Researchers And Registrars Debate E-Voting

Paper Trail writes "There's a fascinating discussion going on right now over at SiliconValley.com. A group of computer scientists, journalists, voting activists, and county registrars are discussing the e-voting mess in an online forum that runs all this week. The panel is a who's who of e-voting: Avi Rubin, David Dill, David Jefferson, and registrars from San Bernadino and Riverside, CA. They've even got Scott Ritchie from the Open Vote Foundation. The question they're hoping to answer: "What's your assessment of the risks related to the use of electronic voting machines -- in the areas of verifiable voting, errors, recounts and manipulation -- not in the computer lab, but in a real-world setting? And how do those risks compare with current voting systems and other low-tech options?""

Comments

said it before -- I'll say it again

[Shakrai]

Would somebody please tell me what exactly is wrong with the lever operated mechanical machines still largely used in my state (New York)? The machines are sealed and verified by comparing counter numbers that are tamper-obvious. At the end of the voting day the machine is sealed by the poll workers who write the numbers down and send them to the Board of Elections, who later collects the machine itself.

They are next to impossible to tamper with (it would be glaringly obvious), they work if the electric fails (try that with your touch screen), they keep voters from overvoting just as effectively as a touchscreen does and at the end of the day they can be tallied in a few minutes. What is so wrong with the concept behind these machines that we need to all rush out and buy touchscreen systems? What advantage does a touchscreen offer? It is a closed-source solution that's infinitely easier to rig then a mechanical counting system.

Isn't this one case where we don't need to reinvent the wheel people?

Re:said it before -- I'll say it again

[Orgazmus]

You just raised the most important issue about evoting. Why the hell should we use them?

Re:said it before -- I'll say it again

[garcia]

What is so wrong with the concept behind these machines that we need to all rush out and buy touchscreen systems? What advantage does a touchscreen offer? It is a closed-source solution that's infinitely easier to rig then a mechanical counting system.

People want pretty colors and instantaneous stats. People don't want to worry about counting and recounting. People want to have the voting booth be available in 1000 different languages. People want to have their tax money spent on something that is ever-changing.

I would prefer (and basically demand) that we keep our current voting mechanisms the same. Will they? No. But that's not really something that my single voice can stand against.

Re:said it before -- I'll say it again

[Shakrai]

People want pretty colors and instantaneous stats. People don't want to worry about counting and recounting. People want to have the voting booth be available in 1000 different languages. People want to have their tax money spent on something that is ever-changing.

You can get near instantaneous stats out of these machines. And what multiple languages do you need? They see the names of the people running -- if they don't know "George W. Bush" is running the President and not the local Assembly seat that's their problem. Besides (in NY anyway) any voter can be helped by anybody else except for a boss or union official. The poll workers themselves can even enter the booth with him provided you have one poll worker from each major political party enter the booth at the same time.

Re:said it before -- I'll say it again

[JUSTONEMORELATTE]

Ok, I'll start:

• They break, and spare parts are expensive since they're not in current production
• The numbers on the counters are manually recorded, then manually transferred to a central registrar. That's two places with human intervention, and opportunity for error or, more remotely, fraud
• The manual processing takes time, and like it or not, people want to know results sooner than the morning paper.

--
What would it take?

Re:said it before -- I'll say it again

[iplayfast]

Trust to mechanical (or electrical) things always involves trust of the person behind it. There is nothing wrong with your system, except that it might be possible for it to break down during an election without the operator (voters) knowing it.

Maintainence on these machines must be certified, etc.

Pen and paper (drawing an X in the appropriate square) have worked for years, but again trust is given to the people tallying the votes.

Your system to me sounds like a better solution then the touch screens. More easily verified as working. Less likely to fail through wear & tear. The advantage I can see for your machines is the speed of tallying the votes. (Which are done as the vote is cast).

Re:said it before -- I'll say it again

[Martin Blank]

I can give you some reasons. They're not necessarily good reasons, but they are reasons used.

1. Virtual elimination of mechanical breakdown. This can be an issue with some of the older equipment. It doesn't address the electronic systems breaking down or crashing, though.

2. Rapid collection of stats. This has less to do with anything useful and more to do with people getting impatient. In most cases, the results are pretty obvious within hours of the polls closing. In other cases, we get a little tension for a few days as things come down to the wire. (In still other cases, we get a lot of political infighting for the next four years.)

