Asymmetrical Information

It is obvious absurdity. You can see my posts on the subject. The same set of Democrats think Bush "lied" when Bush was saying the same set of phrases that every other countries intelligence was saying, Kerry was saying, Gore, Clinton, Daschle etc. etc. The same set of Democrats think that Bush could've prevented 9/11. The same set of Democrats think that Rumsfeld should've resigned because he was responsible for the prison scandal in some way.

BUT I think there is a HEALTHY dose of criticism that needs to be stated about computerized voting. I have stated them after Occam's Beard stated his. It's not about who "internal" to a manufacturing company of these voting machines will set up the "fix". It's about all of the external forces and the need for paper in our voting..... I believe. And I am and have been a senior computer person for 8 years (been in computers for 15).

Personally, I think fewer elections will be thrown by people than by the bugs that the lack of validation is permitting to prosper. It's statistically likely, IMHO, that by now at least one candidate has been 'chosen' by a bug rather than the electorate.

Speaking of which, I want to take advantage of this thread to press my own fanaticism, the need for paper backups and active random consistency checks of paper versus electrons. See E-Vote Reform Proposals So Far Too Weak, http://www.centristcoalition.com/blog/archives/000357.html

I probably betray my ignorance here, but aren't the WW-I vintage voting machines sort of clock-work? I mean, to rack up even one vote, doesn't somebody have to pull some sort of lever against the non-trivial force of a big spring? If so, then to fake up a whole bunch of spurious votes somebody would have to pull the level like a little old lady at a Las Vegas slot machine with a bucket of quarters.

That seems to me to be a different sort of effort from jiggering the electronics of a more modern machine. I mean, if every push of the Bush button increments not the "ones" bit register, but the "two" bit register, then the number of votes is doubled. Or more subtle, if that only happens every 16 seconds ...

Personally, I still like paper marked with ink.

Especially for "write in" votes.

One could easily fix it so that, for example, votes for Republicans simply wouldn't register; it's the same lever every time. Of course, this would be time consuming, and one might get caught, and there's a much more parsimonious explanation for Democratic dominance, which is that everyone I know is a Democrat.

This whole "right wing conspiracy" crap is muddying the waters, and making people think it is only tinfoil-hat wearing conspiracy nuts who are worried about e-voting.


The old "clockwork" mechanical voting machines are no comparison to PC based voting machines. They are a lot harder to tamper with than a piece of software, networking protocols, storage media, etc. And don't most of them store a paper tape with each vote recorded?


I wish there were more tech-oriented people here, because the devilk is in those details, not what party employees of Diebold belong to

Tech oriented people and less "blame", "accuse", "conspiracy theory" oriented folks who seem to tend to be to the left of John Kerry.

I think it's fine to make procedural objections to the voting machines -- say they're not a good way to implement electronic voting, or that we shouldn't have electronic voting. But I think the political objections to having Diebold make the machines are silly.

Lev,

That's kind of where I came in, by saying I didn't understand why e-voting problems (glitches and/or fraud) necessarily favored either party over the other. (My first post on the previous thread.)

Tampering with the results in the fashion you describe could as easily be done by one party as the other. Clearly we would need strong safeguards to address this concern. I would put this in the "glitch" category, in this case, a security glitch that allows tampering with originally valid results (as opposed to something intrinsic to e-voting that generates fraudulent results in the first place).

I just never understood the reasoning behind those who view e-voting/Diebold as inherently favoring Republicans, but I think I now have my answer.

It's completely ridiculous to say that only "nutjobs" want an audit trail on electronic voting, or are concerned about tampering.

Do I trust ATM's? Yes, and I've used them thousands of times with no problem. But I keep the paper receipt. Why is generating such a receipt in a voting machine a big deal? Why is it looney to want verification? What are you going to do without it when some machine reports ten thousand votes for Nader and two each for Bush and Kerry in a precinct with one thousand registered voters?

And while we're at it, could we note for the record that the standard libertarian crutch of reputation effects does not solve all problems of business ethics?

