Just Say no
TO e voting
With just a week to go before the US presidential election, academics, politicians, and voters are voicing increased distrust of the electronic voting machines that will be used to cast ballots. In early balloting in West Virginia, Texas, and Tennessee, voters using e-voting machines made by Nebraska-based Election Systems & …
Let's get the paranoia out of the way first. This doesn't sound like a fraud to me - after all if you wanted to rig an election, would you let the voter know that you were doing it?
It sounds more like an HCI issue. Without seeing the machines it's hard to say what could be going wrong, but I'm going to guess... if these are touch screen machines either they've never been calibrated or the calibration has drifted so much that presses are being misregistered. Alternatively it it is an ATM style machine with hardware buttons, it's possible the layout of the on-screen ballot is just lousy.
Sorry Tuija, try to test it more again - maybe it turns right some day? But we Finns are stubborn, we will test and test and test it until we get the correct result or there has been so many tests that the number of incorrect results is statistically insignificant and can be ignored - problem solved!
You have to click on your choice, then wait about a full second or more, meanwhile staring at it until you see that it did the right thing. _Then_ you can safely move on to the next voting item.
That's what I eventually came to understand after voting on these beasts two or three times.
Luckily, I'm a career software tester, so I can figure these things out and actually vote. But your mom is hosed.
Just shows how criminally inefficient and/or corrupted the voting system is in a USA that boasts of freedom, and prides itself on technical achievement. We have giant profit-making cellular and Internet businesses, space-shuttles (sometimes), submarines, anti-satellite missiles, and yet our fumbling leaders can not guarantee an honest election. They (the government) have allowed crooks to sell faulty and easily-hacked voting 'machines' to local government, and have allowed the continued fractured national "vote".
I presume the point is that the researchers violated obvious and routine safety checks - so that in order to swap the chip, they'd have to break a seal, and the seals are checked before the machines are used. You have to presume *some* level of physical custody and competence - the same with 'legacy' voting machines, too.
The real way to manipulate the vote is to hypnotize the 60-something bluehairs that invariably run the polling places!
Voting 101:
1 x Voting ballot paper (yes, paper)
1 x Pen or pencil
1 x Booth
Method:
Voter gets 1 ballot paper each
Voter name registered as having voted
Voter marks ballot paper appropriately
Voted folds said paper and drops in sealed box
Sealed boxes opened later
Ballot papers are counted while being scrutinised
Candidates with the most votes win
Result
Too easy.
As with the till fiddling (restaurant owners change totals on electronic tills to avoid tax):
http://www.nytimes.com/2008/08/30/technology/30zapper.html
So it comes down to the voter verified paper trail, the till roll printed in the till, a copy of which is handed to the customer.
But also the rest of the tricks:
http://www.gregpalast.com/block-the-vote/
I personally reckon that GOP won't try to get McCain elected by fraud, however there are other elections going on at the same time, don't forget the states that are electing Senators, Congressmen etc.
How fucking hard is it? HOW HARD CAN IT POSSIBLY BE?
I make interfacii all shitting day and they talk to a database. Now, these guys screw it up when they have an entirely closed system, known hardware and *their own choice of input mechanism*.
But no, choosing one thing out of a goddamn list is prone to "flopping"?
Someone should be shot, and then shot again.
And @AC re:Pencil + Paper.
Using paper opens it up to editing. Using boxes opens it up to fuckup. See "hanging chad" fiasco. If someone can't push a button or punch a hole I failrly doubt they can see the difference between one empty tick box and another.
AC becuase I spent too long typing at work :/
If the video of the 'vote flipping' was anything to go by, this seems to be simple user error.
Due to a little bit of parallax the user was pushing what (to them) was the correct button, but a side view showed that the tip of their finger was actually hitting the button above first. Thus 'flipping' the vote.
A bit of GUI redesign (and possibly a bigger screen to host it) is what's needed, the number of candidates available seems to have ended up making the selection buttons too small, certainly smaller than I'd tend to use for a general purpose touch interface.
It doesn't look like there was anything deliberate or malicious in it, just an implementation cock up. The same happens with paper too.
