accommodate the blind people who don't vote at the polls anyway. UGH! Thanks ACLU for standing up for those who didn't ask for and don't even use the tech.
Forgive my anon... I like my job.
>My thinking is that it could be programmed to reject valid votes
That's easy enough to avoid, and I believe most paper&canner polling places do so: Have the voter feed their ballot into the scanner, which then immediately confirms or rejects it. That way the ballot is rejected right in front of the voter, and they can fill out a fresh ballot if there are any problems.
--- Most topics have many sides worth arguing, allow me to take one opposite you.
Paper ballots are pretty damn open-source.
Just because a voting machine is supposedly running open-source software doesn't preclude tampering - hardware or software.
I can remember one wise lecturer in my computer science course gave a challenge to come up with a system to solve a customer's problem. Being CS students we designed everything requiring the use of a computer. At the end he asked us if we had considered whether a non-computer based system would have actually have done a better job. While in the particular case the answer was no, it did show us that sometimes we use technology for technology's sake and not to solve the problem in the best possible way. Voting machines should be approached in the same way and the opti-scan mention by another poster certainly seems to strike the right balance between solving the problem and not throwing the wrong technology into the mix.
Jumpstart the tartan drive.
there will be little room for 2000-style "dimpled chad" and "interpreting the voter's intentions".
There should already be little room for this. "Voter demonstrates intentions by poking hole in piece of paper. No hole, no intention to vote." Very simple. The failed assumption is that every person who cast a ballot intended on voting for every position and if there wasn't a hole there was a mistake. People who had no intention of voting for any candidate got their vote counted anyway.
This voting system was approved by both parties prior to the election. It wasn't a surprise dropped out of Heaven on an unsuspecting public. The time to say "gee, it's too hard to poke a hole" was before the election, not after.
We shouldn't have to find an excuse for preventing that kind of nonsense. It should be SOP that people can refuse to vote for a particular office and have it honored. It should be SOP that those who followed the instructions get their votes counted and those who don't don't.
Bingo. Having computer assisted voting that produces a ballot that is both machine and human readable is a must. Without this paper trail, you have absolutely nothing. Even with crypto, crypto doesn't protect against erasure, and an "accidental" erasure of votes on a voting machine can sway an election.
I was working on an e-voting prototype using Java in the late 1990s. No matter how it worked, there was no way to secure it, so I gave up on the project, because if the device couldn't be hacked, the data on it was destroyable. Distributed storage could easily be hacked/tampered with, and would be hard to admin by volunteers. The hardware could be made more secure, but it would completely destroy voter anonymity.
Instead, David Chaum's Verifiable Voting system is the absolute best thing out there. It provides not just anonymity for votes, but validates ballots were done correctly.
Back when I had punch card voting (Benton Co., OR,1970's) there was a serial number on a tear off strip on the card. When the punch card was voted the serial number strip was torn off and saved. After the election the election officials could verify they still had every serial number they had printed either on the unvoted ballots or the torn off SN strips and they had no duplicates. They could also make sure the number of torn off strips was the same as the number of voted ballots. It was a decent system and kind of the high tech of its time. After all it was computer punch cards just like I used to run my FORTRAN programs on the mainframe.
My local precinct uses electronic black box voting machines. On election day (Nov. 4, 2014) I requested a paper ballot. They told me I could not vote with a paper ballot. I told them that I had checked with the election division of the secretarty of state office for my state and that I did in fact have the right to vote using a paper ballot. They asked me to wait, they called a supervisor. The supervisor then called the secretary of state office. After a few minutes they issued an apology to me and said that yes I could vote with a paper ballot and yes my vote would count. I voted using a real paper ballot which I placed into a real ballot box.
If they do not allow you to vote using a paper ballot then they are denying your right to vote.
It is not possible to verify a vote with an electronic voting machine.
They can be hacked, fooled, or in any number of ways made to provide bad numbers.
Paper ballots find their way to peoples home for safe keeping only to stuff the boxes themselves (cite:/. article)
But paper ballots are actually the only way to go and how I vote. but by mail only, I'd much rather stand in line to fill out my "paper ballet; I'm never sure it's filled out properly and counted or even made it to it's destination.
Yes electronic voting being outlawed long ago is my feelings on the matter; now voting by mail running a close second.
Note: I live in Washington State, I knew Charter lost (and crying) as I was headed to the school to vote, I still voted and the laws have since been changed to prevent that from happening again. But that was just the news services attempting to beat each other and just wrong.in so many ways.
I am an officer of election. We use DRE machines. We black-box test them before each election, the exact same technique that is the gold standard for software testing by any reputable QA team. A properly-tested machine is far better than any paper-based system. It provides immediate feedback to the voter. It can display in large fonts. It can be used by the disabled. It has almost no moving parts. It cannot get jammed. It does not have to guess what a vote is, it is unequivocal. Paper is a miserable medium. It gets lost. It gets wrinkled. It can be marked haphazardly or incompletely. NO ONE counts paper ballots by hand. Humans are TERRIBLE at repetitive tasks, that's why we invented machines to do this kind of work for us. Even optical ballots are counted by machine, so it is no different than a direct-recording machine -- except it has many more moving parts, and has to GUESS at the voters intent.
I miss New York City's steampunk mechanical voting machines. Designed in the 1920's, and still in use through the new millenium, though I think they have been phased out by now.
Back when I worked the polls in the '80s and '90s, California was using the punched cards. Not only did we have to keep and return all of the stubs, we had to destroy or deface all unused ballots as part of closing down the station. If there was a low turnout, this could be the longest part of the job as we used pencils or pens to write a big X across the face of each ballot to render it unusable.