I have a brief discussion of eVoting from a technical perspective here.

Based only on what I've read in the report referenced at the link above, Diebold should be disqualified on technical incompetence alone.

If you thought the Diebold-is-owned-by-a-Republican conspiracy theory was OTT, see this.

It's completely ridiculous to say that only "nutjobs" want an audit trail on electronic voting, or are concerned about tampering.

Do I trust ATM's? Yes, and I've used them thousands of times with no problem. But I keep the paper receipt. Why is generating such a receipt in a voting machine a big deal? Why is it looney to want verification?

Bernard, you've sharpened my comments beyond recognition. As I said explicitly, I support the idea of an audit trail ("paper record" were the words I used), and hence do not consider those who want one "nutjobs." An audit trail is only prudent.

Perhaps I wasn't clear, but my "nutjob" characterization referred to those who see e-voting/Diebold as part of a Republican plot.

Audit trails, you bet. Hermetic security provisions, absolutely. Concerns about technical problems and competence, of course.

Corporate conspiracy to fix elections, pass the tinfoil.

Some of us aren't alleging any conspiracy to commit voter fraud. We are concerned, however, with the appearance of inpropriety that Mr. O'Dell's partisan activites creates. This guy isnt' just a run of the mill contributor, He's a Bush pioneer, and contrary to the claims made on the other thread, he has not resigned. He has simply suspended his political activites. The problem is one of appearance. It's entirely possible that Mr. O'Dell and Diebold can do things on the up and up. But his activities, combined with his statement that he will do anything in his power to deliver Ohio's electoral votes to Bush, do create at least the appearane of a conflict of interest. It's my position that the job of counting votes should be done in as non partisan a manner as possible. If something goes wrong in Ohio, for example, it simply provides ammo to the conspiracy theorists and creates unnecessary controversy and taints the electoral process. In a judicial proceeding, it is of course possible that a judge with a bias toward one of the parties can still effetively hear the case, but ethically, we would ask that judge to recuse himself based on the APPEARANCE of bias, not on any actual impriopriety that we know of.

Eamon,

Do you realize HOW MANY people would have to be in on the conspiracy for Mr. O'Dell's "interest" or conflict of interest to actually deliver Ohio's electoral votes via these machines? Softare developers, testers of the software, others including managers.

You really put yourself way out there. I have to commend youu on that. :)

The Florida judges who were appointed by Democrats didn't recuse themselves and neither did the New Jersey judges who allowed the Lautenberg switch out.

And why are we just talking about Diebold when there are other manufacturers including the ones who supplied LA county with thier error prone touch screens?

Pat, that is not my argument. I am not arguing any impropriety, what I am oppposed to is the APPEARANCE of bias created by O'Dell's partisan activities, including his language. I doubt very much that there is some grand conspiracy to steal electoral votes in Ohio or anywhere else, but the appearance is very bad and ought to be avoided, as it just gives ammo to those who would challenge any disputed election conducted with Diebold Machines.

As a tech oriented person who is very concerned about electronic voting, I'll go ahead and comment.

Despite being to the left end of the spectrum, I'm not actually concerned about the CEO being a Republican, for the exact reasons you mention ... he doesn't alone affect the code, etc. Though, it was a rather stupid move for him to declare himself "committed to delivering Ohio's electoral votes for the president". :-)

But given how critical voting is to our democracy, e-vote machines scare the sh*t out of me. I am more concerned about:

1) A rogue programmer. If you've ever worked as part of a large programming or engineering team, you know it's trivial to bury items that others never see.

2) Poor adherence to standards and testing regimes. There are already numerous exampes of e-voting companies patching software *after* a state's audit and verification procedure, but before the election. This makes it more likely that concern #1 could actually cause a problem.

3) Hacking. Almost no computer is hack-proof, and an analysis of parts of Diebold's code posted to the web a couple months ago found dozens of hackable flaws. One of the worst was with the remote vote collection system - anyone with a copy of Microsoft Access who knew the IP address of the collection center (which aggregates votes from polling places) could alter the vote database untraceably. No password required.

