Feds to Recommend Paper Trail for Electronic Votes

flanksteak writes "The National Institute of Standards and Technology is going to recommend the decertification of all electronic voting machines that don't create paper records. Although it sounds like this recommendation may have been in the works for a while, the recent issues in Sarasota, FL (18,000 missing votes) have brought the issue a higher profile. The most interesting comment in the story comes near the end, in which the author cites a study that said paper trails from electronic voting machines aren't all they're cracked up to be."

Re:Paper records

[Anonymous Coward]

>But if a paper copy is given to the voter, then lies are caught

Um, no. The part of the program that prints the paper does not need
to match the part that records the votes. Check out the HBO film
"Hacking Democracy" for an example of this. http://www.hbo.com/docs/programs/hackingdemocracy/ [hbo.com]

As a matter of accounting....

[3seas]

It seems to me that ATM machines get a great deal more use and account for a great deal of money. They give a receipt and also have an internal paper log. It takes two people to open one up legally ....

You'd think this was new technology in light of the voting machine problems.
But ATMs have been in use for at least a quarter century.

Re:Paper voting!

[QuantumRiff]

Remember the vote for leader of Iraq that Our government bragged about? The one with the election that we setup to be fair and honest? People voted by dipping their thumb in ink, and then voting for the canditate by pressing their thumb next to the candidate.. No duplicate voting, since its obvious when someone walks in with a purple thumb.. Simple, effective, and fair... And people were told to "Twist" their thumbprint to make sure that the print wasn't readable..

San Francisco machines are the best IMHO

[geneing]

I think that the voting machines that San Francisco uses are the best design so far. We get a paper form on which you mark your vote (by drawing a line next to your choice). If you are voting in person, you feed the paper into a machine that 1) Alerts you if you overvoted or didn't mark anything 2) Counts the votes and 3) Stores the marked paper ballot.

This system has all the benefits: the preliminary results are available immediately from the electronic machine, there is a complete paper trail, you know if the machine couldn't read the ballot, and absentee ballots look exactly the same as the ballots in the precinct. Why isn't this system used everywhere?

RA storage vs WORM type storage

[Odinson]

'He pointed to a system devised by Ted Selker, co-director of the CalTech-MIT Voting Technology Project. "The state of the art systems aren't even on the market."'

Warning RANT!

Then the people creating the current systems should all be fired. What kind of computer scientist doesn't understand that with any random access storage there is a risk of accidental or intentional destruction or alteration, at any time, in a random fasion. That's why it's called uhh random access. Hello? This is like a CS 101 second week quiz question. They even still call it RAM!

Any write once technology will be infinately better. Which one is academic. You can use a variety of write once technologies with a diverse amount of write confidence levels, number of rereads possible and techniqiue used, and cost. Just write the votes at they happen, in a sequential fasion, in a way that you cannot backtrack and rewrite.

• a dot matrix printer?
• a laser printer?
• a cdrw?
• a writable dvd?
• a WORM tape drive?
• Sevral of the above?

Why the hell are do Sarb-Ox and Hipaa require worm tape and encryption in many cases, yet our voting systems have nothing but the seat of their pants.

As an aside Bruce Schneier [schneier.com] chimed in on this recently. I wonder if this had any effect on NIST's comments.

optical scan is the real deal

[The Monster]

Currently I lean toward optical scan, filled in by the voter and not by machine, with readers set to reject invalid ballots with helpful error messages ("Looks like you voted twice for Congress")and trigger a shred-it-log-it-replace-it procedure.

This is pretty much what we do now in Kansas City, Kansas, after we replaced the old machines (in both senses of the word). There are so many advantages to this method:

• Since the time each voter spends at the optical scan machine is just a few moments, there is no waiting while the person behind the curtain flipping levers on a mechanical machine works through all the races on the ballot. We use small folding tables with privacy shields, all of which can be folded up into a very small space for storage between elections.
• The machine can, as you suggest, catch overvotes and alert the voter and election officials to allow the ballot to be redone.
• In the event of a failure of the scanner, or to randomly audit to prove there's no hanky-panky with its software, a hand count can easily be done, without any damned chads to fall out during the process. Ballots can be temporarily held in an old-fashioned locked box, then transported to the county courthouse, where another machine can be used to count all the ballots en masse.
• Absentee and provisional ballots can be executed on the same form as regular ballots, but processed as appropriate, reducing costs and eliminating a possible source of problems.

