The Diebold Machine

An article in The Independent looks at certain… anomalies… in the last midterm election. In certain states where new computerized voting machines were used, there were huge discrepancies between the pre-election polls and the outcomes, as well as quite a bit of apparent procedural sketchiness. It's important to take the appropriate sized grain of salt considering the source, but still… somewhat disturbing stuff. Read the piece before saying "ah, it's The Independent."

Comments

[Gene]

Are they really saying that the vote counts are being purposfully changed??? If that is true, how would the companies do it?? Companies like this maybe run by corporattions, but the people who do the actual work would know if someone was tampering with data. There would surely be a "whistler". I think its just "sour" grapes and an effort to muddy the waters.

[TSPUR]

What's wrong with The Independent as compared to any other paper? I find it a well written source of news and opinion, the first overseas paper I read everyday -- they carry the great Robert Fisk among many other benefits though you have to slap down some quid to get him...

[Dan]

At the risk of sounding predictable -- ah, it's The Independent

Comments like "this is a hack" and "this doesn't really work" are common in the source code for large projects. Dividing by an integer "1" is a way of insuring that the resulting number is also an integer (it's a dippy way to do it, but since when do all engineers do things the sensible way?).

The evidence points to electronic voting being shoddy and insecure (ie, like most software), but there's zero reason to suspect foul play. That the companies (like most companies) are owned by Republicans amounts to little more than an ad-hominem attack.

And this line really takes the cake:

The possibility of flaws in the electoral process is not something that gets discussed much in the United States.

What the fuck planet is this dolt from? It's gotten discussed, ad nauseum, on TV, in newspapers, in magazines, and on the web, in just about every local, state, and national election since "Bush v. Gore". Then again, this is the same paper who thinks no "public dissent" has been allowed since 9/11, so maybe they consider anything less than "daily headlines in 150pt font" to be "silence".

[s.m. koppelman]

What is wrong with the Independent? It's not like it's the Revolutionary Worker or Granma or something. Heck, it's not even the Guardian.

If anything, the article understates the potential for tampering on the Diebold machines. As used in Georgia in 2002 and in many other localities even now, their voting machine software is a rudimentary Microsoft Access-based application with no password protection on the database, no transaction logging--electronic, printed or otherwise--on the database to record changes to data records, no role-based access control to allow the software to record a vote with a timestamp but block the system in the course of normal operation from changing a timestamp or a vote. Nothing. These are not difficult things to do, and though MS Access isn't a great database engine out of the box for high-security, high-integrity applications, there are plenty of ways to make tampering with and fabricating data tough. Small-town plumbers have written more accountable, auditable little database programs than what's in Diebold's voting machines.

I'm still not sure what's wrong with mechanical voting booths or even handwritten ballots. They seem to have worked pretty well for a long time. Recounts are straightforward, and ballot integrity can be tracked through the reliable, low-tech solution of letting representatives of each party watch the ballots get carried from the machine to the tables where they're counted.

Noise in blogland and inchoate complaints about November 2000 notwithstanding, I think it's fair to say there's been surprisingly little concern about the perils of electronic voting in its current form. Just the fact that Georgia machines were running a "patched" (several times over) version of the software that election officials didn't review (and in fact were barred from looking at!), in violation of the state laws that require machines to be running software that's been certified, should have been enough to prompt the state courts to nullify the entire election wherever the Diebold machines were used.

[Warren]

Man: [carrying large stack of paper] Here you go: the results of last month's mayor election. All 48000 voters and who each one of them voted for.
Lisa: I thought this was a secret ballot.
Man: Ehh.

Bart: Oh my God...the dead have risen and they're voting Republican.

[Nicholas Weininger]

It would be nice to know how often swings of the magnitude Gumbel cites occurred before the introduction of electronic voting, or how often they occur in other industrialized countries with different voting systems. Surely the data must be available. My intuition is that such swings would actually be quite common; even if polls are usually right on, in a large country with a great many elections going on at the same time, it's likely that a few results will be way off the polls. But a comparative study would really help here.

[Nicholas Weininger]

And another thing: it's not actually an ad hominem to point out that the companies are owned by politically active Republicans. An ad hominem is a statement about someone's motives for *making an argument*, presented in lieu of a criticism of the argument; it's wrong because the truth of an argument does not depend on the motives for making it. Pointing out that someone has a motive to *commit an action* is not an ad hominem, because the question of someone's motives for committing an action *is* relevant to deciding how likely it is that they committed it.

[Gene 6-Pack]

The owners are Republicans? But the programmers are all libertarians or democrats so it balances out.

[JDM]

Reading and thinking a little more about this, it continues to fall apart. The machines in question can print out each voters vote, and the county officials seem to be choosing not to do that. The password issue is moot. The "GEMS" application using the Access database is an application which tallies the votes, not the application which registers the votes. I didn't find anything about whether the voting machines themselves use Access or not. It is unimportant whether the actual file used by Access (the ".mdb" file) is password protected or not, because the file can't be accessed without having the password to the machine in question. The GEMS application is clearly intended to be run on Windows NT, which can be set to log all file access. If the county or town wanted to they could set up the GEMS machine to do just that and invalidate any election in which the files or machines in question were tampered with.

So the election procedures used here are not incredibly secure, but the whole evil Republican corporation angle is total bullshit, of course. The machines and the corporations themselves are not preventing anyone from holding an election which is as secure and auditable as a paper ballot election.

