Slashdot Mirror
California Grills Diebold Over E-Voting Foul-Ups
orthogonal writes "Electronic voting machine producer Diebold admitted today that 'thousands' of voters were turned away from the polls during the Super Tuesday Presidential Primary because of flaws in Diebold's machines. Diebold Election Services Inc. president Bob Urosevich said 'We were caught', and answered 'yes' when asked 'Weren't [California voters] actually disenfranchised?' Today, California officials may recommend decertifying some or all of Dielbold's machines for the November General Election." Reader TargetBoy adds: "Diebold knowingly used uncertified software in California elections. Especially interesting is the comment that, 'The law firm's memos reflect a corporate defense firm on a $500,000-a-month campaign to protect Diebold.' Wonder how much it would cost to just fix the problems?" Apparently India is having evoting problems of its own: purple writes "The world's largest democracy is in the midst of a 4-month election marathon. Except this time around the whole thing is run electronically. And, surprise surprise, things seem to not be working perfectly. Some polling booths have been ordered to re-poll due to malfunctions in the electronic voting machines. In another article, 191 voting booths were ordered to re-poll. Other polling locations seem to be operating on voter lists from 2001. I suppose the good news is that these errors were caught before they could have really screwed things up."
The simple fact is that, while Diebold does indeed care about producing accurate voting results, they are more concerned with making money. If Diebold is forced to choose between increasing their profit and making the system better, they'll choose profit.
If you put voting machines in the hands of the private sector, the private sector will try to maximize profit. Corners will be cut. There simply isn't any way to avoid this, so long as the people making the machines are doing so to make money off the venture.
So long as the design and development of voting systems is left to the private sector, voters will be disenfranchised for the sake of profit. That's all there is to it.
Obliteracy: Words with explosions
Did India outsource its voting machines? Seems like maybe it's not just a matter of incompetent programmers. Maybe e-voting is actually hard to accomplish.
Complication does not equal sophistication. Sometimes, a number 2 lead pencil really does work best.
I think it's disgraceful that people forfeited their right to vote because of "foul-ups" Let's just hope that, if they plan to use these machines in the presidential election, that all the bugs would have been ironed out.
The Erogenous Zone
See, the sick part about all of this is that nothing will actually happen. Diebold will stall and complain and fling their influence around, The Governator will promise to look into it and do nothing.
"The general election is too close to fix anything now! If ONLY we'd learned about it sooner!"
Machine voting isn't the problem, Diebold is. They've created a horrible, insecure system. It's simple enough to create a more secure system that it's hard not to believe Diebold is deliberately enabling fraud.
A system where votes were printed to a machine-readable piece of paper, verified by the voter, then deposited in a secure box, would be simple and secure. By printing votes you create a self-verifying system -- voters can check their vote is correct, and an audit can easily verify that votes were recorded as voters intended. Management of the printed records would be just like the ballots we already are using, but without the reliability problems of punch-card systems. Tallying could be done mechanically, as a barcode could accompany the printed text.
The whole system is very simple. Even if they just used an ATM style of security (printing to an internal paper log) they would be far superior to Diebold. But using logic is difficult in this case, because Diebold is clearly making absurd claims, and it's difficult to refute absurdity.
EVM 2003 is trying to create a complete open source voting system (not just machine). I wish them the best of luck. This is more than just philosophy about copyright and IP, it's the defense of democracy from those that want very much to take away even the slight accountability that currently exists. They've already made it into office with one fraudulent election (2000), and very possibly kept control of congress with another (2002, with many states being won with unverifiable votes that didn't match up with predicted results).
Personally I don't really care about glitches, crashes and other problems with the machines. What I do care about is the use of uncertified software and the fact that these companies are more or less getting away with it. It sets a bad precedent for the future. Who cares if a few voting machines get decertified if you get to rig an election as a result? Any use of uncertified software should bar that company from ever producing voting machines in the US again. Do we really have to wait until someone is caught rigging a major election before real efforts are undertaken to stop it?
For the same reason as with many other things in this country: States Rights.
This topic was widely discussed during the 2000 election, with lots of questioning about why there wasn't a standard mandated by the Federal govenment. But the elections are run by the municipalities, and not by the national government.
Once again, I have to ask - what is the big goddamned rush to get election results that requires electronic voting machine? Why are people so frickin' hard to get the results of an election, like, on election day.
People should just chill, let a team of little old ladies count PAPER BALLOTS marked in PENCIL or PEN, and get the VERIFIABLE RESULTS a week or so later.