3. Standardization of interfaces. I've only seen one e-voting system, so I'm not entirely sure how possible this is, but it seems to me at least theoretically possible that the presented screen can be relatively standardized across a state, at least in terms of basic layout (since county- and city-specific issues will be different, of course).

Personally, I miss the lever system that I used for about ten years. The 'ka-chunk' feeling of the ballot being marked seemed to give a tactile and auditory sensation to the emotional satisfaction of having expressed my opinion.

Re:said it before -- I'll say it again

[Anonymous Coward]

Also, they can fail to rollover from 999 to 1,000 (or similar counts) because the machine requires additional leverage to turn all of the counter wheels (just like an old analog odometer), and if the parts are worn, there might not be enough force.

So it's easy for votes to be lost due to mechanical error.

Re:said it before -- I'll say it again

[abb3w]

Would somebody please tell me what exactly is wrong with the lever operated mechanical machines still largely used in my state (New York)?

My current state doesn't use them. I used to live in NY, so I'm familiar with those machines. They were excellent. A trivial update of the design could allow electronic reading of mechanical vote tallies, if anyone cared to, while still keeping the old "seal" method for recounts. They are substantially better than the punchcard methods (used locally prior to last years touch screen purchase) or the electronic scams^H^H^H^H^Hschemes being suggested (and currenlty in use locally).

And I would say they are MORE effective than the touch screens for preventing overvoting. They give tactile feedback; you try it, and you realize the lever can't move.

Re:said it before -- I'll say it again

[truthsearch]

People want pretty colors and instantaneous stats.

When they watch TV, not necessarily when they vote. People don't care if they vote by pulling a 20 year old lever. They only care that it's quick, easy, and counts.

The masses aren't screaming for computer graphics on the windshield to aid in driving or a drive-by-wire joystick. People experiment with it, but the flashy stuff isn't what people call for in practical situations.

And making it colorful won't get more people to vote. People aren't running to the polls to check out the computers.

Re:said it before -- I'll say it again

[Zeveck]

The advantage is not to the voters, but to the company supplying the new systems. There is a ridiculous amount of money to be made by reoutfitting the entire United States with new voting technologies.

Hell...look at Diebold. They made their voting machines without a confirmable printout. Why? Just about everything else they make (from ATMs to cash registers) has a confirmable printout. But hey...look at that...now they can get paid AGAIN to go "upgrade" all the faulty machines they've already deployed. And then they can get paid again to fix the "bugs" in the machines.

Even if that is a little too cynical for you, the fact remains that the companies bringing out the voting machines are making a lot of money.

It is up to those companies to convince the public that they need and want new voting machines. It doesn't matter whether the existing technology works - they'll focus on its flaws and potential abuses and tote their shiny new products as if they are sleek and bugfree.

Create a sense of fear and then offer a remedy that appears to address it. Works in business. Works in politics. Works in just about anything really.

Re:said it before -- I'll say it again

[Shakrai]

They break, and spare parts are expensive since they're not in current production

Then put them back into production. And if they go out of production what do you suppose is cheaper? To contract some local guy in a machine shop to make a part or contract some coder to try and figure out a closed source system because you needed a replacement part and it broke something?

The numbers on the counters are manually recorded, then manually transferred to a central registrar. That's two places with human intervention, and opportunity for error or, more remotely, fraud

They are phoned into the Board of Elections by the poll workers (at least one from each party) after the polls close. This isn't the "certified" tally but it's the one that is released to the news media for the nightly news/morning papers. Once the Board of Elections receives the machine back they open it up (again with a supervisor from both political parties present) and certify the tally. With at least two people doing every task the odds of error are small -- and fraud is damn near impossible.

The manual processing takes time, and like it or not, people want to know results sooner than the morning paper.

This is why you have the unofficial count from the poll workers and the later certified results. Electronic touchscreen systems will not change this. The count isn't "official and certified" until they manually count the absentee and challenge ballots.

And as far as fraud goes -- which system do you trust more? The system that relies on two public servants sworn to uphold a scared trust or the system that relies on private vendor companies with lovely quotes like "I'm committed to delivering Ohio's electoral votes to the President". This is a no brainer people.

