Slashdot Mirror
Buggy Voting Machines
dkleinsc writes "The NYTimes is running an article arguing in layman's terms that voting machines are inherently buggier (Sperm sample required. Sorry ladies) than most software systems because they are not tested properly. A fun quote: "Extensive discussions are under way at sites like VerifiedVoting.org, CalVoter.org, and the "news for nerds" forum Slashdot.org about inexpensive, practical ways to make automated voting as reliable as, say, buying books online. Their recommendations make sense."" We makese sense? Wah?
Saying slashdot makes sense is just a way to discredit the anti-voting-machine crowd. People will check out the dot, and they'll think it's a bunch of kooks complaining about this stuff.
Ever since I was a kid, my mom has been taking me into the voting booth (either to teach me the importance of voting, or for lack of babysitting). This year I voted for the first time, and it felt great. I don't mean emotionally, I mean physically. When I pulled the levers and flipped the switches, I was actually convinced that my vote counted. It was the neatest feeling.
It was rigged. Look at the Berkley paper backed by MIT.
a. The Supervisor of Elections has unreasonably delayed providing information.
b. The certification was based on inadequate and incomplete information regarding the election results.
6. Some or all of the information requested on Nov. 2, 2004 by Black Box Voting is still missing from 59 of the 179 voting precincts, including portions of or all of the voting machine tapes for those 59 precincts, which are a vital part of official paper record of the election results from those precincts.
7. Complete information on problems with the voting machines prior to and during the election has not been provided.
8. Complete information relating to memory card failures during the election has not yet been provided.
9. Only a partial list of the transmission logs from the Accu-Vote optical scan server has been provided. Despite repeated requests, the Elections office has refused to provide to the Volusia County Democratic party the official election results, now stating that those results will not be available until December 1, 2004.
10. The Elections office has provided incomplete data regarding Early Voting and Absentee ballots. The Supervisor of Elections, for example, reported that the total number of absentee ballots and Early voting ballots, combined equaled 89,999 votes, yet the published figures for those totals is 84,100 votes, leaving over 5,800 votes unaccounted for.
11. In addition to the pattern of delay in providing the requested information, the true election results are in doubt because of numerous violations of election law procedure and unanswered questions concerning the results.
12. The polls were opened early and closed late during Early Voting.
13. Many public records, including one signed results tape from a voting machine were found in the trash. Many of the requested records not furnished by the Elections office have been found in the trash. Results from the tapes found in the trash do not match the results of the copies of tapes furnished.
14. An email from Mark Earley, of Diebold Elections Systems, Inc., to the Elections office was provided which asked the recipient for an explanation of why Volusia County had more memory card failures than all of their other Florida customers combined, and then asked why the 17 memory card failures which the Elections office reported on November 3, increased to 25 before November 12, 2004.
15. The reported memory card failures were significant and troubling and included reporting zero votes after one week of voting, requesting permission to upload votes before the voting began, and messaging whether the card should be reformatted.
16. According to a statement by the Supervisor of Elections on November 17, 2004, the GEMS computer is not networked, and is "stand alone." The furnished computer logs show evidence of at least two attempts to remotely access the GEMS central tabulator, which is claimed to be secure. A computer screen shot printout on November 17, 2004 (found in the trash) shows that the GEMS computer at that time had two networked hard drives.
From Chuck Herrin's info sec website: It's an amazing demo. Be sure to check out the associated FAQ which is as easy to read for the layman as for the techie, and full of citations. Share it with you friends and family !
For you party-liners out there, Chuck is a Reublican who wants you to understand that this is not a partisan issue.
"Whoever would overthrow the liberty of a nation must begin by subduing the freeness of speech."--Benjamin Franklin
Gone! are the days of waiting in endless lines just to discover you were in the wrong presinct and you must waste more of your Tuesday just to be confused by candidates you don't recognise and legal measures with words you cannot pronounce all on a ballot whith confusing instructions like "punch out the third chad to vote for the second person on the left".
With modern voting and polling technology, ballot booths are a thing of the past. Automated computerized voting booths, using ultra-complicated statistical methods, can tell how you will vote even before you do.