Personally I prefer the old fashioned 'mark the box' method, not perfect but better than most of the alternatives!
Take a piece of paper, use a pencil, fill in the bubbles next to photgraphs and names of the candidates and/or parties.
Modern systems should be able to handle checks, dots, etc...
Place a stack of cards in the voting booth and allow print a big sign which says "if you make a mistake, toss the card and start over on a new one. If this is too difficult, ask for assistance".
Allow a maximum of one card to be inserted into the machine per voter... I.E. the booth will have to be manually reset by an adminsitrator between voters.
Now we have resolved the problem. If using this system people still can't get their vote right, then we might be able to agree that they wouldn't get it right even if someone were to specifically interview them and do the work for them.
Best of all, there's a paper trail and there's electronic counting. The electronic counting can run real time as the votes are cast. Then at the end of the day, the cards can be automatically fed.. MULTIPLE TIMES into a counting machine to count, recount, and RE-RECOUNT every vote for final verification. There should be the EXACT SAME NUMBER of votes every count. If there isn't, then someone should be arrested for malicious tampering or stupidity.
When it's ALL OVER, after 10 years of physical storage and safeguarding, the paper can be recycled.
To start with, if the complainers come from West Virginia, I think we can draw some conclusions already (PEBKAC...)
Secondly, the big question is not whether e-voting should be used but how much we're willing to pay to NOT use it.
Bear in mind the time delays (Media issue) and expense (manual counting) of the pencil-and-paper process.
Ultimately, most 'anti-evoting machines' stories seem to rely on someone going to inordinate lengths to either change a vote or eavesdrop on a voter. Neither strikes me as a serious issue other than in an academic environment.
As Georg W proved, there are easier ways to steal an electrion.
because its not that simple as in a lot places they vote for ;
President,Senator,Members of Congress,Governor,state senator,state reps,Mayor,councillor,
judges,police chief,school board,dog catcher etc etc
all at the same time - lots of bits of paper and lots a ballot boxes and lots of stuff ups, or a huge sheet of paper and lots of stuff ups - or a couple of bits of paper (feds, state & local) and lots of stuff ups
To put it in terms a pommy could understand it would be like voting for the queen, prime minister, local commons member , couple of members of the lords, mayor and councilors, queen of scotland, head of the scottish assembly, scottish assemblyman, chief constable, head of local NHS trust, and local dogcatcher all at the same time.
As my dear old dad you say "OTHER people are stupid, if you remember that the world makes sense"
Don't use them! The problem is humans are not necessarily that keen to vote with a machine. Many older people are not comfortable using even a touch screen and may even not trust a machine and many younger people will go into the booth and not read/follow the instructions.
Voting 101 is correct it's simple and providing the counters are doing a decent job it's going to be pretty accurate.
Since these machines are intended for Democracies, perhaps they should have a vote on whether to use voting machines or not.
They waited until a week beforehand to start raising concerns?!??!!!!ON1E!!!
Maybe the should have raised concerns beforehand, or didn't they realise that there is an election on? Over here in the UK, we've hardly heard about it, so maybe they haven't either...
(Please just make it end, so we can get the recriminations out of the way and get about three years of peace, before the next one...)
Like most dittoheads, you suffer from premature posting. Next time, try to wait 10 or 15 seconds before your, er, emissions.
The truth is simpler. You need to get an imagined human fault into the front pages of the press and wedged into the public's dim consciousness before election day. When the exit polling is again 2σ removed from the "counted" vote, you have your reason to jump up and point - Remember! Regisimastration errors! Dumb Voters! Regisimastration errors!
Just try not to be still foaming when you blurt that out next Tuesday.
Nice touch with the real fake name, though. Show you read all the way to the bottom of Rush's manifesto. Unless you simply fell for an earlier dittohead - proving that it works.
You say:
> If someone can't push a button or punch a hole I failrly doubt they can see the difference between one empty tick box and another.