Good, inexpensive web hosting
Probably driven by the media since they want a result as quickly as possible so they can sell more tv-time. I have no idea how the presidential election works in the US but I assume here that the president elect doesn't take over directly, it probably takes some months before he/she can take office anyways so time should not be an issue for the election in it self. And also since the result is to last for four years, having a result in seconds seams quite useless.
Many years ago our company was asked if we could develop electronic voting systems for elections (we do, in fact we invented electronic voting systems decades ago for conferences and audience interactions, so we basically were a logical choice). The customer intended to buy a complete electronic voting infrastructure for a whole country, so this was very tempting. I was tasked to research into this topic, and have examined this very thoroughly from every angle possible.
My conclusion: There is no, absolutely NO way to get the level of democratic voting quality from electronic ballot systems that is comparable to classic paper ballots. The risks are immense, the gain neglectable.
The electronic system is in no way verifyable by the average voter or voting administrator. Anybody can look into a ballot box before the vote starts and see that it is empty, people can watch over the whole thing to verify that everybody casts only one vote, and wittnesses and recounts can see that every vote from the box is counted for the right candidate. But nobody can do this in an electronic voting system. Yes, they can click on a button and the system tells them "0 votes in ballot box", but they cannot verify this. The voters cann press "A", and the machine tells you that your vote was cast for "A", thankyouverymuch, but internally it could just drop the vote or count it for "B" or "C". Nobody could check this. At the end, the machine would display some numbers for A, B, and C, and you have to believe them.
And this is just the logic part of the problem. On top of that there is the question of technical reliability and user errors. There have been voting systems with touchscreens that needed to be calibrated before use, and there have been several cases where mis-calibration led to votes being cast for the wrong candidate/party (just as an example, whoever knows a technical system will know thousand ways it could fail). How does the system cope with a power loss during voting? Has the vote you just cast been counted or not? And what about the ease of vote? You and I can cope with "press candidate button, verify choice, press submit button", but an astonishing number of people can not (anyone who ever did tech support will not be that surprised).
All the key requirements to a democratic vote cannot be established simultaneously with an electronic voting system: Verifyability, integrity, secrecy. Yes, you can do a lot in the realm of integrity (like they do in Vegas for the one-armed bandits), but the stakes are way higher and so is the temptation to fix the game in a way that will go undetected even by the toughest inspection (and you cannout tough-inspect every electronic ballot box after every election!). And if you want a really reliable system, you will loose the secrecy factor. If you want secrecy, the verifyability and integrity will go down the drain. It is in fact worse than the business classic "Iron Triangle" (Fast, Good, Cheap, pick any two), it is more or less a "pick one". And for a true democratic vote, you will need all three.
The only advantages that an electronic ballot system can give are the results seconds after the closing of the ballot station and no problematic votes where people have to decide whether a vote is valid or not. Thats why the politicians LOVE electronic voting - it gives them nice results in time for the evening news. But do you really want to sell away the integrity of the last democratic instrument left for the citizens for saving a few man-hours in each ballot station? And I'd rather wait for the morning paper with the final results from a paper-based, democratically obtained election result than seeing grinning polititians congratulating themselves in the evening news, claiming their win from a quite doubtful, error- and manipulation-prone process.
In the end, I had a long and intense talk with our company founder and CEO and could convince him that electronic voting is a bad idea for democracy, and he communicated this very result to the customer. And as the customers intention was to have a democratially sound election system, he agreed.
I've read the stuff from David Chaum, and it is bullshit. Sorry to be so harsh, but you'll lose secrecy if you want verifyability - this is part of his method, and he even states this so. So even in a perfect world, this would not work. It might give you a verifyable vote, but not a democratic one when compromizing a key factor. Sorry, David, but you'll need a spoonful of reality.
In Sweden we have had Paper ballots forever. It has always worked. Very steady system. Doesnt cost a lot. Always reliable.
I don't understand why US and Canada are putting a weakness into the democratic system by using electronic stuff. It is just stupid.
That can count?
The only problem I see with computer printed ballots is that it allows for easier ballot box stuffing. When everyone manually fills in the circle or completes the line you get a lot of variation even with that simple action. Printed ones look all time same and would be more difficult to verify as cast by humans.
I used to work on casino games, and gosh, with all that money at stake we never lost track of a penny in the field.
Also the American elections are insanely complicated. Elections every 2 years where every second one is more important so voter turnout varies a lot and they vote for multiple offices from dog catcher, judges, prosecutors, sheriffs up through the regular municipal staff such as school boards, mayors and such, county officials, State representatives including sometimes governor and a good chunk of their federal government including President.
It is simplified by only having 2 political parties so rather then taking the time to learn about all the candidates they can just vote a straight ticket. Still ballots are complex and vary on location with just a couple of miles being enough to mean a different ballot.
I'm Canadian, we have municipal elections which are the most complex, especially since often there are no parties, then Provincial elections and separate Federal elections, both where you just tick off one name. Parties can still vary at the provincial vs federal level so new parties can arise.
https://en.wikipedia.org/wiki/Inverted_totalitarianism
Here in Sweden we have different papers for each party, that is instead of crossing of a name you put a whole paper for party x into an envelope. Pro is that this makes the counting much easier and also is not complicated for the voter. Con is that it's very expensive for a new/small party to print and distribute papers to all voting places (if you get over 1% in an election then the state will pay and manage the distribution). You can also write the party name on a black piece of paper if you want to (or if the party you want to vote for doesn't have papers in your voting place) which of course negates the pro but it's rare enough to not make a real impact.
If there's a yes/no vote done then each such vote gets their own envelope and there is a yes or no paper to choose from, so no crossing their either.