Should voting be high-tech...or low?

[thatcrazycajun]

So, a week from tomorrow fully one-third of US voters will be casting their ballots electronically, using machines made by Diebold, a firm with a documented bias toward the Republican Party, with no paper trails to verify votes and security systems that are still evolving at best. Nervous? Johns Hopkins University comp-sci professor Ari Rubin is, and he tells us why in a BusinessWeek Online op-ed published today. Read it here and then discuss: Are these newfangled touch-screen doodads any better or safer than those old, clunky mechanical monstrosities they used to have to lug out to the voting precincts (and still do, in some places)? Or should we just say "hellwiddit" and go back to the old reliable No. 2 pencil?

Comments

The Canadian System. Takes Longer. Works Better

[shelleybear.livejournal.com]

Paper ballots.
Representatives from each party at each polling place counting the votes at the end of the day.

Re: The Canadian System. Takes Longer. Works Better

[dglenn.livejournal.com]

"Takes Longer. Works Better"

From what I've heard in other fora where this has come up, "longer" is a matter of only a few hours. Seems hardly worth mentioning. Am I misinformed?

("Works better" is, well, unsurprising given how screwed up things are here right now. Electronic machines may well ultimately work better once the bugs are shaken out ... or it may turn out that "as good as" is as much as is possible. Personally I think a combination is best: electronic machines to do input validation and handicapped-voter assistance at the polling booth, which print out voter-verified paper ballots that are the official ballots of record. The electronic record in the machine serves as a check against old-fashioned paper-ballot methods of cheating, and the paper ballots serve as a check of the integrity of the electronic system. Designed right, the techniques required for fooling both simultaneously should be different enough to make successful fraud unlikely. But we won't have that for this election, and I'd be happier going back to paper-only temporarily, until Analysts Who Have A Clue can design a proper electronic voting system. There's no shortage of clueful analysts; they're not not who got hired by the companies that were awarded voting machine contracts so far. *grumble*)

Re: The Canadian System. Takes Longer. Works Better

[shelleybear.livejournal.com]

I'd be open to the idea of a dual system until we got ALL the bugs out of the electronic one.
And yes, I LIKE the idea of handicapped-voter assistance.
We need to run out elections on the K.I.S.S. principle.

[faxpaladin.livejournal.com]

Forget security for a moment; the machines used here (dunno who makes them) have a design flaw: If you cast a straight-ticket ballot, it goes straight to the verification screen, skipping any propositions/bond issues at the bottom of the ballot. If you know where to look, you can go back to them from the verification screen (there's still another step before your ballot becomes final), but still.

This is causing some consternation for the local community college, which has a bond issue on the ballot this year. They've started taking out ads in the paper telling voters how to find them on the electronic ballot. You didn't really need to do that in the paper-ballot days, unless you were running a write-in campaign or some such...

[mneme]

You need a paper trail. Or, at the very least, -some- external validation that isn't easily compromised. The Diebold machines...don't have that. Interface issues are another whole big pile of fish.

Foxtrot had a good strip on the voting machine issues here:

http://www.gocomics.com/foxtrot/2006/10/29/

[zsero.livejournal.com]

Paper. If I were designing the system I'd have smart machines and all that good stuff, but what they'd do when the voter was done is print out a paper ballot which the voter would then fold and deposit into an old-fashioned ballot box. The only function of the fancy software would be to do error checking, avoid the problem of ambiguous hand writing, scribbles, faint marks that counters then have to decide whether they were intended to be votes, etc. The output ballots could be machine readable for a quick count as soon as the election's over, but also countable by hand later. And my rule for recounts would be - keep counting until you get the same number twice in a row.

Meanwhile, while the machines are wide-open to potential fraud, actual fraud is going on, and to stop that we need photo ID to vote, inked fingers, proof of citizenship to register, and fingerprint verification of absentee ballots.