Voting Machines and 'Calibration Drift'

An anonymous reader writes "Tuesday saw elections for school boards and city officials throughout Kansas. In Saline, ES&S voting machines in several locations were 'mis-calibrated,' and when the voter touched next to one candidate's name, the 'x' appeared next to another one. One person I talked to said he tried to vote three times before going to the 80-something-year-old election worker, who told him 'It was doing that earlier, but I thought I fixed it.' From the story in today's Salina Journal: 'The iVotronic machines used in Saline County are sold by Elections Systems and Software. In October, the Brennan Center for Justice at New York University School of Law notified 16 secretaries of state, including Kansas Secretary of State Ron Thornburgh, that the machines are known to record votes to the wrong candidate.' The county does calibrate the machines the day before each election, but, '... in conversations with ES&S on Thursday, [the county clerk] was told that the calibration might change during the day. "What they've seen is calibration drift on a unit," Merriman said. "They're fine in the morning, but by afternoon they're starting to lose their calibration."' There was also coverage of the problems when they occurred two days ago."

Comments

a new low for /.

[Anonymous Coward]

One person I talked to said he tried to vote three times before going to the 80-something-year-old election worker, who told him 'It was doing that earlier, but I thought I fixed it.'

What does that have to do with anything? It's not in TFA. Am I supposed to just take your word for it? Even so, what's it supposed to mean? Old people can't calibrate newfangled voodoo touchscreens?

Calibrate Per Use?

[hal2814]

How hard would it be to just calibrate per use? I know on things like a Palm Pilot you just touch three places and it's good to go. Why not do that for each voter (or at least offer it to each voter)?

Re:Calibrate Per Use?

[Anonymous Coward]

I agree. A system with a touch in the top left corner, bottom right corner, then the center of the screen would add only a very small amount of time.

After calibrated, the machine could show 4 lists of 4 items, and have the user select 1 highlighted element on each of the 4 lists to ensure that the calibration was correct. If they could not select the 4 items, a light could go off on the station to alert poller assistants.

I'm as frustrated as you are, hal. This is the type of thing that would come up in even the most basic systems testing. Even some of the worst programmers that I went to college with would have made sure that the system was calibrated properly.

Re:Calibrate Per Use?

[TheRaven64]

Or, for a more reliable solution, do the same thing they do with cash machines and, rather than using a touch screen, put a row of buttons next to the screen and get users to push the button next to the candidate's name.

Re:Calibrate Per Use?

[TapeCutter]

Or, for an even more reliable solution; The pencil.

Re:Calibrate Per Use?

[getuid()]

Even if it wouldn't be too hard, it's not how it's "supposed to" work.

A person is supposed to go into the booth and vote, not start to hack on workarounds for obviously faulty hardware design. To bring in the car analogy: it's like having electrically adjustable car lights, and having to re-set them every time you turn them on because they wouldn't remember their position when turned off, or simply "drifted" during you drive.

Besides: even if you and I and most Palm users are able to perform such (granted, relatively simple) tasks as calibrating a touch screen, not everybody is (think: grandma). And while one may argue whether this is good or not, one cannot argue about one thing: the constitution gives *everyone* above a certain age the right to vote, not only to those who can calibrate touch-screens.

And: while it was your *choice* to own a palm, it was not everybody's choice to vote electronically. It is (was?) imposed on us. So if somebody is making me vote electronically, they at least ought make sure the damn technology fsck'ing works. It's not like it's rocket science, and it's not like there wouldn't be easy ways to make it work reliably -- worst case, for example by using regular buttons left and right of the screens instead of touch screens (think ATMs of most banks).

Re:Calibrate Per Use?

[conureman]

At the polling places I've worked, the (touchscreen) AutoMark machine (for voters with disabilities, &c.) marks a paper ballot, which is counted by the M100 scanner. After the polls close, we seal the marked ballots up in boxes which never get opened up unless there is a problem with the computer's count. The protocols (except for the software) seem fairly robust and transparent, and skeptics are welcome to watch. That's Contra Costa County, YMMV.

Re:Calibrate Per Use?