Such reverse-justification disqualifies you from future UI work. The job of any UI is to let the intended person do what they want/need, in a failsafe, obvious, easy way. And to design one, you therefore have to find out what those people know, like, understand, hate. And if you, AC, don't appreciate that, unlike you, significant swathes of the voting population hate and fear using electronic devices, and will get things wrong no matter how well they're designed by you (or, rather, someone other than you), but that these people are all still very happy indeed putting a cross in a box as they have all their lives, then you, sweetie, are part of the problem.
Stick to the database side of things, if you reckon you're talented at that (but do ask for a second opinion).
I've said it before and I'll say it again. Everyone trusts lotto machines. It's both paper and electronic based. You mark who you want to vote for the machine reads your vote and prints you a receipt and retains the original paper version. Simple. Pleanty traceability there.
The benifit of loto machines also allows for you to make a quick pick..... They're reliable too because people would get very upset if they didn't get their numbers.
Electronic voting machines can NEVER be made 100% trustworthy. It's impossible. Not just very difficult, actually impossible. Think "undoing a multiplication by zero"-type impossible.
Let's consider the process of a pencil-and-paper election in purely abstract terms.
Every voter has a unique token (their name and address on a piece of paper in front of the presiding officer). This is spoiled (name crossed off) and exchanged for one of many tokens (ballot paper) which are indistinguible from one another. The non-unique token is then marked by the voter to indicate their preference. There are several possible ways the non-unique token can be marked (boxes to put an X), but all tokens marked the same way are indistinguible from one another (all Xes are the same).
The process of exchanging a unique token for a non-unique one ensures that each voter can only cast a maximum of one ballot. And the indistinguibility of similarly-marked tokens makes it impossible to determine from a ballot already cast, who cast it. There is thus no way that the results can be subverted before the count.
With e-voting, the token exchange is done "invisibly" and it is impossible to verify the integrity of the process. So some people might be allowed to vote multiple times; and there is no way to detect this, let alone prevent it. It is impossible to verify the indistinguibility of the second token, so someone might know who you voted for; and there is no way to detect this, let alone prevent it. It is also impossible to verify the integrity of the count, so your vote might be recorded for another candidate; and there is no way to detect this, let alone prevent it.
Paper trails bolted onto e-voting do not help one iota. This is because, with e-voting, there are no physical tokens to be exchanged, marked, transferred and counted. Everything is done by copying, and the integrity of the copying process is unverifiable. And publishing the Source Code won't help either, because there's still no way for the voter to be certain that the Object Code running on the voting machine was derived from the published Source Code that they have checked and found OK. The possibility always exists for a vote to be undetectably subverted before the count.
So much for subversion *before* the count. There still remains the problem of subversion during or after the count. The only way that voters can be sure that their votes have been counted properly is for all parties involved in the count to have an adversarial relationship with full mutual distrust. If nobody trusts anybody else, the only way they can agree is on the truth.
(It's possible that a direct-recording, purely mechanical vote-counter might satisfy the requirements of a fair voting machine. A separate add-only counter with numbered wheels would be used for each candidate. The machine's housing would have to be transparent, with simple distorting lenses over the counters so that the voter could not see what the figures were; only how the machinery moved and that one, and only one, counter changed as their vote was cast. The machine would be primed by the presiding officer to accept a single vote, by means of a Bowden cable, as the voter entered the booth.)
The reason that these machines work so well is that in each case there is an adversarial relationship between the user and operator of the machine. Any bias is ultimately detrimental to the manufacturer, regardless of which party it appears initially to favour, and the only stable equilibrium is fairness.
* A one-arm bandit that pays out too much will return little profit, and so not be popular with amusement arcade / casino / fish and chip shop owners. A bandit that pays out too little will become unpopular with punters, and so return little profit for amusement arcade owners. Either way, the machine will most likely be replaced with one from a different manufacturer.
* A cash machine that pays out too much will be unpopular with the banks as they end up losing money. A cash machine that pays out too little will be unpopular with bank customers, who will complain to the banks who will then end up losing money in compensation and bribes to keep quiet.
* A lottery machine that routinely misrecords punters' numbers will end up costing the lottery organisers money, either in falsely-won prizes or in compensation for prizes that should have been won but weren't.