So why bother with election day when 2,000 randomly selected land line phone owners without caller i.d. and fifteen minutes to spare can do the work for you? Just another wonder of the computer age.
at least here in manitoba, canada, we use a process where you a) get a piece of paper with straight lines connecting the checkbox with the person you are voting for, b) you place a checkmark on the candidate you wish to vote for c) you feed it into a machine which records your vote, and d) if there are any discrepencies, all the ballots can be counted by hand. all the ballot stations are manned by volunteers representing all parties. I can't understand why places like the ukraine and the USA have made this process more difficult than it should be. It's not like any one party is beter than another. Perhaps its just nations that start with the letter U.. Ukraine, USA, Uganda .. Uruguay.. heh. Uruguay.
I'm not a very experience programmer by any means, but why the hell would it be so damn hard to make a voting machine that works properly? It seems like a simple concept that even beginning programmers could do a decent job of creating.
I think that is because a VOTING system has to be designed for ALL kinds of users. From the youngest voter to the oldest, from the illiterate to the PhD'ed. If there were problems with users then it has to be considered that it's the fault of the voting machine or the instructional/assistance systems.
;-)
Regarding that comment about a Kerry win and the docile nature of the opposing party: It also helps that there are actual choices on the ballot too. Though I doubt very much that nobody in the Republican party would have questioned a Kerry win. Logic seemed to escaped YOU.
LoB
"Anyone who stands out in the middle of a road looks like roadkill to me." --Linus
This article compares India voting machines vis-a-vis Dieblold.There was also a previous slashdot story on this. These machines are much simpler and hence lesser prone to bugs.As discussed by the article , faith in this machines have been established simply because they have been used over the past few years by over 670 million registered voters in elections at national as well as state levels.
This simple article explains the EVM's used
http://www.openvotingconsortium.org/.
An open-source system that runs on commodity hardware, with an encrypted, anonymous ballot. Definite paper trail to allow for recounts. Why there isn't a clamor to get this off the ground is beyond me. A similar system has been working in Australia for years.
Knowledge is power. Knowledge shared is power multiplied.
Bruce Scnheier:
Computerized systems with these characteristics won't be perfect -- no piece of software is -- but they'll be much better than what we have now. We need to start treating voting software like we treat any other high-reliability system. The auditing that is conducted on slot machine software in the U.S. is significantly more meticulous than what is done to voting software. The development process for mission-critical airplane software makes voting software look like a slapdash affair. If we care about the integrity of our elections, this has to change.
http://www.schneier.com/blog/
No wonder it's intelligent and no particular surprise it mentions Slashdot -- the article was written by James Fallows, who as a long-time writer for The Atlantic was also a long-time technophile, or at least one who appreciated the productive uses of technology. I venerated him because he was a very public user and proponent of Lotus Agenda, a product which was unfortunately orphaned way back when and whose intelligence and functionality have never been duplicated.
Yeah, well this year if you had gone to vote in queens, bronx, or any other inner city district,
the experience would probably have gone more like this (if you were trying to vote for Kerry)
1) When you get to the front of the line, be told that if you want to vote a straight Republican ticket you can use any of the machines, but if not, you have to wait a little longer because half of the machines are "stuck on the republican side"
2) Get in the booth, pull down a lever, and have it not quite click. Or refuse to go down.
3) Notice that for some strange reason, you can only vote for Kerry as a Democrat, not on any of the other party lines, because the levers are broken (New York allows third-parties to nominate a major party candidate, so votes for that candidate get counted for the purposes of party viability. The Dems hate third parties.)
Someone had to do it.
I followed the problems with these machines on the news and apparently one of the most important problems is that there isn't a way to verify the vote count on each machine. These machines have a function where they print the total votes cast and that's it. No audit trial.
... And at the end of the day if for some reason there is a problem with the voting machines you can always go back to the transparent ballot boxes and count each individual vote all over again.
/obviously this calls for a reliable printer mechanism, like the ones with see on ATMs or Cash Machines, at least!
Why can't they attach a printer to each machine where the voter will see a paper ballot being printed at the same time he/she submitted their vote on the screen? They will hava a last chance to see their vote before it is automatically dropped into a see through box.
If there was a problem with an individual vote the person will call for assistance immediately and with a proper procedure in place, the vote could either be cancelled or approved.
Makes sense?
