Overheated Voting Machine Cast Its Own Votes

longacre writes in with the results of a report on voting machines that malfunctioned in NY during the 2010 mid-term elections. "Tests of a number of electronic voting machines that recorded shockingly high numbers of extra votes in the 2010 election show that overheating may have caused upwards of 30 percent of votes in some South Bronx voting precincts to go uncounted. WNYC first reported on the issue in December 2011, when it was found that tens of thousands of votes in the 2010 elections went uncounted because electronic voting machines counted more than one vote in a race. A review by the state Board of Election and the electronic voting machines’ manufacturer ES&S found that these 'over votes,' as they’re called, were due to a machine error. In the report issued by ES&S, when the machine used in the South Bronx overheated, ballots run during a test began coming back with errors."

Comments

Scrap them all

[Fned]

It's clear we're just not ready for electronic voting. Let's stick to paper ballots and re-visit this idea in twenty years or so.

Re:Scrap them all

[TubeSteak]

It's clear that we hired the wrong people to build our electronic voting machines.
Instead of the guys who build ATMs, we should have hired the guys who build slot machines.

Re:Scrap them all

[Anonymous Coward]

Why not just change the machines? Brazil uses them for more than a decade, without any big problems.

Either that or Brazil isn't as good at discovering there were problems after an election.

And before you get offended, I'm Brazilian. I'm also an electrical engineer and software developer, which means I don't trust voting machines, at least not voting machines without a paper receipt to be used for recounts. Which I know the Brazilian machines do not have.

Re:Scrap them all

[Anonymous Coward]

No. We should have hired the guys who print money and other secure papers. Paper voting is superior in every way to electronic voting, except possibly price - and shouldn't we be willing to spend what is ultimately a pittance compared to what we spend on everything else to ensure one of the cornerstones of democracy is eroded away?

Re:Scrap them all

[yakatz]

Except when slot machines are hacked, the developers usually keep it a secret so the casinos will not be investigated. (Mitnick, K (2005). The Art of Intrusion.)

Oh wait, the voting machine companies probably try to do that too.

Re:Scrap them all

[Daniel Dvorkin]

ATMs are incredibly reliable these days. The fact that these POS voting machines are built, in large part, by the same people who build ATMs indicates strongly that Occam's Razor beats Hanlon's (or Napoleon's) Razor here; malice, rather than stupidity or incompetence, is the simplest and most likely explanation.

Re:Scrap them all

[Darinbob]

There is a fundamental problem with elections in the US and many other places, regardless of electronic versus paper. The problem is that once the election is over it is OVER. There is no re-do if someone finds a mistake. In this case the cause of the mistake is discovered 18 months late and the next election cycle has begun! But even in a normal case in the US we have elections early in November and winning candidates take office in January. That leaves no time to invalidate results and hold a brand new election if something goes wrong. We don't have wiggle room like calling for early elections or rerunning them if there are problems. Generally when there are disputes they're not resolved until after it is too late, so we just cross our fingers and hope it doesn't happen again. The cases where a result is held off for more than a couple of weeks is very rare, and always because the counts are very very close. I've never heard of anything being delayed merely because someone thinks there were far more invalidated votes than are statistically expected (or because of evidence box stuffing for paper ballots). The election is a juggernaut and is not slowed down by inconveniences.

So how do you resolve problems like this. It's been 18 months, do you pull the elected officials from that district out of congress and have the state assign a pro-tem replacement? The governor of the state would just appoint whichever candidate belongs to the same party. But we've had 10 years of these problems without things substantially getting better.

Re:Scrap them all

[Anonymous Coward]

Your statement is unproveable because, as the previous poster said, you don't get a paper record of your vote. With paper, you can do a re-count of the actual ballot papers used to determine the result, you can do it any time you like. You HOPE Brazilian elections are clean, but you don't KNOW.

Re:Scrap them all

[JoshuaZ]

Or laziness and a problem of incentive. These companies get in a lot of trouble if their ATMs are hacked or broken into. They don't have to pay much if their voting machines screw up or are easily hackable.

Re:Scrap them all

[MightyMartian]

Paper voting means a physical verifiable record. As to hanging chads the issue is complex and poorly designed ballots.

As to the speed of counting ballots, so what? You have to wait a few hours, or on tight races, a few days. Sounds like a reasonable sacrifice for not having fucked up elections due to equipment failure.

Re:Scrap them all

[MightyMartian]

Actually, in Canada, if you can demonstrate that the irregularities were high enough to have brought an election result into question a judge can order the election results vacated and a new election runs. I'd like to think that if 30% of the votes were lost that the *independent* (there's a keyword right there) election commission would go to a judge and ask exactly that, that the election results be vacated and a new election called. And Canada may find out soon, as evidence of robocall interference may call the results of at least a few ridings into question, which means even if it ends up being a year or more since the election, those results can be discarded and a new election fought.