4) "Black Box" effect. On the face of it, it's patently ridiculous to have a non-examinable box where you press the vote button and at the end of the day it spits out a tally of total votes. Any flaw, malicious or not, can end up altering an election.

One election in 2000 resulted in one candidate getting 140,000+ votes ... in a district with fewer than 30,000 voters. It was a bug, not malicious. But the only reason the error was caught was that the vote total was totally implausible. If that bug had produced an erroneous but plausible result, we would never have known.

We have now had tens of thousands of elections held with similar, nonverifiable, e-voting machines. Are you prepared to argue that only that one bug happened, ever? I think it's extremely likely at this point that some official somewhere is in office because of a simple code bug.

There may be several. There may be dozens. There may be others who are in office as a result of malicious hacking. With the current system, there's absolutely no way to know. Ever. Period.

For the record, I don't care one whit which party that official is from - it's a travesty either way.

An auditable, verifiable voting system is an absolute prerequisite for a valid democracy, both practically and philosophically.

We're not ready for e-voting. At the bare minimum, existing machines should print verifiable paper receipts (that are kept by the machine, not by the voter).

dOccam's Beard:

I'm not worried about Diebold stealing an election, but, to help the tin-foil brigade, I'd offer the following pieces to fit into the conspiracy theory of choice -

1. People who manage elections do immoral or illegal things all the time. There's reasonable evidence that LBJ stole a 1948 election to Congress by stuffing a ballot box. Ed Rollins allegedly offering preachers money to not get out the vote. Kennedy's dad running a second candidate against his son with the same name as his son's opponent to split the vote. Kerry's almost certain coordination with MoveOn.org. Bush's almost certain coordination with those in SC who engaged in the vile whispering campaign. The "unfortunate" trimming of African-Americans off the voter roles in FL (because they were mistaken for felons). Daly's deliverance of Chicago to JFK. Really, all machine politics before 1968.

2. The only questions worth asking about cheating are: (1) how easy is it to do, (2) how easy is it to get caught, (3) what do I get if I get away with it, and (4) what do I get if I get caught? As I understand your point, there would be relatively little benefit and it would likely be sussed out. In response I'd say that I've never been that clear about how easy it is to prove election fraud - people try it all the time with very limited success. And this WH is about as good bet as any to depend upon for deliverence from punishment. Seriously, why hasn't Rumsfeld (whom I love) been fired? Why hasn't Cheney been fired? Why are they so damn secretive about EVERYTHING (recall that one of the first EOs that Bush signed made it harder to get FOIA access to Presidential materials in the future). I mean, if a more accountable President lets you trade arms for hostages and get away with it (welcome to Pardonville), surely a little election fraud would be pardonable.

But ultimately I think you're convincing on three points - (1) technically, it isn't clear to me how easy this would be to do secretly (I don't have the background), (2) the benefits seem uncertain (though I note that it would hardly be the first time a CEO and his minions ran a company into the ground for their own benefit), and (3) I don't know why the Republicans would have to be the ones to get away with it (I too would like to think we have the coder community).

(Jane, you've done it again, thank you.)
I still support the concept of electronic voting, just not Diebold's implementation of it, for reasons IdahoEv mentioned. I don't care if evoting machines are based on Intel 386 chips and Xenix operating system. I do need to see all the hardware and software specs released for public review, so a nonpartisan commission of folks like IdahoEv and Pat in CA can pick it apart.

We don't need voting machines that can play movies, animate tutorials, or offer surround sound. We need voting machines that can accurately, reliably, and demonstrably record and aggregate votes.

President Kennedy promised to get us to the moon in the 1960s. We succeeded; not by continuing to harness emerging technologies, but by building complex systems from then-stable technologies. I am convinced that if we take the same approach to e-voting, we can succeed there too.