Having said that, I don't know why they don't stick with paper ballots. It leaves less opportunity for bureaucratic incompetence, and (sorry) tinfoil hat ranting.

Why is it that journalism is so shoddy when it's about something you know well?

[StMack]

What absolutely astounds me is that the function and security of voting systems is based on software with such an incredibly poor record as Microsoft's. By 2004 when these systems will be even more widespread look for a virus attack resulting in a presidential election victory for the Green party.

[Rob]

StMack:

I think you're the only one here who's getting it right. If the state and federal governments want touchscreen voting, they should form (for want of a better word) a consortium to build a completely proprietary system to do touchscreen voting. It's not as though running a touchscreen is some major technical breakthrough. Sure it would be more expensive than building your app on Windows, but it would be a completely self-contained system. Just pick a processor (only one; no need for JVM here, thanks). You don't need half of the services that ship with Windows to run a application that presents a display, logs the user's choices, and prints out a slip of paper for verification. And I can't imagine how the nature of picking candidates from a set of lists is likely to change radically in the future.

I don't want to sound like an open source zealot, but the code should be completely open. There's no way in hell I want to vote on a system that I can't inspect. I write software for a living. I've seen the kind of crud that can creep into a large application if you're not careful.

[StMack]

Rob,

I use software for a living and I too have seen the kind of crud that can creep into a large application (Microsoft Office), but I'm not sure I can get on board with the open source idea unless every voting kiosk was a stand-alone.

I have to ask the question of how many of the current voting sysyems have you inspected? Have you inspected the code that runs the machines counting punch cards or other machine read paper ballots?

[amr]

I think that by bringing up the Minnesota vote, where the incumbent Senator died, and where we have optical-scan voting, undermines the point of the article. To me it says, similar changes happened in a state where I doubt there were any shenanigans about voting systems.

I do think it's silly to move to touchscreen instead of optical scan, however. What do you do for a recount? Just look at the totals again? One thing I think our MN sec of state (R) did good was bucking the '00 CW that touchscreen was the way to go. We used plain paper (hand-count) ballots for Senate due to Wellstone's demise.

It would be interesting to see how the Georgia vote for Governor and Senator compared to votes for lesser offices and previous voting. Holding it up against previous polling does not convince me, because I've seen some really bad polls (including that '02 Senate Race.) I don't have time to check Georgia's votes myself, though.

[Jason Ligon]

This is whako. You mean to tell me that the evil corporations forced programmers to elect Republicans?

What is with the corporate paranoia of the left? Why is it an accepted good for Ma' and Pa' to have a store that sells jelly, but a an evil for Ma and Pa to incorporate to sell jelly in 23 states? ARRGH!

[Jason Ligon]

I might add that it is the left that is driving the race to more technologically advanced voting methods. Somewhere between chads and diabolical software engineers, they will find a system they can live with, which is to say one that elects all Dems.

[joe]

"I'm still not sure what's wrong with mechanical voting booths or even handwritten ballots."

As I understand it, they were abandoned for more modern methods not to insure greater accuracy, but to make the counting process faster and cheaper.

Which hasn't exactly worked out.

[Jesse Walker]

When I left North Carolina for Michigan to go to college, I was impressed by the voting machines I found there. Back in NC, we still used paper ballots. Michigan was obviously high-tech, what with these levers and gizmos and such.

I didn't realize that the machines had been around for about a century and that still other states considered them as archaic as the Carolina system.

I didn't encounter punchcard voting until 2000, when I voted for the first time in California. Within a few weeks, I was hearing how outdated that system was, thanks to Florida.

Which to me meant they should just go back to paper ballots. Sigh...

[Kevin Carson]

There are some touch-screen machines that create a paper trail for recounts, I believe. After you vote, the machine gives you a printout for verification, and you put the printout in a locked box. That wouldn't be so bad, assuming most people would bother to read the printout before they turned it in.

[Skip]

Does the article mention anything about the exit polls for these elections? Wouldn't the election results have to be wildly dirrerent from the actual total for suspicon to be created? These polls, which many complain about, is actually a pretty efective check sum for us.

[JDM]

"The reason that emerges from the e-mail is that he wanted to make the software compatible with WinCE 3.0, an operating system used for handhelds and PDAs; in other words, a system that could be manipulated from a remote location."

This is absurd fear mongering. Pocket PC is an operating system for PDAs *based on* WinCE. WinCE is MS's operating system for any type of appliance, such as cable boxes, embedded devices (network routers, etc.) kiosk systems (such as, let's say, voting machines.)

If this is actually an Access application, however, I would have to say that it is probably not such a great idea to trust it without a paper trail.

[Andrew Gumbel]

Andrew Gumbel here, the author of the piece. Pro or anti, I certainly don't intend to stop you talking about my piece. One point of information, though, in response to Skip's question about exit polls. (This got edited out of my piece for space reasons.)

As you may recall, the Voter News Service, the media consortium charged with exit-polling last November, suffered a (never explained) computer failure, did not publish its results and subsequently went out of business. There are all sorts of unproven conspiratorial theories of foul play out there, which one can take or leave, but one question seems worth asking:

Whatever happened to the data that was collected? Even if the computer crashed and it could not be tabulated in a timely fashion, there is surely no reason why it can't be done more slowly after the fact. Why can't we see it now to see what VNS found?

[Goodman Erika]

EMAIL: [e-mail]
IP: 62.213.67.122
URL: http://www.pills-for-penis.com
DATE: 01/25/2004 11:09:00
Morality by consensus is frequently morality by convenience.