Re:Scrap them all

[Anonymous Coward]

Hanging and pregnant chads had to do with Florida not cleaning the chad out of their punch card holders. The jurisdiction I worked for for over 10 years ran this technology and it was understood by the people, was transparent and used OPEN SOURCE ballot counting technology. I could give you the interpreted scripts that ran the counting software. This technology was used all over the US until Florida effed it all up.

The old "chaddy" technology counted ballots at a rate of 1,000 per minute. The "new" technology counts them at 100 per minute if your really pushing it hard. And it jams and tears and rips and still misreads stray marks, smudges and even paper imperfections on the ballot. Congratulations, you've taken the technology (punch cards) that still runs many payroll systems and helped Hitler hunt the Jews and thrown it in the trash. :O

California is now heading a major push towards voting by mail to alleviate costs related to polling place operations. This means another line of fraud. I register by mail, I vote by mail, you ask for ID the first time I vote and I give you a utility bill (unverified, could be faked). All of this subverts the intent of the legislature to require a form of ID (State ID or SNN) which is verified against state databases BEFORE your ballot is counted for the first time.

Authority: 10 years programming, running and testing ballot counting software for a major California Jurisdiction.

Re:Scrap them all

[AK Marc]

Have you never been to a magician's stage show? He gets 500 people to all look at the wrong thing at the same time with close to 100% accuracy. And you are claiming that a well timed car backfire won't make people look. Really? Really? All it takes for voter fraud to be easy is for stupid people to think they have a fool proof system, when they are the ones that are the fools.

Re:Scrap them all

[ClickOnThis]

Have you never been to a magician's stage show? He gets 500 people to all look at the wrong thing at the same time with close to 100% accuracy. And you are claiming that a well timed car backfire won't make people look. Really? Really? All it takes for voter fraud to be easy is for stupid people to think they have a fool proof system, when they are the ones that are the fools.

Your speculation pales in comparison to the kinds of hey-nonny-nonny that can be committed with binary bits on a computer system by someone with a malicious intent.

Re:Scrap them all

[Mashiki]

Indeed. Canada's federal elections are all paper ballots and it's very simple. You have a name, a party(now), and you mark in the big circle with an X who you're voting for. We do have electronic voting, but to be honest most people don't like it, and refuse to use it. Paper trails are good.

Re:Scrap them all

[Mitreya]

Instead of the guys who build ATMs, we should have hired the guys who build slot machines.

ATMs are very reliable! Because if an ATM were to spontaneously spit out money, you bet that owner bank will hold the manufacturer responsible and make them pay! So ATMs don't really screw up, ever
This is a symptom of no one holding them accountable. If every lost/wrong vote cost, say, $1000 to the manufacturer, such crap would not happen.

Re:Scrap them all

[ClickOnThis]

No, it doesn't. But there's nothing I can say to get any of the lying idiots to compare a reasonable implementation of paper vs a reasonable electronic solution. It's always the best paper vs the worst electronic. So there's no reason for me to argue the points, other than tell you that you are wrong.

Even the worst paper vs. the best electronic still puts paper on top. Because with paper, there is macroscopic evidence of what has taken place. No matter how well you design an electronic system, it's still too easy to hide tampering in the ghostly phantoms of ones and zeros within computer systems. Go ahead and tell me I'm wrong, it doesn't matter. Anyone who truly understands both systems knows who is right.

Re:Scrap them all

[mwvdlee]

Does anybody ever read /. journals?

Re:Scrap them all

[eastlight_jim]

The argument Stallman uses against this is that we, as voters, have no way to know whether the code actually running on the machine in front of us is the same as the open code that we have reviewed. Ultimately there will come a time when a very select number of people are responsible for compiling the code and putting it on the machine. If those people have a vested interest in some outcome or other then they could tamper with the machine and no-one would know any better. In fact, we would all be thinking it was a secure system because of the "open" nature of it. These things aren't like our PCs, we can't just install VotingMachine From Scratch and be done with it.

Re:Scrap them all

[flyingfsck]

Aha, so what we should do is make the vote machine pay out a $100 bill with each vote receipt. That will ensure that they are designed and built right and it will cause all elligible voters to go and vote.

Re:Scrap them all

[Sique]

Pen-and-Paper-voting is the one system that can be made secure quite simple, and that can be verified by about everyone without actually tampering with voting secrecy.
There are (at least) three conditions which are not easy to align: equality and secrecy of votes.

Equality (each vote counts the same) is only possible if one guarantees that the counting process is open and verifiable.
Secrecy (no one except the voter knows how he voted) is only possible if no one else can watch the actual act of voting.
Integrity (no one can tamper with the vote once the voter cast the vote) is only possible if the votes can be watched without actually knowing the votes.