There is no such adversarial relationship between the organisers of an election and voters.
I hear there will be riots if he doesn't win.
Oh, and for the oblivious amongst you, the last broohaha occurred in a democrat precinct with a democrat running the voting booths, who designed the allegedly "hard to read" ballot and compounded the mess by failing to follow the procedures IBM had delivered some decades ago which clearly stated the machines needed to have the chads from the previous ballot removed to ensure proper functioning for the current ballot.
yes, we vote here also for more than just a party in elections,
OK, so we're not voting for a massive amount of people, but there is no reason you can't have multiple boxes on a single sheet of paper.
president
[ ]1
[ ]2
senator
[ ]1
[ ]2
[ ]3
it's hardly rocket science...
that said, that's not the point.
Electronic voting should make voting/auditing easier. so it really sounds like what is really needed is better UI design
like push buttons instead of
[push here]1
[push here]2
a bigger space
[push here] 1
[push here] 2
and frankly if you don't have to co-ordination to push a simple part of a screen inches away from another part of a screen...
on the other hand if it's hardware/software error then still better design is needed, really a kid of 15 could probably design a voting system with a few buttons that populated a database for their high school work.
give them a touch screen monitor and I'm sure that they could even make it work without need for a mouse keyboard as well.
I mean, how stupid can people be? Really. Both the ones who design the system and those who use it. Now it's that time when we pity the Americans. And you see, it can not be lack of technical ability on their part, so...
Contrast this with elections in Brazil. We just had some 120 million people or so go vote last Sunday for the municipal elections -- I wasn't there, but I've seen this before. We have been using electronic only, in the whole country, same machines everywhere, for more than 15 years. Nobody complains of anything being difficult, because it isn't -- unless you are retarded. And you can imagine how much worse is our literacy there compared to the US or Europe, right? A freaking picture of the candidate (and we have lots of them), name, etc. and you pick the one you want. Another screen asks (written, of course): you picked this guy (with picture), confirm? You do (or don't). And that's it, goes to next thing. Does not sound too different from what I hear the system here is, it seems. Older people take a little longer to do it, maybe a couple minutes, but nothing too dramatic. The time it takes has not been extended compared to the old days when it was paper-based. Actually, there are never significant lines to vote, and I don't remember ever seeing complaints of people not being able to vote. Contrast that with the US again.
Before going to sleep on Sunday night, we already had some 99% of the votes counted -- some machines have to travel by boat in the jungle or other remote places to get to the counting centers, so those can take a few hours longer to get counted. There has never been any suspicion of fraud -- and we have such a large amount of parties there, it's not even remotely funny, so it would be impossible to get everyone to play along. OK, so it might be that fraud is undetectable? I doubt it. The election results have never really been "surprising", like they sometimes are in the USA. Incumbents/their favorite have won and lost in different places as predicted by the polls, and results are very seldom so close that a little fraud/error would change the results.
So, as someone said above, if there is fraud it is subtle, as much as it was in the times of paper 20 years ago -- when sometimes counters would carry tiny "pen-rings" or whatever to either invalidate votes they didn't like, or create a vote when there was a blank ballot, etc.
Anyway, here in the US people don't even get to vote for president, I don't know why all the "democracy" talk...
I started out voting (many moons ago) on a large paper ballot that required the use of a round, inked stamper to mark my selections. this worked, but the paper and space required was cumbersome, to say the least.
This was replaced with the punch-card system that was small, easy-to-use and I never had a problem with "hanging chads" as I punched my card, pulled it from the machine to place in ballot box and always looked at it first.
I moved some years later to a new voting district which used a lever machine where you had to flip levers for each possible issue/candidate and then pull a master lever which recorded your vote. I was not fond of this machine as a means of voting as I could not see the end result.
That same district finally went to the punch card and I was happy again with a small, easy-to-use system where I could verify all my choices (and no, without seeing the count; I couldn't verify that my vote was actually tallied at all...on any of these...get over it).
I moved again and went to another district where we used punch card for years and then the new "voting machines" with the touch screen. Besides all the basic problems with security and lack of verification of vote (even after they added a paper trail), I don't trust these devices.
