Centerfield

(same lurker as before)

what's your opinion on the bill HR 550?

When it comes to e-vote systems, there are only two groups who may be expected to aver that they trust them:
- politicians and others who know nothing about computers,
- employees of companies which sell them.

Certainly nobody who works with computers professionally would dream of trusting a systems with neither an audit trail nor any means of verification that the output matches the input.

Personally, I would suggest that your threshold of 2% for an acceptable error rate is vastly too generous. I would suggest that anything over 0.01% errors (a mere 4 sigma, rather than the 6 sigma that businesses aspire to) should be grounds for cancellation of contract and penalties. I do like your suggestion of penalties including the costs of reverting to another system -- although I think that should be only a part of the penalties, rather than the whole of them.

Paper ballots. Hard copies. Tangible records. Review available. Recounts possible.

No hardware costs. No system freeze-ups. No lost files. No corrupt files. No viruses. No remote hacking. No tech support department. No electricity. No brainer.

I would remind everybody that there's a million ways to fiddle with elections using paper ballots, too.

Remember all the spoiled "overvotes" in Florida? The allegation was that some ballots had been intentionally spoiled after the election by punching out a second vote for President on them. That's just one way. There are many, many more (hanging chads, anyone?). There's a reason elections officials have been moving away from paper for a long time. Louisiana has used the big lever voting machines for decades. Of course, they can be rigged, too. There's a joke that our old Governor, now in prison, was elected President of some central American nation after we sold them some old machines.

I'm coming to the conclusion that the best result is computer voting, like this: You pick your candidates. The screen shows you a summary of your choices and asks, in clear terms, if this is how you want to vote. If not, you can pick which races you want to change your vote in. Once the summary screen shows your final choices and you are satisfied, you press the "cast vote" button. That also causes to be printed (on thermal paper for ease of maintenance) a paper summary of the candidates you chose. You look those over and, before you leave the voting booth, deposit it in the machine. A scanner reads the barcode to verify that the ballot has been deposited. If you don't deposit the paper print-out before leaving the voting booth, an alarm sounds so that the commissioner can assist you and make sure you deposit the paper properly.

After each election, even those won by a large margin, the elections officials pick a random sample of machines from a random sample of precincts, and manually tally the print-outs to make sure that the manual count matches up with the computer-reported results.

That makes sense, in that the computer system is only being used as a printer/counter. But notice that there's really no trustworthy way to get around the need for paper verification.You still need to make a hard copy, and file it for back-up purposes.

It sounds pretty good, and might well represent an improvement,though system sounds quite expensive to implement and keep up-to-date. That makes me a bit skeptical of the value added.

And notice that it doesn't even eliminate all the fraud possibilities. People could still get into booths under phony n@mes. And even though it sounds far-fetched, it could even, for the daring, conceivably make fraud easier. You print a bunch of ballots for the candidate of your choice and then use a well-designed program to electronically check off the n@mes of people who didn't vote to make the numbers reconcile on the bottom line. It might not survive a very close look by a very skilled investigator, but why chance it? I just don't have a ton of trust in the idea of counting votes using a sophisticated system that only a small handful of people truly understands at its most basic level.

On the issue of chads, let's not forget that this system was driven by the desire to have easier counting, too. When I say paper ballot, I'm talking simple and more idiot proof, not something that requires you to use a machine to readt it, operate it, or cast it.

IMO, every presidential election should be voted upon using a very simple separate, every-state-roughly-standard ballot with each qualified ticket listed, a box to check off, a write-in line, and that's it.

Pat,
Surely there are lots of ways to fiddle paper ballots. But at the end of the day, if you have a paper ballot you have a chance (whether a good one or a not so good one) to look at the ballots and perhaps detect a fraud. With a strictly electronic system, you have nothing to look at. To take the simplest possibility:
- the system fudges the results.
- at the end of voting, the computer system erases the program that fudged the results and (re-)loads one which would count correctly.
Result: false results, and nothing to show that the program which falsified the results was ever there.

I'd really rather have a method which gives me a chance (even if only a slim one) to spot fraud than one which gives me no chance.

I agree, wj. I point out the possibilities for fraud with normal paper ballots to show why we began looking at electronic ballots to begin with, and why we shouldn't give up trying with them.

I used to be opposed to paper print-outs with electronic voting, because I see potential logistical problems there, too. What if the machine runs out of paper. Am I still allowed to vote? If I am, then suddenly the machine tally and the printed tally won't match.

Brian, the problem with using a separate piece of paper ballot for Presidency is the logistics get very complicated very quickly. Not only is there confusion (I have to vote again by marking this piece of paper? But I just voted in the machine for all those other races...), but that's an awful darn lot of paper, combined with all the other paper ballots for all the other races.

And we saw in Florida the significant potential for uncertainty deciding what constituted a vote even on what amounted to plain old pieces of paper. As somebody suggested elsewhere, if you write in "the governator" in California, is that a vote for Arnold? If it's a check-the-box system without allowing write-ins (remember, somepeople in Florida both punched a chad and wrote in a n@me), what if the voter puts a check-mark in one box and an x in the other?

wj... mentioned to finish my response to you by saying that the concerns you express, having been uttered by enough people, have convinced me that, even if the electronic system is in fact 100% trustworthy and tamper proof, a significant mass of people in this country will still never trust it, because its trustworthiness cannot really be fully proven. So I changed my position and now support a paper verification (which is never allowed to leave the polling place).