And here pen-and-paper-voting shines, and no other voting system comes close. Nearly each part of the voting process can be in the open: the ballot box can be opened to the public, controlled by everyone to be completely empty, sealed and be watched all the time by everyone who likes to watch. The breaking of the seal and the counting can (and should) be performed in public, and again everyone who wants to can watch it. The result for the local election office is announced publicly, and publicly written down into the forms and sealed, the votes are put back in the controlledly empty ballot box and sealed again, and the ballot box is then transported (and accompagnied by whoever wants) to the central election office, where the votes according to the sealed forms are tallied and the complete result is announced.
And the actual act of voting can be performed in the voting booth, no one else can watch it, the ballot is folded by the voter, which preserves secrecy and personally put in the ballot box, which preserves integrity. So all three: Intregrity, Equality and Secrecy are preserved.

No electronic or mechanic voting system can come close to this. Each of them has at least one element which should be open closed (for instance counting with a computer is actually a closed process, because it is much faster than anyone can control count, so you have to trust the system to count correctly), or one element which should be closed in the open (the paper record which allows backtracking to the ballot).

tl;dr

Each voting system which performes at least one act of the voting process faster than the human eye can watch it can be tampered with and should not be trusted.

John Connor

[cloudmaster]

I presume that the vote was cast for Skynet, or at least against some relative of John Connor?

Re:John Connor

[Sparticus789]

Nope, Bender. He already tried to win the DC School Board elections. But now we know he started in the Bronx!

Wrong Approach

[sincewhen]

This reminds me of what I was thinking after yesterday's article about Java security problems.

I think society has taken the wrong approach to deploying computers. We execute untrusted code we receive from the internet. We build complex, computerised devices to perform a simple task.

I think that sometimes we should accept that less is more.

Who, honestly, thinks e-voting is a good idea?

[Baloroth]

Seriously, why the hell are people even trying these things? No permanent record of any kind, little to no public oversight of the process, and of course glitches and the possibility for "glitches" on a massive scale that can completely overturn the entire election process. At least with paper voting, cheating is a) moderately easy to catch and b) moderately difficult on a large scale. Mistakes can be corrected afterwards, by examining the paper trail. An e-voting machine? No trail, and a single alteration the code can allow anyone to change the result in absolutely any way they want, with almost zero possibility of detection, and with a single commands.

They are a terrible idea, and honestly any politician/bureaucrat who pushes them should be regarded with strong suspicion, if not of attempting downright fraud, then of bowing to special interests (i.e. the machine manufacturers). Possibly both. And, even if they are really clean of both the preceeding, then they are technologically stupid and shouldn't be trusted to make decisions about these kinds of things anyways.

Re:Who, honestly, thinks e-voting is a good idea?

[Darinbob]

Elections in the US are messy. They were messy before these machines too. Essentially all the elections are run at the local level; not at the state but in each individual county and district. These election districts have very little funding and they're always being beat on to do better, have fewer errors, report results faster, reduce number of complaints, and save money. There is no national standard for how elections should be run, and not all states have standards either.

With the Bush v. Gore circus in Florida a lot of people panicked. Suddenly there was an urgent desire to upgrade the paper ballots even though almost nobody used systems similar to Florida's. At the national level some political pressure came to change things and there was even some funding. So in the madness of "omg fix it!" tons of districts purchased electronic voting machines with very little in the way of rigorous evaluation. But then the money dried up. In the absence of a national emergency things were back to the way they had always been. Problems cropping up here and there were disasterous enough to capture the nation's attention, these were just "glitches", and besides there was no money fix the machines or get new ones. Add to this that the elections were faster and recounts took seconds and no one had any incentive whatsoever to pull out the old dusty paper machines.

Why So Many Problems?

[semilemon]

I'm not particularly knowledgeable on the subject, so I'm hoping someone here can provide some insight. Why do electronic voting systems seems to have so many problems? Yes, they obviously need to be designed for 100% accuracy, but computers and electronic equipment take care of so many other, more complicated operations like flying aircraft and recording financial transactions, all of which should be much more complex but require the same level of accuracy and precision as counting votes. Are voting machines really that bad, are news reports skewing my opinion of them, or am I just unaware of how many problems a paper ballot system has?

Re:Why So Many Problems?

[SlippyToad]

Why do electronic voting systems seems to have so many problems?

I think the employ of Occam's Razor would be quite useful here. There is an un-holy appeal to any designer of such a machine to be able to artificially control the output. We already have the CEO of Diebold publicly promising to deliver votes to George W. Bush, so any protestation of "naw, people who build these things are so trustworthy, nobody would ever actually think to rig an election by deliberately designing a machine to do so.