Mod down people who tell people how to mod in their sigs
Also, while I feel bad for the folks who are having their ability to vote without assistance taken away from them, it has to be better than having broken machines at which no voting at all can take place.
Your mileage may vary, but mine is constant.
Allin all, how difficult would this --really-- be? At least getting the part right about who's allowed and who's not allowed to vote? I'm a programmer, I've studied cryptography, I understand the problems associated with voting, but what if they made an open system, hired good programmers, and hired other good programmers to check the first programmers work, without having a private company do the work. (or at least force the private company be open).
.... test in in some *local* elections for a few years, and when those work, start moving it up to larger (ie: statewide) elections ....
Lave the code open, let people look at it themselves, fin problems or what not
Jesus, people have created some insane stuff back in the day, what's the problem now?
-
Nothing about apologizing for the problems with the product, or the fact that they didn't work. He appologizes for getting caught.
Which speaks volumes about Diebold as a company. Using the phrase "We were caught" implies they willfully put the bad machines out, etc. Having the head of the company say this makes it very hard for even the most forgiving of souls to trust them.OMG this is so stupid. How fucking hard can it be to make a distributed database and client software to increment a couple hundred variables? You don't see these kind of problems with web polls. Sure there are different candidates in different elections, but if you split the problem along the lines of particular races, other than the shiny happy interface to use a snazzy hi-tech touch screen to touch the face of your fav candidate, it really comes down to a VERY simple database call to increment a count for each candidate. I could do this whole system with one single database table for results.
Seriously, I could put together a system to increment a couple hundred variables in non-realtime over a weekend reliably enough to stake my life on it's accuracy. Most of us could.
The classic example of stable, solid transactions would be the ATM machine. Maybe if some of the rocket scientists at Diebold would talk to someone who had actually worked on an ATM machine they would know how addition works in computers, even across globally distributes databases, and learn how to do things like paper receipts, reliable atomic transactions...
Huh? Diebold already makes ATMs? Oh..
But what's so hard about e-voting, again?
I can't imagine that the computer end is so difficult. The code cannot be that complex. You need a numbered menu, human input device, some switch statements, and increment counter. Maybe check for buffer overflows, etc. Save the results with high-grade encyption that requires a password to access. Give the voter a receipt with a printed confirmation hash for verification. Plus, computer hardware is such commodity items that there should be sufficiently good drivers.
Seems to me that humans are demonstrably capable of mucking things up all on their own -- even with paper ballots.
To-do List: Receive telemarketing call during a tornado warning. Check.
Spell Check
>Today, California officials may recommend decertifying some or all of Dielbold's machines for the November General Election.
Sadly, this will include the Diebold optical scanners used in my county. Like much associated with this issue, this would be JPFN. The optically scanned ballots are much like the machine scored tests used in university classes everywhere. You fill in a bubble with a black felt pen to vote for a candidate. Simple, quick, readable with either the optical scanner or the Mark I eyeball in the event of a power failure.I am totally at a loss to understand this rush to electronic voting. As a citizen, I demand that my vote be:
- Secret
- Subject to verifiable recount
- Free from fraud
I realize that these are the ideal and that abuses have occurred under all forms of balloting yet used. However, the paper ballot and voting lists have stood the test of time. Reducing costs is not be a valid reason for mucking about with the very foundation of the democratic process.Some mornings it's hardly worth chewing through the restraints to get out of bed.
...Oh, wait... ...these machine don't provide a means for a recount, do they?
Never mind
"How to Do Nothing," kids activities, back in print!
Yeah, Texas doesn't have the greatest record of fair voting, either. Just google "Ballot Box 13" and texas (add an LBJ in there for good measure) ;)
You'd think that there'd be some impetus towards a minimal standard.
Returned Peace Corps IT Volunteer
San Mateo County went to mark-sense machines years ago, and has had very little trouble. The ballot boxes consist of a lid with a scanner locked to a big plastic bin, so every ballot scanned is locked inside the ballot box should a recount be necessary. At the end of the election, the scanners are plugged into a phone line and transmit results to election HQ. They can be re-read later, and the ballots counted and matched against the scans if necessary, one ballot box at a time.
Other than generating huge amounts of paper, there seem to be few problems with this.
The problem isn't directly with Debold (although they definitely are culpable).