Although the machines may well be buggy, other sources of error may be more commonplace and more insidious. A prior /. article shows that some bugs occur in the metadata configuration created by officials of the particular election. Vote counting is really more vote interpretation than simply doing Votes[Candidate]++. And if the people configuring the software for a particular election make a formating mistake, the wrong bits will be counted for the wrong candidates.
These types of errors are hard to test for because it is not testable until the ballot is set and every new ballot demand a new round of testing. These types of errors won't be solved by better testing of the machine or by OSS. At best, the voting machine software designer can provide easy-to-use tools to ensure that the ballot layout and voting interpretation/tallying software is in sync.
Two wrongs don't make a right, but three lefts do.
It's been pretty well established that we won't have a fair and functional voting system until we have a considerably greater level of transparency and accountability.
You won't have transparency until every part of the voting process has been moved into the open source domain for thorough examination and auditing. The current systems are all closed source, and the system which "prevents" cheating is controlled by the same people responsible for gerimandering, and is readily bypassed via "emergency" updates.
Furthermore, we shouldn't have to file Freedom of Information Act requests in order to have ballot results released. This information should be freely available, preferably on the websites of the various counties that do the tallying.
Also, a person's vote absolutely must be recorded in a non-electronic manner at the time of polling. Paper ballots are essential. Even if those paper ballots are printed by the voting machine after the voter casts their votes, it must be produced. Otherwise, a recount is no different than refreshing the calculations on a spreadsheet.
While this is all a good idea, it isn't like a system like I described actually exists. I believe MIT formed a group to produce such a system, but four years later they've mostly just produced research papers. There is a group which is currently working on such a system, but they are currently suffering from severe under-funding and various bits of social blockage. They're the Open Voting Consortium. I strongly urge everyone to go check them out.
Wake up - the future is arriving faster than you think.
Voting may seem similar but it is VERY different. You get a statement every month to reconcile against your personal records. There is an individual trail that you can take to the bank and say "see, you fucked up!!! Give me my money back!!!"
No such burden exists for voting systems. The customer does NOT receive a statment in the mail.
Furthermore, I would suggest that the "once a year" model of "use" should NOT be a problem since these systems are SO FUCKING SIMPLE!!! Any developer worth his salt could design tests to find errors. And any company worth it's salt would EXTENSIVELY test their software before deploying it to the field.
The "private" nature of voting means that any system designed to allow a voter to "check", will probably allow others to "check" as well. The best solution I could think of is smartcard driver licenses that digitally sign your vote. But even then, the motor vehicle dept will have your "private key" as well as any other personal parameters.
I guess one could add randomly seeded keys to the voting machines and randomly generated numbers to hash each vote ID. But those to seem succeptible to precinct worker mischief.
In the end, the easiest solution is to BAN, the on-screen vote verfication phase. Vote verification takes place after a ballot is printed behind glass. If the voters rejects the ballot it is visibly voided in some way and the voter get to change their choices. If it's accepted, it's automatically placed in the ballot box.
-------- -------- Support Wesley Clark for president!!!
Yeah, but not before the author bashes two elected bodies, denies voting problems in the 2004 election, and throws around Microsoft's slogan "trustworthy computing."
After all that, then we make sense.
The best e-tally I've seen in the last while (and I've voted in 3 general elections in the past 5 months) is a paper ballot which you put into a cardboard sleeve when leaving the polling booth (so no one can see your vote). The sleeve with ballot is taken to a counting machine. It looks a lot like an electronic sheet feeder. You place the sleeve with a bit of the ballot sticking out face down into the feeder. It pulls your completed ballot out, and electronically records your vote. There is a small digital display showing your vote for 3 seconds. You can confirm that it scored your ballot correctly. The display blanks, and the paper copy of your ballot is stored (the machine sits on top of a large box which holds completed paper ballots). Electronic reporting is complete and exact, and there is a paper trail for recounts if necessary.
Yes, they performed an AUDIT, not a recount. And the headline was spun to "legitamize" the election. Indeed, if you knocked out all the ballots that the machines ignored, Bush DID indeed win the election.
If you used the standards of counting a hanging chad where no other chads were displaced, and you counted obvious write ins, Gore was the winner.
But I am glad to see a Republican who mistrusts these machines. Primaries can be hacked just as well as genreal elections
-------- -------- Support Wesley Clark for president!!!