To PatHMV at November 13, 2006 10:23 AM

Does this sound like your comment? (i'm not very bright, as you can tell):

The Voter Confidence and Increased Accessibility Act (H.R. 550) will:

* Mandate a voter-verified paper ballot for every vote cast in every federal election, nationwide; because the voter verified paper record is the only one verified by the voters themselves, rather than by the machines, it will liserve as the vote of record in any case of inconsistency with electronic records;
* Protect the accessibility requirements of the Help America Vote Act for voters with disabilities;
* Require random, unannounced, hand-count audits of actual election results in every state, and in each county, for every Federal election;
* Prohibit the use of undisclosed software and wireless and concealed communications devices and internet connections in voting machines;
* Provide Federal funding to pay for implementation of voter-verified paper balloting;
LINK from Holt website

An open letter to voters clarifying H.R. 550
From Rep. Rush Holt (D-NJ)
LINK

it had 219 co-sponsors
(about 14 R's)
; in your opinion, should this bill be re-introduced?

I think we are in agreement, Pat. Let me commend to your attention the optical scanner ballot that is used here. The voter gets a paper ballot on arrival at the polls, just like any paper ballot system. The voter marks it with a pen, and inserts it into a counting machine.

The voter knows that what is on the paper is what he really voted for (a potential weakness in systems where he never has the actual paper physically in his hands). The results are recorded electronicly, for rapid counting and reporting of results, and the paper is available for recounts or certification.

The down-side? You need to take care to fill in the bubble on the ballot, not just put a check mark in. Fortunately, we get broad-tipped flow pens, so even shaky hands can manage it pretty well.

Lurker,

I vaguely recall that bill from a year or two ago. I think at the time I opposed it because (as I noted above) of my then-opposition to the paper backup (to which more thought needs to be given... I don't think the statute designers have contemplated all the potential problems with the paper backup yet).

Generally, I suspect I am in favor of it now, without getting into too many of the details. One thing I am adamant about is that the paper verification can NOT be allowed to leave the voting booth. There can be no chance that the voter is allowed to carry away from the polling place any piece of paper (or electronic code) which could be used to verify how that voter voted. As long as the paper verification stays at the polling place in a secret ballot box, that's ok with me.

Simplicity and audit trail. I find the optical scans to be the most efficient method. The forms are clear and usually easy to fill out. The only computer and power needed is the ballot box itself. When you have touchscreens, you also have to worry about keeping them powered. No power, no voting. At least it is easier to handle a no power system with optical scan. Add in the fact that the majority of poll workers are not the most tech savvy people, you really need a simple system for them too. I have seen the "calibration" needed for the touch screens. It is way too easy to be fouled up. Plus they are overly expensive.

Voting is not that difficult a situation that we need to go overboard on the technology. The more complex a system gets, the more there is to break. Just look at the most complex system we have ever made, the space shuttle. Voting does not need complexity. Even if it takes a little longer to count the vote, the process should be more focused on getting it right and making it simple. KISS.

Pat, I agree. The slip has to stay. It makes no sense for it to leave. What good is the paper if it is not left behind for an audit trail. A printed piece of paper is no more assurance that the machine recorded the vote right than an "I voted" sticker. For the purposes of recounts, the paper printout should be considered the official ballot.

I read the comments on the hyperlinked story ("whoopsie"). This comment is telling:

Wait a second this is all digital - THERE SHOULD NOT BE SAMPLING ERRORS!.
Statistics has nothing to do with this - or else you will find that 3+2 = 6 some times and 4 other times. On average you'd still get 5 but...

And this IMHO is the heart of the matter. Statistics has EVERYTHING to do with it. And yes sometimes (from a statistical standpoint) 3+2 can equal 6. We will not feel good about any method when the result is within the margin of error. And of course, we will only question the value of any method when the election result is close and by definition "within the margin" of error. Its like measuring microbes with a yardstick.

PS And as someone correctly pointed out in the discussion at the linked story missing 1 vote in 36 doesn't mean that you'll miss say 100 votes in 3600 or 1,000,000 in 36,000,000.

PatHMV at November 13, 2006 01:44 PM

Thank You for your response!

"the paper verification stays at the polling place in a secret ballot box"

according to the bill, that's what would happen. the ballots would then be used for recounts.

I know folks may not believe it, but we've NEVER had anything like "100% trustworthy and tamper proof" voting, or even close, and likely never will. You're dealing with people, not robots. Voters make mistakes, and sometimes screw things up on purpose.

Over time the trend is towards less error and less fraud. But each and every method will have its own vulnerabilities, and provide both means of error and opportunity for fraud. Absentee ballots sound good, but provide an easy way to "steal" votes in bulk. The nursing-home scam is considerably older than I am, as are the ACORN change-of-address registrations and the "one party only" registrations bit. Also poll-packing, roll voting, zombies, etc. All are separate challenges.

Every method will have its own holes.