Slashdot Mirror
ACM Eyes Policy Position on Electronic Voting
while(true) writes "The ACM is preparing to take a policy position on electronic voting in government elections. It has a poll page up to get feedback from it's members and where they also explain their proposed position. The proposed position calls for a paper trail to ensure a physical record of the vote. Go there and place your vote if you are a member. The ACM Public Policy Committee could be a valuable ally in many questions that are dear to Slashdot readers in the US. They have already spoken out on issues such as the DMCA, DRM, and private policing of P2P networks."
E-Voting will not be trusted for quite a while. There are difficulties with authentication, and spoofing. I am glad to here ACM is trying to do something about it, as opposed to other companies just trying to scream out "I am better Uncle Sam, give me funding !", can anyone else guess why this project isn't opensource yet. Possibly because money talks more than common sense ?
*pets the troll*
The ACM has clout, considerably more than a bunch of unwashed geeks who troll slashdot all day. They're the closest thing the software industry has to a union.
The ACM isn't a bad thing to look into-if I had a regularly meeting chapter within 50 miles I'd probably attend. They're a good deal for students, getting them internships at conferences and hooking them up with lectures and talks. I hadn't even heard of a "public policy" angle to them, but I think it's a good thing.
If my answers frighten you, stop asking scary questions.
They've already spoken out against the DMCA, DRM and private policing of P2P networks, eh? I guess that shows how little their opinion matters then, doesn't it.
I'm far more interested in what the EFF's official stance is, considering they're the ones with the real legal and lobbying power (miniscule as it may be when compared to the twin bohemoths of the MPAA and RIAA)
1. Ballots would be cardstock. with 'complete the line' method for marking choices. Each voter would get one blank vote card. The vote card would NOT have their name, voter number, or any identifiable information on it. The cards *should* each have a unique 'serial number' for accountability, plus some preprinted mark identifying what 'election' it was for.
2. Voters can either complete the card by hand, or use an electronic terminal (which can be closed or open source, doesnt matter), which they would insert, and the terminal would compelte the lines for them, and give them back the card. This terminal would neither record or count any selections. All it would do is complete the paper card.
3. Immediately after either completing the card by hand or using the terminal to complete it (and having the opporunity to verify what the terminal printed), the voter would step to a tally machine (which would be immediately adjacent to the terminals and hand-filling our area, to avoid any possibility for anyone other than the voter to see or confirm what they have marked on their vote card) and inser the card on the tally machine. The design of both the 'vote card', and the tally machine would be required to be fully open, documented, and auditable by any concerned party.
No closed-source software, nothing hidden or proprietary. Vendors wishing to provide the 'terminals' for filling out the vote cards, would be required to supply tally machines for the cards as well, and they would be required to supply 'extra' tally machines, to allow for both redundancy, as well as for the voters not using the terminals. (Im thinking something like 5 tally machines for every 4 terminals)
The tally machine would
Not at all pointless. The paper trail is not going to be used for a recount in every district. The idea is that you only recount a couple(ie, enough to have reasonable statistical confidence in the machines not checked) of random districts or machines for each vendor, to verify that the machines are accurate. This is a means of quality control. There would be no doubt or delay in these checks because of things like hanging chads, no basis for wondering what the 'intent' the voter had, because they will be printed by a machine.
Note that the possibility of this kind of quality control by election officials will give voting machine vendors more incentive to be really sure things work and are accurate and secure.
I really wonder why people are so furious about paper trails.
Remember the ghastly voting issues in Florida 2000?
Well, they actually had paper trails, and it didn't change a thing. As it turned out, the courts ruled the recount illegal.
It seems that legal deficiencies of the US voting system are a much bigger problem than missing paper trails.
Don't forget that paper trails aren't immune to counterfeiting in any way. It's probably very easy to print a lot of paper trails with a standard PC and very little extra equipment.
It can't be that difficult built an electronic voting system, that is about as secure as the normal paper voting.
On the other hand, I don't really get why voting machines are so sought after in the first place. Here (in Austria) all vote counting is done by hand, in the local communities, with members of every party in the voting committee.
You only need a few people for a single day, and counting is insanely fast.
Since they start counting when the vote is still going', they have about 50% of the votes counted by the time the vote ends (usually 5 pm). At 5:10 pm, the first estimates are aired on TV, at 8:00 pm about 80-90% of the votes are counted and at 10 pm the Bundesministerium für Inneres announces the final results.
But in the U.S., most ballots are much more complicated. We (in the US) have a tradition of wanting the citizenry to speak out/vote directly on a number of different issues, and having seperate local and state elections. It's a pain to setup a poll, and a pain go to a poll, so a voting decision is actually more complicated for US citizenry than a non-US citizen might think. A vote might involve federal election (a President, House member, and a Senator), state election (a state senator / representative / governor), local election (county/town board, mayor, school board, sheriff, judges). It probably also involves multiple bond decisions ("shall the state take out a loan of $X to do Y"), and proposals to change the state constitution in various ways. When I go to a poll, I'd be shocked if there were fewer than 4 choices, and there are usually many, many more.
As a U.S. citizen, I'm used to it, and even like it -- it allows me to participate more directly in various decisions than citizens of some other democracies. And the multi-tiered approach to democracy is deeply embedded in how U.S. politics works. But the more complicated ballots, along with the sheer number of people in the U.S., make it the purely manual approach more painful. It's still possible, of course, but some sort of automation is desirable.
Untrustworthy automation is a terrible idea, of course. Hopefully various organizations like the ACM and Verified Voting will change the system so that we can actually have confidence that our votes are being fairly counted.
Oh, and the problem in the 2000 election wasn't that recounts are illegal. Recounts happen occasionally in the U.S., they're even required in certain cases. As I understand it, the problem was that recount rules are supposed to be consistent and clear before the election, and Florida's setup was revealed to be an absolute travesty. Of course, these unauditable electron-only voting machines have exactly the same problem; there's no consistent and clear way to do a real recount, because there's nothing that can be independently recounted. Instead of creating a recount travesty, they need to make real recounts possible. And a computer-printed (and human-verified) paper vote would eliminate the nasty problems in the Florida 2000 election, where it was incredibly difficult to figure out the voter's intent from a card with multiple hanging chads (with more hanging chads created through handling!).
- David A. Wheeler (see my Secure Programming HOWTO)
I looked on Drudge Report and it seems a couple of Representitives have requested UN oversight in the next election. It's Drudge, so who knows how accurate it is, but it would be wild to have UN observers validating the election. Personally, I think the elections should last a week and for an independent (publicly funded) group make sure every single eligible person votes. Or better yet, cast your vote on your tax form...heh.
[RIAA] says its concern is artists. That's true, in just the sense that a cattle rancher is concerned about its cattle.
If they can't trust electronic voting, how will they trust thier online poll?
Easily: they're the ones running it.
If you let me personally set up every computer counting votes, I would trust electronic voting too. Unfortunately this probably isn't a solution to our election problems, since other people don't seem to trust me as much as I do.