Tomorrowist,

Believe it or not, I think it could be done also. But no matter if I pick it apart or not there would be these underlying concerns:
1) Is it networked? If yes, new security vulnerabilities are found on just about every application and operating system every month. No matter how many Pat's "pick it apart" initially there will be someone smart who can then be posed as a "janitor" and affect lots of systems across the nation and maybe even clean up the trail so that nobody knew it was done.
2) Will the things need to be replaced on a 4 year replacement cycled like the millions of PC's that are replaced every year? If it is on a 4 year replacement cycle it would be very very costly.

I would be FOR the following:
1) A simple system that has a simple "operating system" like a microwave. Simplicity would keep the costs down. It doesn't need to be fancy.
2) A system that is not networked and would print out a reciept maybe (I'm not sure about the receipt idea). But the system could have a little unit that detaches and then a driver could drive the unit to the precinct central area just like they do now with punch cards that get tallied.

Eamon, your post seemed plausible UNTIL I got to this line:

It's my position that the job of counting votes should be done in as non partisan a manner as possible.

Yes, that's true, but how does the CEO's rabid political leanings meaningfully factor into the debate? Giving conspiracy theorists (blank) ammunition isn't high on my list of concerns, since they'll happilly tilt cap pistols at windmills whether I help them along or not, and for the more rationally minded, "tainting the electoral process" is going to require more evidence than that.

There may be other valid reasons for distrusting the security of an electronic device (I can think of many, and other posters have listed some concerns specific to the Diebold units), but this isn't one of them.

Pat in CA,
I agree that e-voting machinery would need the simplicity of network appliances. And your fake janitor is a security concern.

But I suspect there would be a way to have detachable programmable computer cores which could be kept under tight security until just before the election where they would be programmed and transfered under seal to individual precincts for use in software-free voting machines. I am satified that such machines could be networked with PGP technology.

The fake janitor would be no more of a concern then than he is now; he cannot manipulate anything he has access to.

The keys are to have open architecture throughout, limited areas of vulnerability, and tight controls over vulnerable areas.

Right now the janitor can affect only those machines in that building not a whole network of machines.

I agree that e-voting machinery would need the simplicity of network appliances. And your fake janitor is a security concern. But I suspect there would be a way to have detachable programmable computer cores which could be kept under tight security until just before the election

Nothing electronic is simple. Even if the software could be verified to be safe, one would then need to verify that that OS and all of its modules was safe. And, one would need to verify that the microcode inside the CPU was safe.

Then, one would need to verify that the equipment used to verify the microcode inside the CPU was safe.

Then...

As for the detachable cores, what if the displays were altered to show Choice #1 and Choice #2, and vice versa?

And, while AFAIK no one has broken into Fort Knox I don't think these machines would be under such tight security. Someone somewhere could obtain nod-and-wink access to them and change the code. It might only affect that set of machines, but it could swing an election.

Just use a voter-verified auditing system. After the election, count a sample of the receipts; if there's a discrepancy with the electronic results, then all the receipts would be counted and that count would be the official tally.

"Choice #1 and Choice #2, and vice versa"

should read:

"Choice #1 as Choice #2, and vice versa"

But you'd probably already figured that out.

> "I mean, to rack up even one vote, doesn't somebody have to pull some sort of lever against the non-trivial force of a big spring? If so, then to fake up a whole bunch of spurious votes somebody would have to pull the level like a little old lady at a Las Vegas slot machine with a bucket of quarters."

> "The old "clockwork" mechanical voting machines are no comparison to PC based voting machines. They are a lot harder to tamper with than a piece of software"

Giggle.

One can design an honest mechanical vote counting machine with or without a reset mechanism (a way to change the counts without pushing a lever many times). The reset mechanism is vulnerable, and the lack of a reset mechanism has different vulnerabilities.

And, it's relatively easy to rig an election without hacking the machine at all, and without running any risk of being caught by a recount.

Technology can't turn an open-loop situation (aka "trust required") into a closed-loop one, and elections are fundamentially open-loop.

> As for the detachable cores, what if the displays were altered to show Choice #1 and Choice #2, and vice versa?

