I can now confirm that Diebold machines are teh suq

[vvvexation]

Of course, the Alameda County registrar of voters is also teh suq. Here's a letter I just wrote to the Tri-Valley Herald:

As an Alameda county poll worker, I'd like to explain a couple of things you mentioned in today's Herald article.

First, you said that "Poll workers failed to post notices or otherwise inform unaffiliated voters that they had to request a Democratic or Republican ballot in order to vote in those primaries." I can tell you that this was not a failure on the part of the poll workers. We were specifically instructed during our training not to ask voters if they wanted those ballots; we were only to give them to those who spoke up and requested them, though it was clear that most voters wouldn't know that they had to do that.

You also said, "Voters who did not have the mayor watching over them said poll workers would not let them void mistaken ballots and vote their preference." Now, there may have been inflexible poll workers in some precincts, but I can also tell you that it was often impossible to let people cancel their ballots and get new ones. Once the "Cast Ballot" button had been pressed, for example, the ballot absolutely could not be canceled, and many voters didn't complain until after they'd pressed the button--somewhat understandably, since requesting assistance would have involved either yelling across the room or leaving the voting machine unattended while they came to fetch one of us. This is in sharp contrast to the paper system, where voters could easily take their half-voted ballots back to us and request new ones; with these machines, voters can only get their cards back if they press either "Cast Ballot" or "Cancel," and during this election, incredibly enough, we found at least once at my precinct that the "Cancel" button didn't show up anywhere on the screen.

I hope this helps to clarify things.

---

For the record, most of the other things he says in the article are quite true. Our card encoder was one of the problematic ones; it showed a "serious error" message when we turned it on and it took several phone calls to the ROV over the course of a couple of hours before I was put through to a technician who could walk me through fixing it (some sneaky things had to be done in order to delete some files that evidently should've already been deleted. For what it's worth, the sucker was running Windows.) It also took a couple of hours for them to send us a substitute poll worker to replace two of ours who didn't show up, and a while beyond that for them to send us extra provisional ballot envelopes as we'd of course run out of them before we got the electronic system working. Then again, the same thing had happened a few years ago, before the touchscreens were even in use; I was working a precinct that was mostly dorm-living UC students who often hadn't bothered to change their registration (or they had and it hadn't been recorded correctly), so we ran out of envelopes twice and had to turn people away while we cooled our heels waiting for more to be sent over.

Now, all of that was merely due to the ROV being either understaffed or just incompetent, I'm not in a position to say which, but we also had problems with the machines that were scary--not just the disappearance of the "Cancel" button, but the disappearance of local measures and other things from ballots that should've had them, as mentioned in the article. And of course there was no way to fix this. Before I'd worked closely with these machines I was merely worried about the potential for tampering; now I know they're unreliable enough even without being tampered with. I'll certainly be complaining.

Comments

Birds of a feather

[zyxwvut.livejournal.com]

It occurs to me that you and abenn would probably have a lot to talk about.

Z

P.S.: How're you feelin'?

[capnkjb.livejournal.com]

Now I'm doubly glad we didn't have the machines. Of course, it would've been hard to have them, given that the polling place was some old fogey's garage on a steep hill.

But yeah, I'll keep all that stuff in mind. I registered as non-partisan, mostly so I wouldn't get too much party mail clogging my mailbox, but I piped up about the ballot because I wanted to weigh in on the Democratic nominee. They were nice old people and gave it to me. I had to tell them, "No, this is not my first time voting, I just dig it HARD". :D

I like 'teh suq'

[persnicketus.livejournal.com]

... and might have to steal it.

I didn't know you were a voting person dealy! Er, I'm sure there's lots of things about your life that I don't know, and I'm fine with that, but I give mad props to you for contributing to anything to do with voting. It's the thing we need to do as a democratic society. Those machines sound terrible, though.

[abenn.livejournal.com]

Diebold sucks big time. Those encoders were horrible.

I like your letter; you stated the facts from the poll worker side well. The problems you had must have been incredibly frustrating. Didn't your precinct coordinator show up to help you? If not, that person may not have gotten paged in a timely fashion-- it took the office almost two hours to page me about a VCE failure in one of my precincts, and they never paged me about a second one at all. After that I broke protocol and gave my inspectors my cell number so I could respond to them faster, which did become necessary as the day went on.

You sound like a dedicated and competent inspector. Mine were all good, too. I hope that the problems with the job don't scare any of the good ones away, because we'll really need the competent ones in November.

[terpsichoros.livejournal.com]

V - I'd like to pass this on to some political commentators who've been discussing the Diebold machines, but you may not want complete strangers poking through the rest of your LJ. How may I do this? I can post your comments with your name, or as "a Berkeley precinct worker",

[flwyd.livejournal.com]

In engineering, there are several key requirements. There are user interface requirements about how people interact with the product. There are functional requirements which dictate what the product needs to do. There are performance requirements about how quickly/cheaply/etc. the product must do its job. There are acceptance requirements about how often and in what ways a product may fail.

From the information I've seen, on each of these points, Diebold either has not done a good job gathering requirements or has done a lousy job meeting them. It seems like the ideal solution would be to hold several reasonably-sized simulated elections, with actual people from disabled to barely literate, and then do an entire system redesign with input from participants at all levels. Unfortunately, what we get is a product rushed to market without proper functionality and usability testing. While this is a natural stage in the lives of many products, something as fundamental as voting technology deserves something more akin to aircraft design than software design.

Now if the software was Open Source and volunteers could suggest improvements...