The fault lay with the requirements produced by California (and any other state trying E-Voting). That is to say, none. They just said "Give us E-Voting, whatever that means"
The agencies should NOT be allowing the machine producers to define the voting method. They will invariably produce a mechanism that maximizes profit potential. How do you test the machines once you get them. You don't even understand the process since you didn't develop it. As a developer I can tell you NEVER let development produce the requirements. We miss everything and when you think you found a bug we just say "it works as designed".
The various election agencies need to come up with a definitive set of requirements for what an E-Voting machine should do. The level of detail should be excruciating.
The agencies also need to define and publish policy and procedure around these devices as well. You don't actually need to devices to do this. If they are built to your requirements then the procedures can be followed.
The kinds of requirements need to cover things like:
A paper receipt must be produced by the voting machine with human and machine readable type. If the machine readable type is not the same as the human readable type, the code produced must not be unique per voter or voter session (i.e. I can't transcribe the code and use it to prove who I voted for or you cant prove who I voted for)
The executing code must be certified (Open or not) and must then be cryptographically signed. The certified cryptographic checksum must be published 30 days before the election and each voting machine must display the checksum at all times during operation in a place that is visible to voters (i.e. I can write down the checksum and verify that the machine I'm using is using the correct version of the software)
When setting up a voting area each machine must be checked for the proper software checksum. (potentially a matching of software checksum and hardware specification, a use for Trusted Computing perhaps?)
Each machine must be able to produce test ballots for every candidate and the test ballots must be accepted by the designated reader machine. The test ballots will be conspicuously marked in a human and machine readable way. The reader will display the candidate indicated on the test ballot when reading (could be a screen, 7-seg display code, whatever).
Lots more, in much more detail that I went into...
=Shreak
There were no "screw ups" with the paper ballots in Florida.
The Republicans descended on the state in waves, filed gobs of lawsuits, and fired PR bullshit cannons at the TV talking heads.
The "problem" was the hysterical Republicans at the counting tables. At the doors, rioting. At the courts, delaying. On the TV, bloviating and installing doubt into people's heads.In the Supreme Court, concocting a BS, only-this-one-time decision to install their boy.
The recount of the paper ballots was going fine, no hitch, no fuss, until the U.S. Supreme Court stopped the process. They held up the decision until 30 minutes before the deadline they themselves set up as sacrosanct, then invited the state to finish up in that half hour.
It hascome to light that if Gore had won the initial count, those armies of drones were instructed to demand recounts until our anal sphincters bled. And to question the legality of Gore's presidency for the next four years. The cable channels would have been a non-stop illegaltheftoftheelection telethon for the Republicans until the people would have impeached Gore just to shut the right wing the hell UP already.
And, oh yeah: a recount was made, afterthe election,by a consortium of newspapers and the parties themselves. If the votes had been recounted, with the overvotes added in (candidate checked off, and the name written in as well -- a common practice, apparently) rather than discarded --
Gore won.
Finito.
What ever happened to good, old fashioned testing? I've seen the problem with companies rolling out software into production before it has been fully tested and ended up paying the price. I've had to clean up the mess of other engineers who didn't test something and told them about it every time. I asked if they tested it. They answer "No, it should work. It always has before." When I ask if they are always 100% confident that nothing was missed, they say yes, but obviously this isn't the case. When it comes to something as important as an election, in my opinion, there is no excuse not to test, fix problems, repeat ad infinitum, then roll it out once everyone is satisfied there are no errors. If this takes 20 years, fine. Just make sure it works correctly before rolling it out.
But why is the rum gone?
We're using eVoting machines here in Brazil for a couple years without major problems. It's better for many people here are iliterate and eVoting machines carry photos and speak the name of the candidate. I see many saying that paper can be trusted more and computers, but hey, we're talking about humans and voting here, at least in Brazil, nothing was more fraudable (does this word exists in english?) than paper, all it takes was a huge shotgun, a corrupt landlord with an easy tone: "mark here..." At least eVoting is making it harder here for fraud.
-- Por mais que eu ande no vale das trevas e da morte, meu PowerMac G4 Não Travará!!!
Technology wins our wars, so it must be flawless.
We have a permanent hard-on for technology. Just look at Slashdot, for crying out loud. We're a country that worships technology for its own sake.
If it's newer, sleeker, faster, shinier, and eliminates interaction with people, we're all for it. We want a permanent state of newness. We don't care about history, or precident, or any of that bullshit.
Ready, fire, aim! It's the American Way.
Read the EFF's Fair Use FAQ
So, Diebold gets off with a half-assed apology, sorry about yer democracy, Mate! My bad!