Uhh, if the display is either raster graphics or row-and-column text, there's really no way to do this. And how could a voting machine display be otherwise, given that the actual contents of the display change from one election to the next?

I find the overall discussion of e-voting very tiresome, as we (collectively) seem to want to make every possible mistake instead of of learning from experience or from careful thought. My biases are (a) the need for open, reviewable designs and (b) the need for paper trail. The second of these is admirably served by optical scanning voting machines, so why are other forms even under discussion?

Sadly, this conspiracy while serving to get Dems behind paper trail touchscreen voting has made it a more partisan affair when it should be all of our concern. By screaming about Bush stealing the election, these idiots are looking for a political weapon rather than concern for our voting process. Bev Harris is the exception to this though she is a Democrat.

I think there should be a paper trail. The ballot prints out, you stick it in a box in case of need of recount and that's it. It's not so hard.

Having grown up in and spending most of my adult life in Illinois, I am very aware of vote-stealing (or whatever you want to call it.) It is done by both parties for similar reasons: to keep control; to benefit current players; to keep out outsiders (party, individuals, doesn't matter), and so on. I am aware of vote rigging and stealing schemes on any system you want to list. It's gonna happen. Dukenfield's Law: It it's worth winning, it's worth cheating.

Usually, vote fraud is not punished very much because, in the US, almost all voting is controlled at the local (usually county) level. In other words, the same folks committing the fraud are in charge of the investigation, trial, and punishment of the fraud. Remember, all those protested hanging chad ballots were under the control of local Democratic officials, not state Republican officials, no matter what the Gore campaign tried to convince you of otherwise.

Vote fraud usually is only useful in relatively close elections because large-scale fraud is harder, is easier to detect, is more likely to be punished, etc. The evidence of fraud in Florida in 2000 was laughably low compared to places like Philadelphia, Detroit, and St. Louis.

Most vote fraud currently requires largish numbers of conspirators. It is frequently an open secret when it happens. Reason no one acts, see above. The problem with e-voting for me is that it greatly reduces the number of conspirators needed to be successful AND it makes it more useful/likely on a large scale than previously. (Even the large scale vote fraud sometimes still seen in Chicago is actually an aggregated set of small scale frauds, co-ordinated loosely at the top.)

Attention Democrats: You may have John Ashcroft as AG because he had a close election stolen from him in St. Louis (probability approaching 1).

Still, we should protect the vote as well as we can. Paper audit trails are essential (this from a guy who teaches Public Budgeting -- paper trail, paper trail, paper freakin' trail).

I would feel much more sanguine about all the arguments and the conspiracy theories if those involved were equally interested in going after current fraud in Democratic strongholds as they are in going after Dieboldt.

One of the worst was with the remote vote collection system - anyone with a copy of Microsoft Access who knew the IP address of the collection center (which aggregates votes from polling places) could alter the vote database untraceably. No password required.

It's worse than that. Microsoft Access is not secure in any sense of the word, and does not track modifications to the database; IIRC. So password or no password, the ability to acertain that the data has not been tampered with is 100% impossible.

Being a kneejerk contrarian is something I suffer from as well, so I understand this desire for everyone right-of-center to redeem the good folks at Diebold when wildeyed lefties start foaming at the mouth. And they are good folks no doubt; they just suck at building secure voting machines.

Or possibly the system is secure but Diebold just sucks at negotiating the political minefields involved in certification. I don't know what the standards are for certifying ATM's (their biggest electronic business), but if was a banker I wouldn't be too impressed by certification by some barely-competent government employee. If Diebold's reputation wasn't better than most government agencies, I doubt they'd sell any equipment to banks.

This is not to say that I trust the Accuvotes all that much, especially if they are networked. They're Windows-based, after all, and it's obvious that even Windows' creators no longer understand it, let alone a company whose design engineers used only 8051-class 8-bit microcontrollers up to 1998.