If Kerry had won, would the voting machines still be buggy?
Just a thought.
My favourite excerpt is the following:
Of course, there is no proof that these gentleman have continued their illegal ways. They could have become completely reformed, law-abiding citizens by the time they started work on the Diebold voting systems.
What do your exection boards do when someone marks an X in BOTH spots? What if someone puts a slash in one, and a slash in the other? What if someone circles a candidate's name, and doesn't put an X? What if they put an X over the whole name? What if on the 10th counting, the light pencil marks on a ballot have been smudged off completely? What if they just put a tiny dot in the middle of the first candidate's box (like they rested the pencil there), then didn't mark anything else in either? I'm asking because this is the kind of nonsense that put Florida on the map 4 years ago.
I personally think that the current, unauditable, unverifiable electronic voting fad is a bad thing. I don't, however, think giving people a piece of paper and a pencil is necessarily the answer.
You're right that a paper ballot is a good thing.
There is a lot of good sense behind a two machine system -- One machine accepts user input, verifies user input, and prints a machine-and-human-readable ballot in a consistent and verifiable manner. This prevents the "input error" scenarios, where the voter doesn't mark the ballot properly; it also makes the ballot easy to machine count, and makes the mark more permanent than a pencil. The second machines just read and count ballots.
The voters enter and confirm their choices on the first machine, are given a paper ballot form they confirm (again), then they slide it into a ballot box. The paper ballots are later counted by the second machine, and if there is any doubt, they can be hand counted by the election board with observers from all candidates' election comittees present. Permanent record, recountable, two verifications by the voter (one on screen, one on the paper in their hand).
We apologize for the preceding message. All those responsible have been sacked.
http://thehighroad.org/attachment.php?attachmentid =18516
...it'll auto-start.
.AVI :).
It's a 158meg Bittorrent file (GEMSDEMO.avi.torrent) - if you have a client installed such as Azureus:
http://azureus.sourceforge.net/download.php
Playing time is only 15 minutes. File size is that big because it's in 800x600
I "filmed" it with a screen record utility with audio commentary voice-over. Sound is a bit low, but crank the volume and it'll work. It uses the Intel Indeo codex which I understand is problematic on Macs...sorry. Windoze Media handles it and I would suspect there's some Linux player available?
If anybody here doesn't "get it" yet about how screwed up their "security" (ha!) is, this will do. Makes sense to most non-techies, too.
If Kerry came up a winner, they Republicans would have done the SAME thing they'd done in 2000. They would have voided the results of an electronic voting machine (Volusia County 2000) and taken votes AWAY from Kerry.
This time, their vote theft was so overkill as to avoid that unfortunate event in 2000.
-------- -------- Support Wesley Clark for president!!!
I don't know, do you think they are coming up with all sorts of reasons the election was rigged regardless of the sources creditability, are spewing conspiracy theories, are calling people idiots, and spend all their time complaining about it on slashdot?
Why don't you go do something about it yourself instead of telling people what to do? Maybe despite the vocal outcry on the internet, you might find that there are more people who are not so vocal about it and agree with the election results.
I don't think buy it that the testing cannot be done in near actual conditions. Just have the American Idol voting done on these tabulators. And what's best, you'll have to make sure you get the results right or the public will be outraged.
But either way.... its fucked up
Ax: This ballot is spoiled and is not counted.
This is why I like the solution of machine generated, but human readable paper ballots. I think it can help cut down on ballot spoilage, which is often pointed to as evidence of political bias (and probably IS sometimes caused by political bias).
Before you write back saying that my answer to your first three questions (which was that the ballot is spoiled and is not counted) is unacceptable, ask yourself this: How hard is it to make a single, unambiguous mark (preferably an X as instructed) in a big white circle beside a candidate's name? And yes, to answer another question, for those people that have physical problems marking their ballot, they are allowed to bring an assistant or aide with them to mark their ballot.
Given the difficulty that people have had with paper and pencil in the past, I figure it must be pretty difficult!
I think we're both in agreement, though, that having a paper record (pencil- or printer- generated) is a good thing, and allowing all parties to inspect and recount them is essential for having an auditable and believable election.
We apologize for the preceding message. All those responsible have been sacked.