My very first thought when I read this rigamarole about how the software conveniently malfunctioned to create new votes was, "oh, my god, what a complete bullshit explanation. Overheating CPUs do not malfunction so specifically as to merely add valid data to the processes they are executing. They STOP WORKING COMPLETELY when they overheat, as anyone who has ever spent even a year working with them would know.

So, I'm calling bullshit immediately, and after being fed an incredibly stupid lie about why these machines generated extra votes, I'm inclined to believe the very fucking worst possible alternative explanation. Why else would someone make up such a fucking ridiculous fib?

Re:Why So Many Problems?

[Anonymous Coward]

These machines were scanning PAPER ballots. The scanning mechanism probably malfunctioned in high heat. They scanned the same ballot without errors when cool and with errors (overvotes) when overheated.

This is the worse of both worlds. You don't get the immediate error checking of a true electronic voting machine and at the same time you discard valid ballots because "the computer said so".

OK Enough of this SHIT

[SlippyToad]

Paper and pen ballots.

ONLY.

And while we are at it, let's fix Voter Fraud with one simple tool: a freaking indelible inkwell at the desk where you pick up your ballot. That way, once you've picked up ONE ballot, you cast your ONE vote. People with purple fingers cannot pick up ballots.

Then we can toss all of this disenfranchising "voter ID" crap on the ashpile too. Our elections will guarantee that each person votes just once and every fucking vote is counted. No swinging chads. No overheating vote-generating machines (oh, and does that story smell like ripe bullshit to me -- yes it does!)?

Paper trail. Physically impossible to vote more than once..

Done.

One More Thing:

[SlippyToad]

Move Voting Day to Saturday. The only reason it was on Tuesday was to allow for travel time and to avoid the often-strictly observed Sabbath of the still quite Puritan colonial USA. Make it a Saturday, and make all businesses except essential service and emergency personnel close on that day period, so the people can take their time to vote.

There. That's the last one.

It was voting fraud

[Anonymous Coward]

It shows a cluster of voids in MULTIPLE voting cells in one area. That means

1) it was not random.
2) Multiple machines in multiple buildings all voided?? No, not overheating, you might pretend that this particular part of NY is hot,but different building have different heat characteristics.

That map is a clear voting fraud pattern, it suggests local tampering.

Machines Not Tested

[Mr Bubble]

Norden said so far the machine in the Bronx was the only machine found to have this problem, but itâ(TM)s also the only machine thatâ(TM)s been tested.

God help us.

Re:South Bronx

[game kid]

Anyway, you guys need to come join our wonderful 'write an X on paper' system. We get results the same night, too.

We had mechanical voting booths in the Bronx and NY in general, but then had to change to electronic ones to comply with federal law. (Stupid HAVA.)

Bloomberg called its first use on primary day 2010 a "royal screw-up". I've voted with both old and new machines, and while both seemed to work well, who knows what bits flipped (or were flipped) between feed and count. Personally I think the change was as necessary as the impending invasion of internet TLDs (i.e. not at all).

Get the Choicepoint data

[Anonymous Coward]

Added comment: Get the Choicepoint data, I bet it shows that section of New York votes strongly Democrat or strongly Republican, and it means that someone was trying to change the election by removing that cluster of votes.

Then go subpoena Choicepoint to find out who commissioned political affiliation data for those districts, and start prosecuting these voter frauds.

Re:

[davide marney]

These WERE paper ballots. The thing most people don't realize is that machines are going to be used to count ballots. If the ballots are paper, those machines will be scanners, as in this case in the Bronx. No one is going to count every ballot by hand. Why? Because hand-counting is far more inaccurate than machine counting.

So, here's the thing: if you're going to use a machine to count anyway, it's better to use a machine with no moving parts because they have lower rates of failure. That's how the election officials in Brazil are doing it.

Also, it's worth nothing that according to the report only one machine in the entire district was malfunctioning, election officials were alerted during the vote, and the votes were not close enough for the voided over-votes to have made a difference.

Re:

[sedmonds]

I was a scrutineer for one of the parties at one of the polls in the riding I lived in during the last federal election in Canada. There were two other parties at the poll who had scrutineers. Each of the three of us sat around a table while the deputy returning officer counted each ballot, showed it to the scrutineers, and waited for the scrutineers to not any exceptions. When he was done, the ballots were sealed in envelopes (which the scrutineers were permitted to initial on the seal), and placed in a box for delivery to Elections Canada.

At the end, each scrutineer checked their count against the official count by the deputy returning officer. The vote total was checked against the ballot booklets. All counts were consistent with each other, and the total consistent with the number of ballots cast.

In this polling station there were no irregular or spoiled ballots, and we had a count to report to our candidate HQ, and for the deputy returning officer to report to Elections Canada, in less than a half hour after the polls closed.

There's no need for machines to count votes. And the notion that people can't count votes quickly, and accurately is pure bullshit.