However, security depends mostly on how the election personnel handle the system. Networking is an option, not a necessity. They'll work just fine as isolated machines, with the ballot inserted on a PCMCIA card, and the totals collected by collecting all those cards at the end of the day. If they've got good systems in place to prevent meddling with paper ballots, the same system will work for handling the cards. If the human side of the system is corrupt, no machine can correct that - if some not-yet-invented machine design makes it non-tamperable by the people who run it, you just find 1,000 party stalwarts (or bums who'll take bribes), give each one a voter ID card for every precinct in town, and you can tilt most local elections that are close enough to be tilted without being too obvious. Do it in enough towns and the right towns, and you can steal statewide and even close presidential elections. It probably happened in Chicago in 1960, and it would have been overkill in Florida in 2000...

OB,

The comment Jane quotes says nothing about audit trails. If you support them, fine.

Still, it's interesting that Diebold opposes them. Is that evidence of a conspiracy? No. But it doesn't give me confidence that they are deeply concerned with the accuracy and verifiability of their machines, your "trust" argument notwithstanding.

And there are a few other things worth noting. Vote fraud is hardly unknown in the US. Presumably those who have engaged in it have borne the risks that you say Diebolders would never take, and they have done so because they wanted very much for their candidate to win. So to argue that no one would do this because the risk is too great for the reward flies in the face of history, and I think that to classify those who have concerns as "nutjobs" is unfair.

Bernard,

Jane's quote is an excerpt from a longer post made under the topic concerning Colorado's Electoral College votes.

I have no idea why (or, indeed, if) Diebold resists audit trails. I'd want to hear their side of this before deciding, but at first blush, I'm with you on this one.

Your point re voting fraud doesn't seem on point. Of course voting fraud takes place all the time, all over the country. But the operative phrase is "all over the country," a patchwork of small frauds, a precinct here, another there, some people voting multiple times in one place, the deceased queuing up to vote in another. No ONE fraud is that momentous, each is a series of miniscule frauds (individuals voting twice, etc.) that are not really worth prosecuting on a case by case basis.

Also, the people behind the fraud have something to gain. They are trying to keep their jobs, the local political power to award contracts (with kickbacks), hand out the patronage that is the currency of local politics, and last, to retain political protection from prosecution for the previous acts of corruption they doubtless committed.

So it very much matters to them who wins the election. If those organizing the fraud lose, they are out in the street, no money, no power, and they have to worry about prosecution for any previous bad acts.

Contrast this with the Diebold scenario. This would not involve a whole series of petty frauds, no one of which is worth prosecuting, but rather one gargantuan fraud that would change the history of the United States. This is no ballot stuffing job; if caught, the perpetrators would feature prominently in American history books in perpetuity, and regardless of which party won, they would be prosecuted, absolutely guaranteed.

Moreover, for the manufacturer (i.e., Diebold) to influence the election, the fraud would have to be perpetrated at the hardware/software level, rather than at the local level. Such a fraud, if discovered, would have their pawprints all over it, and they’d be having a tete a tete with the U.S. Attorney in record time.

So the risks are huge. What are the rewards for success?

Precious little, it seems to me. Unlike the local politicians above, what contracts or patronage could the corporate executives who are suspected of planning this fraud possibly hope to receive that would be of interest to them? Or, more to the point, that would justify the huge risks they’d be running? So Bush wins – what do they get? Or Kerry wins – what do they lose, relative to what they have right now?

One parallel would be to murder vs. assassination. A street murder is one thing. The likelihood of being caught is so-so, and if caught, so is the likelihood of severe punishment.

Assassinating a President is quite another. It’s a monumental act that most certainly would be investigated to the nth degree, with a high likelihood of getting caught, and with severe punishment a certainty.

In similar fashion, the occurrence of petty voting fraud at the local level doesn’t speak to the likelihood of a massive, centralized theft of a national election that would change the course of American history.

So, sorry, Bernard, I have to stand by the “nutjob” characterization. Believing that the CEO of Diebold is so committed a Republican (and is stupid enough to let that be known) that he would risk a very nice life indeed to guarantee Bush’s re-election is, in my view, irrational.

