Forgot your password?
typodupeerror
United States Your Rights Online Technology

Touch-Screen Voting Snags Continue 522

Posted by timothy
from the poke-it-harder dept.
micromoog writes "New touchscreen voting machines caused problems last night in the suburbs of Washington D.C.. Several machines failed and had to be rebooted, and nine were actually removed from the site, repaired, and returned, in violation of election laws. The machines also failed to report their results correctly due to network problems. At least one lawsuit is pending. Interesting quote: 'County elections officials said it was the slowest performance in memory for counting votes on election night.'" Read on for more on how the current crop of electronic voting machines are faring.

Not every electronic voting machine misstep comes from Diebold; reader zznate points out that the Virginia machines came from Advanced Voting Solutions (dcw3 butts in: "The slogan on their home page really gives you a warm fuzzy: 'Helping Shape American History for over one hundred years.'"), as well as that the EFF won a decision for an accelerated court date of November 17 in their attempt to stop Diebold from shutting down sites that make the infamous memos available. Let's all hope this is the first in a series of many wins for the EFF against the Diebold folks and crappy e-voting schemes in general. Have you donated lately?"

Reader meadowreach writes points out more trouble on the other coast: "From news.com: 'As voters in California go to the polls, the state is launching an investigation into alleged illegal tampering with electronic voting machines in a San Francisco Bay Area county.' Diebold upgrades software without letting the state know? How reassuring."

Generic Guy writes "CNN is running a story about California not certifying the Diebold voting machines and instead opening an investigation into the use of uncertified systems. Maybe there is still hope for democracy in the U.S."

And from Cambridge, Massachusetts, Peter Desnoyers writes "Cambridge uses an optical scanner system, where you fill out SAT-style ovals with a pen and the election officer feeds them into a scanning machine. From last night's preliminary results on the Cambridge website:

'In two precincts at 7:55 and 7:59pm the memory cards reached capacity. To ensure that every ballot was counted , the Election Commission has decided to rerun the ballots for 9-1, Lexington Avenue Fire House and 11-3, Churchill Avenue. We expect that it will take between one to two hours.'

I interpret this to mean that they took all the paper ballots out of the box and ran them back through the reader. (with a bigger memory card?) In the mean time, voters were able to continue voting and no votes were lost."

This discussion has been archived. No new comments can be posted.

Touch-Screen Voting Snags Continue

Comments Filter:
  • Oh no. (Score:5, Informative)

    by DrEldarion (114072) on Wednesday November 05, 2003 @03:11PM (#7398869)
    Scantron sheets for voting? That's NOT a good idea. I'm currently working for a company that deals with standardized tests, and those things are a PAIN to clean up in the database becaues NOBODY can fill the damn things out correctly. I'd say at least a good 5% of them have messed up bubbles in the user/test-ID field ALONE. The answer fields usually fare much worse.

    These aren't just 2nd graders, either. High school tests are usually WORSE in this aspect.
  • by jdreed1024 (443938) on Wednesday November 05, 2003 @03:12PM (#7398890)
    ... made by Diebold, it should be noted. They are the AccuVote OS [diebold.com] models. This is not indicated in the article summary, however it is the case. I voted in Cambridge last night, and noted with mixed emotions the little Diebold logo as I slid my ballot in, and then the machine rejected it. (It worked on the second try)
  • by Eraserhd (21298) on Wednesday November 05, 2003 @03:13PM (#7398908) Homepage

    ... and talk to your representatives in the house to get them to sign on. HR2239 requires touch-screen voting machines to print a receipt which the voter can read, then drop into a lock-box. This receipt is then used for recounts, and in a mandatory recount of .5% of districts chosen at random to verify the accuracy of the machines.

    While you could theoretically build a cryptographic system to do something similar, I'd rather not have a theoretic democracy!

    (Petitions are linked to at the bottom of VerifiedVoting.org [verifiedvoting.org].)

  • by jdunlevy (187745) on Wednesday November 05, 2003 @03:17PM (#7398967) Homepage
    My favorite part of the Washington Post article [washingtonpost.com]:
    The problem came when precinct workers tried to electronically send results from the 953 new machines to election headquarters,
    unexpectedly overloading computer servers.
    (Italics mine)

    "Unexpectedly"?? What, the servers hadn't been set up with the expectation that they'd be receiving results from lots of new machines at the same time?

  • by DavidH_Mphs (657081) * on Wednesday November 05, 2003 @03:18PM (#7398975)
    ok -- so the fix-it people removed some of the machines, took them away to the fix-it place, repaired them, and then took them back to the voting sites?