For visually handicapped voters, all that is needed is a computer that can give braille, or headphones for oral feedback, and a printer to print the selections onto the same size paper as the other voters use. If it's set up correctly, this printer can print the regular ballots as well, allowing the election workers to begin the day with a smaller number of pages preprinted, and only need to print additional forms if turnout is fairly heavy.

Yeah, well...

[jd]

Ballot boxes go "walkies" all the time, which is important if you have specific districts that are likely to vote in opposition to however you happen to feel. As such, there needs to be much more security on such stuff before I'm willing to take it seriously.

One thing I'd like would be for the electronic machine to generate a cryptographically-secure hash generated from all the votes cast on it. The paper ballots can then be electronically scanned and the same hash algorithm applied to the scanned data. If ALL votes are present and unmodified, then the hashes should be the same. Provided there is no collusion between the voting machine and the scanning machine makers, the probability of the hashes coming out the same in the event of vote-tampering of any kind should be extremely low.

However, knowing that tampering has occurred doesn't solve the issue of what to do about it. I'd simply insist on the election being re-held until all districts came back clean from tampering. Oh, and all sports, adult and cartoon channels would be legally required to stop transmitting until everyone bloody well voted and/or adjudicated honestly. Also, anyone caught attempting (or practicing) voting fraud should be compelled to buy everyone the DVDs of the shows they missed, before being locked up in a psych ward in Romania for the rest of their unnatural life.

Here's a system that would work

[netchipguy]

Computer prints you a receipt with a random number. Your vote is kept with your random number (electronically, plus internal paper roll in case electronic copy is FUBARed). You can check your vote, using the random number, on the web or via a bulletin board at the voting place the next day. Poll workers count the total number of voters at each site, which should "exactly" match the number of records on the web/bulletin board.

Can anyone think of a way to cheat THAT system?

It seems to be able to handle extra votes, dropped votes, and changed votes.

Re:electronic voting idea

[wheelgun]

The problem with this and with any sort of receipt is that it destroys the anonymity of the vote. Someone could be pressured into revealing their ballot reciept or the information required to access it. You can bet your bottom dollar that union thugs and similar underworld types would take advantage of a receipt system, if one was instituted.

Re:Paper voting!

[sn00ker]

Or you could use a pen and an optical scanner like the entire state of New Mexico did.

And like New Zealand does. Sure our voting population is only about two million people, but we get voter turnout in excess of 80% and still have results for most electorates within six hours of the polling booths closing. The official result takes somewhat longer, once all the special votes (taken at hospitals and prisons, registered before the election by persons out of the country on the day, or made outside the country at a diplomatic or military post) are tallied, but it's a rare race that is so close that those results change the preliminary outcome. Given that a lot of polling booths in the US take only a handful of votes, why is it so unthinkable that OCR could work? Even the major inner-city polling stations wouldn't take more than a few thousand ballots individually, and that's no different to polling booths here.

What Problem are We Trying to Solve?

[vtcodger]

Seems to me like Americans have this cultural thing that causes us to believe that complex, untested, technologies must be superior to the old fashioned way. When I was a kid during World War II, I used to thumb through the back pages of Popular Mechanics where they had pictures to these neat weapons that would surely bring the axis to its knees. Sixty years later, we STILL can't build usable versions of some of those things.

If the problem is that people make mistakes in counting, mark and scan technology should produce better results. If the problem is votes from dead or imaginary voters, how can any technology help?

If there is, as I suspect, no real problem at all, why the hell are we stumbling around with all this half-baked technology?

Paper Ballots: Shorter Lines?

[rickkas7]

In my little town in Vermont there was no line to vote because there were probably 20 completely private booths and 40 semi-private counter spots in which to fill in the little bubbles on the optical scan card with permanent marker.

All other issues aside, there is no possible way we could afford anywhere near that many touch-screen machines. Even barring technical problems this is bound to cause a bottleneck as people ponder their vote.