In 2000, I voted in Iredell County NC. We had cardstock quality paper ballots with a "fill in the circle" choice. Like you, we had sharpie markers, and if we screwed up our ballot, we could swap it for a fresh one. When we finished, we fed our ballot to the scanner.

The only question I heard anyone asking was "Which side goes up?" at the feeder. Cheap, simple, and easy to recount.

F'ing bureaucracy.

That is how it was in Destin, FL too.

I voted with a punchcard ballot in (very rural) Ohio in 2000. It was very easy. You could see how you had voted and watch your ballot drop into the bin.

Yesterday I voted in a huge electronic or mechanical booth in urban Ohio. It was like a gigantic vertical board with the candidates and issues overlayed on it. You make your selections, and a light indicates how you are voting. When you're done you push a button that says "vote" and a huge thunk (presumably) records your selections on a ballot. The pros are that you can easily see how you're voting and make changes if you want. The drawback is that you have no idea if your vote was recorded properly, and never see your ballot after the poll worker loads it into the machine.

Personally, I like the method of the machine better, but would really like to hold that filled out ballot in my hand before dropping it into the box. Making this such a big deal is counterproductive. It only makes it more complicated and expensive than it needs to be.

"Bay County is to Palm Beach County was the Taj Mahal is to a run-down Motel 6"

I'm pretty certain you mean that analogy the other way around.

DOH! Thanks, Nathan...

Cast my first vote for Ike in 1956 and since then have lived in various parts of the country and used various voting machines. They were all pretty simple to manage and we felt that our vote would be counted correctly and whoever got the most votes would win the election.

All that's out the window. Anyone who has even heard the word computer knows that 12 years olds can hack into the most tightly guarded systems and wreak havoc upon them. So it is even less than child's work to manipulate the outcomes using electronic means.

It's open secret that in the last Louisiana senatorial election the left benefited mightily using these high tech machines.

No backup except a tape! That'll be a big help.

In my precinct, we use the scanner paper ballot method, and I like it a lot. The machine *does* spit out inproper ballots on the spot, so you can correct it. And the paper ballots go straight into a *locked*, heaavy steel strongbox.

If you have confusing, mistake-prone systems, then you need lots and lots of government staff to keep it a-runnin'.

Conspiracy solved. ;-)

Here in Oklahoma, we've used 'complete the arrow' for as long as I've payed attention to ballots. Which is about 10 years longer than I've been voting.

Chicago (which, for you tinfoil types, was the home of Al Gore Jr's campaign manager) used the punch card ballots. The machine would theoretically kick out any spoiled (hanging or dimpled chad) but in 20 years as an election judge my wife never encountered one.

We were utterly mystified by the huge number of spoiled ballots (aside from the fact that they weren't being discarded) until I read an allegation that people were trying to punch two ballots at once.

Will, I don't often agree with the LA times, but I wouldn't put this in the tinfoil hat category. This isn't hacking the local university (what I nearly went to jail for) but its not rocket science either. Diebold used a combination of software that is known to be weak, and there are no checks in place to detect software changes that would signal a torjan. Compare with slot machines in vegas, which do have such controls, and still get hacked every now and then.

We use the paper ballot with black marker pens here in my town in Rhode Island (actually, I think the whole state uses this method) and it is a good and reliable system. After scanning the paper ballots are automatically stored in a locked compartment of the machine and in case of any question about the accuracy of the totals (or in case of a recount) they can simply be counted. (I'm sure even a resident of Florida could manage to vote successfully with this system.)

From my brother in Leon County FL, I got to see the failed votes in their version of the complete the arrow. (The clerk put a PDF of the ballots up on the web)
Many people circled the area around the two parts of the arrow, completely missing the space the reader is looking for the mark.

There were more problems up in Gadsden County (who are poorer than Bay County) as they only had the ballot reader at the clerks office and so the spoiled ballots were not rejected until after they had been collected from the polling precincts.

Given the enormous shock this country got in 2000, seeing live! twenty-four-seven! up-close-and-personal! how county canvassing boards really operate, your characterization of the possibility of corrupted electronic voting machines as "tinfoil-hat loony" is, shall we say, a more difficult sell than you seem to think.

I've now had the joys of both the complete-the-arrow system and the butterfly-ballot punch-card system (modified in SoCal to the InkaVote system--blah).