    This raises serious questions about the accuracy of the count, no matter how many machines had to be fixed. One machine or twenty machines, if you've got to take one away for repair & then bring it back, the accuracy of the data must immediately be called into question.

    If someone has to physically remove a machine, then something must be seriously wrong with it. What if they accidentally erased the data & then, in an effort to cover their mistakes, 'fudged' the votes?

    On top of that, election officials made a stupid error -- a preventable error. [Some] memory cards were full before the close of the polls.

    Election officials know exactly how many people are registered to vote in a given precinct. Therefore, they have the ability to determine the amount of memory they'd need on the machines. They should have asked the software folks, "how much memory will I need for each registered voter?"

    Instead, voters are left to fend for themselves as inept voting officials stumble their way through technology.

    This is completely absurd & inexcusable!

  • by pixelpusher220 (529617) on Wednesday November 05, 2003 @03:36PM (#7399186)
    from the AVN web site [enfocom.com]

    These things are wireless.

    All those that think this is a BAD idea raise your hands...

  • Re:What's wrong with (Score:4, Informative)

    by SVDave (231875) on Wednesday November 05, 2003 @03:45PM (#7399277)

    This leads to a question I have: why isn't pencil and paper good enough in the United States?

    One factor is that ballots in the United States tend to be quite long, due to our multiple levels of government and the fact that elected officials serve set terms (with terms at the various levels being synchronized). In the general election next year, for example, I'll be voting for three federal office (President, Representative, Senator), a state equivalent of a Representative, city council positions, a county council representative, various other boards (e.g. school board), probably a dozen statewide referendums and maybe one or two local (city- or county-wide) referendums. And it generally takes a month to certify an election, even with an automated counting process (which is why Arnold Schwarzenneger hasn't been sworn in as governor of California yet).

    What's it like in Canada? Does a general election include anything other than federal MP? Do you have separate elections for sepearate offices?

  • Re:What's wrong with (Score:2, Informative)

    by Llyr (561935) on Wednesday November 05, 2003 @03:46PM (#7399288)
    why isn't pencil and paper good enough in the United States?

    Two words: synchronous elections.

    In Canada we elect different levels of government at different times; we also don't have as many elected positions. So the most different ballots that might have to be counted at once are for municipal elections, where you might vote for mayor, a couple of councillors, and some school board members. Contrast this to voting for President, US Senator, US Congress, State Senator, State Congress, various municipal positions potentially including a local judge and sheriff and other local officials in addition to the usual mayor + councillor + school board, *and* various referenda. All to be counted at once. We, on the other hand, have asynchronous elections, and probably reuse the same counting volunteers a lot. Plus we have more positions that are appointed rather than elected, and very few referenda.

    Basically they have a lot more counting that has to be done at once, and as usual repetition leads to attempts at automation. Too many elections to scale well, rather than too many people.

  • Re:What's wrong with (Score:3, Informative)

    by wrenkin (71468) <alex.cookeNO@SPAMutoronto.ca> on Wednesday November 05, 2003 @03:47PM (#7399291) Homepage
    In municipal elections in Toronto, we also have pencil and paper ballots, which are then optically scanned. You fill in an arrow next to your candidate's name:

    Joe Schmo [== ==

    like this

    Joe Schmo [==#==

    And then it is fed through a scanner. Results are available soon after the polls close, and you have all the ballots if you want to recount (which you could do manually if you wanted to.)
  • by buck09 (212016) on Wednesday November 05, 2003 @03:47PM (#7399295) Homepage
    Diebold is using the DMCA to serve takedown notices to service providers. Mirror sites in the .EDU domain are needed to keep Diebold's memo's online and availible to the world.

    Go to why-war.com [why-war.com] and find out how you can help.

  • by color of static (16129) <smasters.ieee@org> on Wednesday November 05, 2003 @03:47PM (#7399297) Homepage Journal
    I voted in the State/county election in question Fairafx Virginia and I can tell you these things really troubled me.

    We have been using a combination computer mechanical system for years which I felt very comfortable with. Yesterday we walked in to find the new "WinVote" machines. They offered no privacy and were actually slowing down peoples vote entry by quite a bit (I saw most people take over 10 minutes to vote compared to one or two I would normally see).

    The officials were telling me about how one machine stopped working and couldn't be revived. The others they had apparently been able to reboot multiple times to keep going. They of course didn't know how the vote count was protected in these cases. I have a guess though.

    Before each person votes, an official inserts a smart card. The application restarts, displays some statistics and proceeds to allow me to vote. My guess is that the results are copied to the smart cards. In that case the state of the machine isn't really in question so long as the tally increases as the voter voted.