Re:said it before -- I'll say it again

[Shakrai]

So it's easy for votes to be lost due to mechanical error.

I could say the same thing about a touchscreen system that suffered a hard drive or non-violate memory failure. Even if you use top of the line name brand components (think Diebold is doing that?) think about a large enterprise network -- how many hard drives will fail when you start taking about thousands of systems? Now expand that scale to tens of thousands of voting machines and you start to see the problem.

Any election system we use (electronic or mechanical) needs to be inspected and tested prior to the election. To blindly suggest that these machines will suffer a breakdown without pointing out measures taken to prevent this scenario is just spreading FUD.

Re:said it before -- I'll say it again

[jbarr]

No, not like you can when you're watching Survivor or an NFL game. Remember that's what people want... Monday Night Football commentary with tickers. Flashing lights and shiny metal. Oooh.

I want to know who these "people" are. Personally, I would give up all of those bells & whistles for a system that is reliable, accurate, and fraud-resistent. Certainly, the integrity of the election process is more important than marketing fluff.

"And what multiple languages do you need? They see the names of the people running -- if they don't know "George W. Bush" is running the President and not the local Assembly seat that's their problem."

You have negated any credibility you might have had with that statement.

I disagree with your dismissal of his credibility. There are four years between national elections, and people have ample time to become informed. We need to develop a voting process that strikes a balance between reasonable accessibility and over-accommodation. Removing personal responsibility is definitely not the answer.

The most important thing in my eyes is to develop an election system that is reliable, effective, and virtually unchallengable (ie: the results are reliable enough that a challenge of fraud becomes virtually unnecessary.) Our current system is so screwed up, so in flux, so susceptible to lawsuits, so prone to varied interpretation that it invites challenge and fraud. One huge problem is that people in general are unwilling to accept a registration and validation process that would provide accurate votor validation. We can come up with all the latest and greatest systems, but unless we can accurately and reliably validate votors, the system WILL break down.

Re:said it before -- I'll say it again

[Joe Tex]

How bout this (All this is pedicated on the E-voting stuff to be handled correctly)

- Verification of ballot - no invalid ballots
- Context sensitive help, less confusion
- Possibility of more advance voting methods (condorcet, IRV, etc)

Re:said it before -- I'll say it again

[Kwantus]

The numbers on the counters are manually recorded, then manually transferred to a central registrar. That's two places with human intervention, and opportunity for error or, more remotely, fraud

Amazing -- 'cause that's how we count ballots in Nova Scotia. I would never trust a machine to do a count. How we get around human intervention/error/fraud at the count:

There are at least four witnesses to the counting: the deputy returning officer and the poll clerk, who are nominated by the two leading parties, and at least two from the public (who are usually, but not necessarily, agents of two different candidates)

All the totals get printed in the newspapers so the witnesses can check their own poll and anyone can check the sums

Your system does nothing like that. Even where the ballots are counted by hand, all the numbers disappear into Voter News Service, which then reports what it wants. (I don't know what's replaced VNS. I understand it was dissolved after it so badly botched the con in 2000. It was a pretty secret society and its successor can be expected to be even more so.)

People want to know sooner than the morning paper

I REFUSE, as a voter, to buy into the horserace psychology. That's just hype created to get the whole scam over with and out of the news cycle. I want a proper count more than I want instant falsified results.

The results don't take effect for months; why the fuck the indecent hurry?

BTW all I get out of the forum anchor is "cyclic link". I guess konqi users are locked out >:(

Re:said it before -- I'll say it again

[swillden]

And as far as fraud goes -- which system do you trust more? The system that relies on two public servants sworn to uphold a scared trust or the system that relies on private vendor companies with lovely quotes like "I'm committed to delivering Ohio's electoral votes to the President". This is a no brainer people.

This question presents a false dichotomy.

The correct answer to the question is "Neither". We should trust no one to be in a position to manipulate our votes. That ideal may be unachievable, but with a completely transparent process, we can get very, very close.