[aynoknman]

I live in Minnesota. Even pencils didn't save us from court appointed observers. Our Senate election is still up in the air.

Minnesota's problem is precisely that they didn't have electronic voting. That would have allowed the result to have been determined before the election.

Re:Calibrate Per Use?

[tedshultz]

The calibration was so far off when I voted this week (on an ES&S automark), that I talked the election official about the calibration. More or less it is done the same way you used to with the palm pilot (touch 4 corners). The problem is that there is also some parallax issues as the screen is about 45 degrees off vertical. The result is the calibration can be a little off. A little off is fine and normal if the interface is good, but on the automark machines, they put the candidates names right next to each other so even a small error in calibration will result in the wrong candidate being selected. I took a video of my self voting, (it's here: http://shultzonline.com/vote/ ). in the video is is clear that not the person I am pressing is selected, and that a candidate only 5mm away is selected.

'Drift' sounds like a rubbish excuse

[Anonymous Coward]

Right. I've been using touchscreen gear for more than 11 years now. Monitors with touchscreens built in, tablet PCs, iMacs with touch-enabling overlays, two cintiqs of my own and many dozens I've sold and supported to graphic artists.

They NEVER 'drift'. I've not seen even the cheapest touchscreen gear 'drift'. What's with this drift excuse? That smells too much like an excuse for throwing elections. Color me for stating the obvious, but sorry that sounds too suspicious.

A small victory against voting machines...

[Anonymous Coward]

...happened in Finland last week. A few municipalities tested electronic voting in the last (municipal) elections and when (unsurprisingly) irregularities occurred (232 votes were not counted properly), the results were challenged all the way to the supreme court, which now decided that the elections must be held again. The lawyer representing the appealing parties has said that he doubts that any politician will ever propose electronic voting in this country again.

That outcome is thus quite positive but it would've been even better if the minister responsible for it had accepted her responsibility and resigned like many people demanded her to.

Not really

[aepervius]

The most probable explanation is extremly shoddy hardware engineering combined with extremly shoddy software engienering ina bid to make as much benefit as possible. I have seen this with another touch screen machine, and although I did not ask the team what was the problem in detail, the aforementionned point were the problem. The old adage probablym hold : Never attribute to malice what can adequately explained by idiocy and/or greed.

Call themselves engineers....

[hengdi]

Touchscreen calibration? I used to work for a company that built quiz machines and the like for the UK pub industry (circa 2000). Essentially they were simple PC's with a touchscreen (the monitor had a PS2 output).

We used to leave those machines running at various sites for YEARS, and I can't ever remember a calibration problem. And trust me, we'd know because when a customer starts to lose money they let the pub know about it all right. The biggest problem we had was the coin slot mechanism screwing up.

So now you're telling me that almost 10 years later and the calibration in a voting machine can't last A WHOLE GODDAMN DAY? That's service so bad it almost makes me believe in the conspiracy angle!

Re:Call themselves engineers....

[maxume]

I would bet $0.15 that the machines are being incorrectly calibrated.

Re:Call themselves engineers....

[VShael]

ALMOST believe the conspiracy?

Geez, what the hell more do you need? A video tape of Diebold executives laughing evilly while cashing cheques?

What the fuck kind of excuse is that?

[BitZtream]

Drift? Seriously?

You mean that kiosks in airports, malls, restaurants, hotels, atm machines that sit outdoors, my iPhone, my Windows Mobile phone, tablet PCs and god knows what else can be calibrated once and last for years, but these voting machines can't last for 8 hours?

Most traditional touch screens CAN'T drift. They need an initial calibration to align the location of touches to match the display to deal with manufacturing and assembly differences, but they don't actually drift, ever.

WHAT THE FUCK are they doing to get drift in the system? The $2 multitouch video on YouTube shows a system less likely to drift than this shit?

Someone needs to be hung. We need to start instituting criminal punishment for leaders of companies that produce crap like this. There is no accountability anymore because everyone hides behind 'the corp'. That shit needs to end now. We can either do it legally, or wait a little while longer and watch the public start taking the law into their own hands.

which way?

[v1]

So, have we found it more common for the calibration to drift to the right, or to the left ?