    What worries me is the use of smart cards. Now these tend to hold a handful of memory (8K to 64K in general), and can run some code internally. My question is, if a machine crashes then could it alter the contents of the smart card? A write only smart card would not have enough room for a busy polling location. A card where a count is updated would be vulnerable to coding or transfer errors.

    Like the user who asks for a database when they need a filing cabinet, I think this may be an idea to early for its time.
  • by rsidd (6328) on Wednesday November 05, 2003 @03:47PM (#7399298)
    The machines' manufacturers in India claim [rediff.com] that the machines are tamper-proof. This is 1980s technology (and the article I link to is from 1999).

    I'm not an expert but it seems reasonable. These machines are standalone units, not networked; they have hardcoded (machine-language) software on their chips, with no facility for modifying it or running an external program. To tamper with them you'd have to replace the motherboard with your own, on which you've embedded your own program, and even then it probably won't work since the machine has various safeguards for tampering. And these machines are extremely rugged and sturdy, and easy to use (I've handled them) and inexpensive (around $100 each).

    Sometimes antique push-button technology is better than the latest cutting-edge stuff (anyway, who needs touchscreens, what's wrong with buttons?)

  • by Chris Burke (6130) on Wednesday November 05, 2003 @04:03PM (#7399478) Homepage
    In the 1800's, life expectancy was around 45; now it is about 74. There were no old goats back in John Adams' day.

    "Expectancy" is the same as "average". In the 1800s the infant mortality rate was much, much higher than today, and this drags the average down. There were plenty of old goats in John Adams day. John Adams himself lived to be 81!

    The problem with averages is that few understand them. As evidenced by the joke you may have heard: "Oh no! It says here that 50% of Americans are below average!" If you think the humor is that someone would exclaim that rather than that the author thought the exclamation would be funny, then I'm talking about you.
  • by jbn-o (555068) <mail@digitalcitizen.info> on Wednesday November 05, 2003 @04:19PM (#7399646) Homepage
    Seriously, why don't we get/license the well working system that was put in place in Australia? Yes, its not domestically produced, but the source is there and can be verified.

    Because there are serious problems with that system. The software issues are virtually a red herring and do not make their machines trustworthy. Although it seems ironic to some, the same issues exist with free software-operated and non-free software-operated voting machines. Wired revealed big problems with eVACS but buried the description of the problems midway into their article and then posted their eVACS article under a misleading headline which is probably why you reached the conclusion you did. I commented on this system in that thread and responded to one of the system's developers when the software trustworthiness question was raised. [slashdot.org]

    The Australian system you refer to does not allow the voter to verify that their vote was recorded correctly and there is no permanent non-computer record of the votes to recount after the election. Even though the article quotes one of the system developers saying as much, this showstopper revelation is midway into the article and then apparently ignored for the purpose of writing the article's title. From the article [wired.com]:

    The [eVACS] machine does not include a voter-verifiable receipt, something critics of U.S. systems want added to machines and voting machine makers have resisted.

    A voter-verifiable receipt is a printout from the machine, allowing the voter to check the vote before depositing the receipt into a secure ballot box at the polling station. It can be used as a paper audit trail in case of a recount.

    Green [Phillip Green, electoral commissioner for the Australian Capital territory] said the commission rejected the printout feature to keep expenses down. The system cost $125,000 to develop and implement. The printouts would have increased that cost significantly, primarily to pay for personnel to manage and secure the receipts and make sure voters didn't walk off with them.

    Quinn, however, thinks all e-voting systems should offer a receipt. "There's no reason voters should trust a system that doesn't have it, and they shouldn't be asked to," he said.

    "Why on earth should (voters) have to trust me -- someone with a vested interest in the project's success?" he said. "A voter-verified audit trail is the only way to 'prove' the system's integrity to the vast majority of electors, who after all, own the democracy."

    As for the costs of securing and storing such receipts, Quinn said, "Did anyone ever say that democracy was meant to be cheap?"

    There's no way to determine if only the software you trust is running on the machine you vote with. Your /. post is vastly overrated (+5 Insightful).

  • Re:GOP suit (Score:2, Informative)

    by dtaciuch (229229) on Wednesday November 05, 2003 @04:29PM (#7399731) Homepage
    The machines in question (in Virginia) are not Diebold machines. They are AVS.
  • by pudge_lightyear (313465) on Wednesday November 05, 2003 @04:45PM (#7399895) Homepage
    There were two big problems with the new machines yesterday here in Mississippi.