OB,

There is something to what you say about vote fraud, and the incentives, or at least the political ones, faced by Diebold. But the case is not so strong as you say.

While fraud in the race for Mayor of Smalltown may not attract the attention of the DOJ, it might attract the local DA, especially if he's of the opposite party.

Similarly, while vote fraud may be localized, it can still affect a major race, even, of course, the presidential election. That kind of small-scale large-impact fraud would, or should, attract a lot of attention.

That vote fraud does not seem to be frequently prosecuted may have something to do with the difficulty of proving it, or pinning it on someone specific, or on the fact that it tends to be a bipartisan activity so it's safer to let it go and just call it politics as usual. And of course electronic vote fraud might, if done cleverly, be exceedingly difficult to establish.

As for incentives, there could easily be more in it than Bush's election. How about massive bribes? Would the CEO risk what you rightly call a "very nice life" in exchange for, not Bush's election but even more oodles of money? I don't know. Ask the Rigas family, or Andrew Fastow.

Part of one's view of this depends, ultimately, on one's opinion of this Administration. If you think they would stoop to anything (and I do) then it's not totally insane to think they would look for ways to steal votes wholesale.

If you want me to trust your electronic voting system, I don't care if it's made by Diebold, Daewoo, or Dilbert&Dogbert Consultants (motto: "You know you can trust us, but you don't know what you can trust us to do"] it will consist of two parts.

Machine A will not be connected to anything electric or electronic except the power line. It will contain a very simple program that displays candidates, one office-screen at a time; a touch screen for voting; and some device for producing a semi-permanent change in some medium ("ticket printer"). The printed or changed object will be the ballot. It will not be human readable. It will be serialized by some means outside the machine system.

Machine B will accept the ballot. It will read out the offices voted for, and ask the voter to confirm: "You voted for John Kerry for President. Is that correct? Yes or No." or "You didn't vote for the office of 12th district alderman. Is that correct? Yes or No." If any answer is "No," Machine B spits the ballot back out and tells the voter to go back to Machine A. If the ballot is damaged or otherwise not readable, Machine B tells the voter that and directs a return to Machine A.

If the ballot is readable and the voter answers "Yes" to all the questions, the machine tallies the votes and drops the ballot into a box in its base. It will tell the voter what the serial number of the ballot was; the voter has then the option of taking note of the number.

At the end of the day, the vote total can be taken from Machine B -- but the ballots (already verified as readable and matching the voter's intent) are taken to another Machine B (perhaps the one from the adjacent precinct) and run through again. If the totals don't match the ballots must be counted by a third machine. If the third count doesn't match the other totals, an appeal is made for voters to come forward with their serial numbers and confirm the ballots they cast. There will be enough suspicious bastards to make a valid statistical check.

Machine A would be dead simple; I could write the operating software for it using a microprocessor. It would not have Gouraud shading of three-dimensional icons, or any of the other folderol -- just text on screen. Choice of languages would be the only control other than the votes. It would not be based on Windows, OS X, or any other commercial operating system. Its code would be available to anyone who asked for it.

Machine B could be anything, including a Windows-based computer. So long as the paper trail exists there's at least the theoretical possibility of getting the votes right, even if someone hacks it. Vote totals would be stored on an EEPROM thumb drive, perhaps.

I can think of half a dozen ways to spoof that, but it at least gives a chance. Having votes recorded only on a hard drive is a stupid idea that positively invites the notion of fraud, and if faced with such a situation I would assume that fraud was the intent, regardless of whether or not I trusted the company that made the machine.

Regards,
Ric Locke

Part of one's view of this depends, ultimately, on one's opinion of this Administration. If you think they would stoop to anything (and I do) then it's not totally insane to think they would look for ways to steal votes wholesale.

Reasoning, circular ....see "Circular reasoning"

Circular reasoning, ....see "Reasoning, circular"

Comments are Closed.