I have always preferred the punch card system since first using it and don't honestly see why there would be a problem with "chads" if the cards were checked and verified before being dropped into the ballot box...a very, very simple thing to do that takes only seconds and will result in none of the election nonsense possible with the other devices.
OK, try this.
Candidate A gets 380 votes. Candidate B gets 500 votes. Candidate C gets 120 votes. Those are the actual results. Every voter gets a slip of paper to take away with them showing how they voted (and possibly to exchange for favours from one of the candidates, which is in and of itself an argument against voter receipts). Every voter's name, address and vote are also recorded in a book which is kept at the Town Hall, because voter anonymity has to be sacrificed to ensure the integrity of the election.
The authorities announce 480 votes for candidate A, 420 for candidate B and 100 for candidate C. Candidate A takes up office.
A few voters think something is amiss, so they go to the town hall with their receipts (which show correctly how they voted). Everyone shows their ID card and gets to look at their vote in the book. It matches their receipt. They leave, unable to draw any conclusion except that the election was conducted fairly.
No ordinary person is allowed access to the whole book, so they can't check that the numbers add up. The "redacted" book -- which omits the names and addresses -- has been subtly doctored so as to add up to the announced figures. And enough people will have lost or deliberately destroyed their receipts that you won't be able to do a full recount based on receipts.
The *only* way to get a fair count is for the counting to be done completely in the open, under close scrutiny of the candidates themselves. Audit trails can be spoofed at every level; they are totally meaningless at best, and downright dangerous at worst.
The bloody election has been going on for almost 2 years, whoever is elected will be in office for 2, 4 or 6 years, and yet they can't wait a couple of hours for the result?
The polls in Virginia will close at 7PM, and the TV networks will predict the result within 30 minutes if the numbers look right, even though people on the West Coast will still have 4 or 5 hours before the polls close!
Phew, being in the technology field, it's nice to know that there's one company in the world, where having had tested something MEANS it will work, has the integrity to acknowledge the infallibility of machines... created by man, those bastards. I mean it's not the company's fault it doesn't work for some people, maybe it's the people who are screwing it up, damn traitors! Throw them in the gulag! Oh wait this is the US, torture them and kill them and their family so we don't have to pay anyone restitution for our humanitarianism of freeing them from the shackles of their miserable lives!
Or were the machines really created by Brawndo!, the thirst mutilator? It's got what voter's crave, shiny flashy beeping things with virtual everything, weeee! FPS voter death match in 2012!
Voter machine error only proves one thing, democracy is the problem. Stupid people wanting to vote. Where as I, a Genius (I'm not a paranoid schizophrenic! Everyone is out to get me and I need to kill them first), should just rule the world. I mean, right? So what if in order to do so millions of people have to be killed in wars of the most importance, like getting me more power so I can rule the world. That's why we classify everyone into little subgroups right, to make it easier to punch someone in the face without feeling any sympathy, now imagine them with a gun in their hand, now that's fun! And if anyone we know or care about gets killed accidentally during all this, it obviously because they must have done something bad in their lives, like poor Hogan's comatose friend.
To summarize, it's people who aren't working right, not machines.
Well, glad I was able to clear all that up.
As Boss Tweed used to say: "Doesn't matter who they vote for.... as long as I get to count the votes"
If these tests are a reasonable facsimile of accurate, John McCain is going to be the president, despite being something like 6 points adrift. The reason for saying that is that, to date, every test has the democrat vote going to the Republican; Jim Crow rides again.
(Jim Crow is a nickname for a whole raft of fairly nasty American legislation aimed at blacks. The intention was complete segregation of the races to the benefit of the whites who got the best of everything including, but not limited to, where they lived, worked, played, voted - whites only, whether they could be unionised and so on. Among the least awful parts of this were laws barring blacks from buying homes - or remaining after dark - in white neighborhoods, from shopping at white stores, even from working at certain trades. In support of this whites murdered, firebombed and lynched.
Oddly enough this kind of political violence aimed at a minority isn't called terrorism, but it should be.)