    1. The lines were very, very long... Now... we did have record turnout yesterday. But, not in the fashion that would cause 2 to 3 hour waits to vote. I was perfectly happy to wait an hour to vote, until I saw what I assumed the reason was that I was waiting. When I got to the front of the line, I saw that most people were standing at the machines for more than 5 minutes each. There's no way the old process would have taken that long. Of course, in my district, and at the time I went yesterday, most of them were older... but IT'S A TOUCH SCREEN WHERE YOU TOUCH A CANDIDATES NAME!!! I think (and saw) many people turn away from voting because of this wait.

    2. I finally made it to a machine. First vote... (don't hate me for it if you don't like him) I clicked on Haley Barbour... his name did not check... instead, a democrat on the other corner of the screen checked. I clicked his name to remove it... Clicked Barbour again... same problem. This happened with about 6 candidates on the first screen. Finally, I was so frustrated that I had to call someone to help (which is bad because of privacy). She said that it had been happening a lot today and was supposed to be fixed. Then, she showed me how to "RUB" the machine so that the candidates would highlight (which worked most of the time, not all). I almost had to vote for a candidate in an uncontested race that I didn't want to. I said, "I don't want to vote for him." A vote helper person said, "He's going to win anyway."

    Total time voting... 10 minutes.

    Do the touch screens work? Not at this time.
  • Re:What's wrong with (Score:4, Informative)

    by iabervon (1971) on Wednesday November 05, 2003 @04:46PM (#7399907) Homepage Journal
    The third section of this article actually talks about Cambridge MA, which does exactly that. Somerville, next door (where I voted yesterday), does that as well. And Cambridge actually did have a hardware problem and rescanned the ballots.

    As for the capabilities of this system, you can actually do ranking quite easily (and Cambridge actually does use ranking in their voting, although Somerville does not); you don't write numbers, but you fill in an oval in the grid for each rank, like on the SAT.

    Cambridge City Council elections have an "instant runoff" system, which uses some algorithm that eliminates candidates who have either won or lost, and moves votes to candidates the voter ranked next until the right number of candidates have been chosen. It's what you'd expect for the city that has both MIT and Harvard in it.

  • by wrmrxxx (696969) on Wednesday November 05, 2003 @05:00PM (#7400074)
    Actually, we've got another well working system down here that we could probably arrange for you to licence. We call it pencil and paper, and it has worked without fail at every election I have ever voted at. The results have always been counted very fast, and there's a good audit trail.

    Every adult in Australia votes (it's the law), so we know the system works well even with lots of voters. The voting system is a preferrential system, which is more complicated to count that the 1st past the post system used in some other countries. The ballot papers are counted entirely by hand, yet we get results out on election day.

    One of the things that helps make Australian elections very smooth (apart from the fact that politicians keep getting elected) is that we have a federal body (the Australian Electoral Commission) to oversee elections. They control the process in every state - we have nationally consistent rules and processes. They seem to be very organised to an outsider like me: they pop up at election time and run the whole show like clockwork.
  • by Anonymous Coward on Wednesday November 05, 2003 @05:02PM (#7400090)
    Why not? Because they're nearly different companies... That is, Diebold bought the firm that makes the voting machines, and I doubt that Diebold the ATM guys made any sweeping changes to the new Diebold Election guys (DESI), especially since they made the purchase in January 2002..... yahoo .com link for more info [multexinvestor.com]
  • Re:What's wrong with (Score:3, Informative)

    by steveg (55825) on Wednesday November 05, 2003 @05:03PM (#7400105)
    At least here in California, I can go to any polling place as long as I can confirm my identity

    Not in my part of California. I have to go to the precinct that contains the address at which I'm registered. They have my name and address on a list, and they check it off.

    If I go to the wrong precinct (last election my polling place handled three precincts) they tell me I'm not listed there and to go to the correct one.

    I suppose if you were registered multiple times at different addresses you might be able to scam them, but you don't just have your choice of polling places. Not in Bakersfield, anyway.
  • Re:Exit Polls (Score:3, Informative)

    by WEFUNK (471506) on Wednesday November 05, 2003 @05:19PM (#7400327) Homepage
    I await the day when the independant exit polls that the media conducts deviate by a statistically noteable amount from some non-verifiable voting machine. What will happen? The media's polls, while "unscientific" tend to be decent approximations, and the sampling error should be calculable. How much would the "real" results have to be off to raise eyebrows or, worse, to raise fists?

    This isn't too far from what happened during the 2000 American Presidential elections, where news services flip-flopped between calling the election for Gore and Bush and then finally recognized that their polls results and eventually the actual results placed each side well within the margin of error. This link [failureisimpossible.com] has bit of information, although it was apparently posted before the final resolution.