And nobody on the federal level is making a fuss because...hmm, now I wonder why?
And it'll probably just tool along all status quo-y until...what? Massive, undeniable fraud? Some kind of grassroots "Hack the Vote" movement?
I think it was Heinlien that said, "It may be rigged, but it's the only game in town."
So keep the pressure on, and hope it makes a difference before November.
(Where's my EFF renewal form...)
What were you expecting?
...from selling voting equipment in the State of California. They have violated the public trust. The ban should be in effect for a time not less than five years.
Start Running Better Polls
Security? So what! The softwar is shit ANYWAY!
People like to harp and harp and harp on how insecure the Diebold system is, and this is very important. But put that aside for the moment and look at where the actual problems have been: software crashes that prevent people from voting, software glitches that produce false data. I don't care how "secure" the system is; if it produces garbage it can be Fort Knox, and who cares! The whole issue of "security" while conceptually important for voting software is in a way irrelevant here until they can make software that produces accurate data while not being tampered with.
"Who are in control, they are not in control of anything - they don't even control themselves!" - Glen Beck
[...] they were instructed to charge HIGH to add [print] capability, if it came to it.
I don't see this as (necessarily) a huge conspiracy to avoid an audit trail. This is likely a simple matter of Diebold achieving vendor lock-in with a limited feature set and then charging outrageous prices to add to that feature set.
Another non-conspiracy explanation: Printers have lots of moving parts and are highly subject to failure. That's especially true if they are used intermittently, with months of storage between uses. Now deploy thousands of them, with the entire corps of news media breathing down your neck over every glitch, and you have an enormous PR disaster in the making.
Given that other voting systems that have been in service for years (such as mechanical voting machines) leave no audit trail, they might have thought that the risks from the printers were far worse than the risks from bad PR over the lack of an audit trail.
That's doubly true because there are a LOT of places - not just the machines themselves - where the count might be corrupted, maliciously or through bugs. Without the trail such corruption wouldn't be detected, with the trail it almost certainly would, and would become big news. Build in a scandal-magnet and you need extra bucks to weather the resulting scandals and achieve the same risk-reward ratio.
But I, for one, am VERY glad to see the conspiracy theories circulate - and that they're naming Republicans. Because IMHO they're likely to lead to increased pressure (especially among Democrats, who otherwise could care less) to add the audit trail - and thus make elections even harder to rig than with the previous, non-electronic systems.
Bantam Dominique roosters crow a four-note song. Once you've heard it as "Happy BIRTHday" you can't NOT hear it that way
The government did demand it, they were promised it, and Diebold lied about it.
Perhaps the government claimed they demanded a quality product, that doesn't mean they really did. If they had, then as soon as they discover the evidence to the contrary they will at least stop doing business with Diebold and at most sue Diebold for failing to live up to their claims and/or contracts. Have you ever bought a grossly faulty product and then continued to patronize the same company regularly afterward? The government does, all the time, and unless hell has frozen over they're going to continue doing so with the Diebold machines which will be filling voting booths in November.
The private sector doesn't work because private companies are miraculously more industrious and efficient than public agencies, it works because there are usually redundant companies supplying the same service, so selfish consumers avoid the companies who do a poor job of it, so selfish companies try to do a good enough job to keep some customers. If companies discover a big stupid customer with deep pockets who will be happy no matter what, how do you expect them to behave?
When you have two parties that have been around for a while in a two party system, it's *going* to approach 50-50. You can't avoid it. Both are tied to their bases and reach towards the center to attain the magic 50 point. If one party underperforms, they react by being less extreme and reaching towards the center more. If they reach too far, they lose their base. 50/50 means the parties know their constituency and the people know their parties.
Third parties try to flip the paradigm and appeal to large cross-sections, which honestly could work, but the problem is the two-party system of the Electoral College ensures that they either have to build enough support to actually win the election in one election, or they risk hurting the people by disenfranchising them away from the likelihood of being accurately represented. (Definition: if there's another candidate that the population prefers overall, on a head-to-head basis, to the candidate that actually won, then the population has been disenfranchised.)
The best way to ensure that there are better choices for voters are to remove the cost of having multiple candidates. This means removing the Electoral College, counting nationwide using a system like Condorcet, and optionally including per-state weightings to protect regional interests in the way the E.C. tries to (but performs horribly at).
skkkoooonnnggggkkk ptui
hummassa ----
How do you avoid losing the whole day's votes if the embedded floppy goes bad at the end of the day?