You're right, the arrow thing is really easy. And it is familiar to some extent to anyone who has been in school sometime in the last 50 years or so. It is very similar to the standardized test forms, where you have to put a mark in the right circle. My high school (20+ years ago now) had it's own optical reader for those types of tests--proven technology. It must be realitively cheap if a high school can afford to have the equipment. It also allows for verification.

I had the butterfly ballot for the first time late in 2002. I was astonished how bad it is. You have to trust you got the stupid punch card in the right location, and then trust that you got the punch in the right hole. No sign on the ballot telling you which hole belongs to which candidate.

You can't verify software, in principle. With enough incentive, and there certainly is that here, it can be made to do something that even the source code says it does not do.

The details are here Ken Thompson Turing Lecture

The short version is that you need not only the software but the complete history of every piece of software ever used to compile and assemble it, compilers, assemblers, loaders, down to microcode, which you will not have. That's not even worrying about deciding that what you do see operates correctly.

The thing to think about is not technical diffculty but the incentive to cheat.

There seems to be a lot of misinformation on the electronic voting machines. They cannot be hacked by a twelve year old because they are not on a network. The twelve year old would have to be in the booth and then there is no way to access the code. And, the code is quite simple. It is not difficult to add one to a count field. The editing to make sure a voter doesn't cast more than one vote in a race is a bit more complicated, but not much.

The danger is from an insider. But Chicago proves that voting machines could be manipulated by insiders. And, excpt for LBJ in Duval county, it is difficult to inflate a vote total. You could, however, drop some votes for a disfavored candidate and call it an undervote. But, it would take a conspiracy, not just one guy.

Rick

Just to clarify, I don't think security concerns about the new ballot machines are loony; I think those concerns are justified and well-taken, particularly in that there's no paper audit trail.

That said, the conspiracy theories about Diebold purposefully programming the machines to produce Republican electoral victories are as nutty as your average Kucinich voter...

Haven't waded through all the comments yet, but I would like to point out some information that may not have been stated yet. If it has, then please disregard.

Computerized accounting systems do not leave a paper trial in the same fashion as a traditional paper accounting system would. One of my electives was in audit/securtiy in accounting information systems (AIS). With the advent of computers entering the business world, there have been many cases of ingenious fraud. Eventually, everything come to light. And, as it did come to light, auditors and software designors have develop the skills necessary to minimize/detect fraud.

Doesn't mean they catch it all the time or in a timely basis.

With that in mind, how insecure are the new electronic voting machines? Of course, I haven't done any studying of the new machines, but when I voted on one on Tuesday, I felt as confident that my choices were recorded correctly as I felt the system was easy to use.

Personally, I'd love to see some sort of comparison of how hard/easy it is to mess with the new computerized voting machines, verses simple pieces of paper.

I would think "loosing" a box of paper ballots, or screwing up paper ballots, or filling out two at once or something similar would be a whole lot easier than hacking into a non-networked computerized voting machine.

Having said that, we use scantron forms that look remarkable similar to the forms we used in high school and college. Just fill in the appropriate bubble, then feed it into the big machine. It tells you right there if it read it okay, and if not, spits it back out.

In Minnesota, I think pretty much everyone uses "complete the arrow" ballots. They're nearly foolproof; almost any problem with the filling-in gets it kicked back, and the judge can explain the process to even the most addled voters.

In a web design class I was teaching in 2000, I used the Florida butterfly ballot for a practice usability and design analysis exercise. The students - most of whom had never done this sort of exercise before - tore the thing apart. It violates SO many key principles of human/system interaction and perception, it's amazing.

There's a big effort to bring decent user interaction design process to electronic voting - both the voting and the user verification. Lots of the problems they're having trace to the shortcuts they took with that effort so far...

Amy, When talking about how hard it is to hack something there is an informal scale that runs from 12yr old script kiddie to NSA/CIA. In this case the info available suggests mafia/small neutral country intel agency as dificulty level (the pakistani ISI could probably do it technically, but they're being watched way to closely to get away with it) Basically several guys working full time with enough money for decent equipment, call it $10 million and 1 year. E voting appears to be resistant to small time fraud, but once the system is owned, its completely broken, and quite possibly with no one knowning better (when the dead vote, people have noticed, when computers vote they don't currently leave an audit trail)