    I remember tracking the results on the web while watching it on TV and CNN's online coverage made it perfectly clear that the results were well within the margin of error at all times. So I think they were well aware of that a margin of error existed, but I guess the pressure to call it first was just too high on TV.

    It still boggles my mind that the new election machine companies are against paper trails -- why is printing out a receipt and putting it in a box _just in case_ such a big deal?

    Exactly because it isn't a big deal, and in fact it seems to be the best way to do it -- but without the computer. For this to work, they apparently need to get rid of the paper -- without the redundancy of computer + paper, it's harder for people to argue that the computer is actually the redundant part. I love new technology, but only when it's useful (well, usually). Throw out the computers and punch card machines, spend more on voter education, and go back to a technology that works better, faster, and for less -- paper and pencil.
  • by Joe Decker (3806) on Wednesday November 05, 2003 @07:29PM (#7401820) Homepage
    Around 8:30, the power went out in the apartment complex. The whole thing. *IF* the machines were still attached, I'm not sure they had UPSes. What does that do to the tallying? What does that do to the data stored? What will a reboot do to the system?

    Power: The Edge machines have internal batteries good for a few hours, built-in. The machines are left plugged-in the night before the election to make sure the batteries are fully charged. You know you've lost power even if you're outside because a yellow bar appears at the bottom of the LCD, it goes red when your internal battery pack gets within ten minutes of failing. We also had a separate external battery pack. We also could call for additional battery power (we'd have hours) if it looked like we were going to run out. Finally, in case all of this fails, there's a backup plan involving paper and the box you dropped your absentee ballot in.

    So, it's unlikely you'd actually reboot due to power failure. Still, Reboot doesn't seem to kill the results card in most cases. We didn't do that catastrophically, but we did turn (in training) the machine off and then on while the polls open/closed switch was left open, the right thing happens there (it notes it on the log, the counts stay there.)

    I wouldn't want to try it during the period immediately after the user presses "Yes, I really want to cast this ballot.", it appears that some sort of NVRAM is written there. FYI, The results cartridge looks kinda like a PCMCIA card, not precisely, but that's the basic idea.

    You don't need extra power to tally the results, etc. The machines have a small printer in the back, they print out totals at the beginning and end of the election (as determined by the turn of a knob.) The internal battery pack could run the printer. You've still got the results card, as well.

    It was a little frightening that when I dropped off my absentee ballot, that there was no lock on the box to go to the registrar's office.

    We had a small blue plastic seal, it's a little plastic doodad that connects the two zipper pulls, you can close the seal through the zippers but you can't open the seal without breaking it. It's just a small #'d piece of plastic, that seal should have been put on after the first absentee ballot was cast (after showing that first abs. voter that the box was empty), and should have only been taken off during the closing of the polling place. It's more likely they forgot (or didn't know to) put it on.

    Similar seals are used to protect against various bits of tampering on the electronic machines and in the final "bag of results" that is driven to the registrar's office (or some local field centers). The one that protects the results cartridge has its code # paid special attention to.

    If you didn't see a blue plastic thingee (it wouldn't have been big, like a master lock), it's worth mentioning it to the registrar, but more likely mistraining of the poll workers than actual malfaesence. (We had a mistake in our own implementation there, we used a yellow one instead of a blue one, but no real harm was done to the system thereby, it's just a piece of plastic.)

    Note that losing an absentee ballot would be possible after the poll-workers open the box after polling, as well. As during the day during aan election, there'd be some chance another poll-worker would catch you at it, unless you're in cahoots. Still nice to lock the ballot box during the election, though. (If nothing else, it saves you from having to tell someone you won't retrieve their ballot when they cast it and then say they made a mistake!)

  • by manchineel (699602) on Wednesday November 05, 2003 @08:12PM (#7402277) Homepage
    Last time I worked at a polling place, lawyers showed up from the ACLU, because there had been rumors of harassment at the polls and voters being turned away for bogus reasons. The lawyers had to be there all day and basically retrain the inept Polling Place supervisor. She was turning away people just cause they weren't on the register (there are many reasons why the wouldn't be on the register, but would be allowed to vote). The scariest thing to me was that the polling supervisor's best defense was "But I have been doing this for 30 years!" Electronic voting machines are horrible, and should not be used until the technical (moral, social, and political) problems are sorted out. But we are very far from having a perfect system. When I lived in San Francisco, there was an election where a whole bunch of dead people voted. Now that's democracy!

Suburbia is where the developer bulldozes out the trees, then names the streets after them. -- Bill Vaughn

Working...