Lots of people have said this before, but I'll say it again anyway. You want a good election system? Here's how:

• Use an electronic voting machine with a pretty, easy-to-use, easy-to-understand touch-screen interface. Big letters, big buttons, voice prompts (using headphones if needed), etc., so that everyone can use it, even if they're half blind, and even if they're not the sharpest knives in the drawer. Use all of the power of technology to make it accessible to all voters.
• When the voter has made all of his/her/its selections, and confirmed them, have the machine print a paper ballot that the voter can read. If you want to electronically tally votes, fine, do it. The paper ballots will allow recounts. If you want to automate recounts, fine, use an OCR-able font, or even print a 2-D barcode or similar containing the data. Make absolutely sure that there is a complete description of the voter's votes that can be read with only the Mark 1 Mod 0 eyeball, though.
• Have the voter drop the paper ballot into a locked steel box, watched by the polling volunteers, including representatives of the major parties.
• Transport, store and open the boxes only under the watchful eyes of party representatives and any interested citizen.
• Conduct recounts as needed, always in full view of anyone who cares to watch.

It's not hard. We *know* how to securely manage paper ballots, and everyone, no matter how technically clueless, or even technologically phobic, can understand paper ballots and see how the process assures their security.

Ink on paper is better than punched cards, it's better than mechanical counting machines and it's worlds better than purely electronic vote-counting systems.

Re:said it before -- I'll say it again

[dahorowitz]

I was trained as a poll worker in NYC a few years ago (I was a student at the time, and had nothing better to do with a random Tuesday), but let me tell you, after going through the process, I have way less faith in the lever machines than I did as an innocent yet ignorant voter.

First off, as people pointed out above, the numbers are manually tallied at the end of the evening. The poll workers use a key to unlock the machine, and little windows appear with the individual tallies. These tallies are transcribed to a paper 'canvas,' and then transcribed again at the Board of Elections office. Granted, the voting machines are not reset until the election is certified, so it is possible to return to the source to review the count, but it is just as likely that an innocent error such as that caused by transposing two digits may never get caught.

More importantly, however, is that similar to touch screen voting, much of what happens with a lever machine during an election is of the Big Black Box variety. Beyond a master counter which is supposed to advance one count for every person who voted, there is really no way of knowing if the machine is functioning correctly. It is wholly possible that a counter for a particular candidate or ballot measure is malfunctioning, resulting in an undercount or overcount. More importantly, no individul record is kept of each vote (similar to the touch screen systems), so it is impossible to go back and 'recreate the voters intent' as you can with paper/optical scan systems.

Further, these machines are getting old and are notoriously unreliable. These machines jam somewhat frequently, and the 'official' method of clearing these jams is to force the red handle (the master voting control that unlocks/registers the votes for an individual voter) back and forth until the machine is operating normally. Of course, each time you do this, a new 'vote' is registered on the machine--presumably, it would be treated as a blank ballot, but once again, with the black box nature of the lever machines, one never is quite sure.

Finally, there is the people aspect. In NYC, there are so few registered Repulicans (particularly in Manhattan) that it is quite common practice for the Board of Elections to designate a Democrat as a 'Republican for the day.' In other words, a precinct that is supposed to have two Democratic and two Republican poll workers may end up having three Democrats and one Republican, or even four Democrats and no Republicans working!

Electronic voting machines aren't the problem

[Frequanaut]

It's the people who hate democracy:

See here for more

And yes, I know it's a partisan site, but it's just collecting news stories, look past the commentary.

Re:Electronic voting machines aren't the problem

[Frequanaut]

How is providing a list of stories concerning voter registration fraud moderated as flamebait? Moderation is screwed.

Re:Electronic voting machines aren't the problem

[Frequanaut]

It's not frothing. These are people actively trying to stop democracy.

Re:Electronic voting machines aren't the problem

[Detritus]

It's a very selective list. That might be grounds for considering it to be flamebait.

Recounts?

[FatherKabral]

Has anyone forgotten what happened in Florida? Granted that recount was a great big cluster, but what happens if the data is lost? No paper copy means no recount...everyone would have to vote again, and that could mean changed votes, fewer votes, etc. If it ain't broke......

Re:Recounts?

[abb3w]

they had to get on with things before January Something-Or-Other for when the Electoral College met and voted for the President.

Or, alternately, fail to certify the vote count, and not send ANY electors. Which would have caused a far bigger stink.

Re:Recounts?

[DigitalRaptor]

No, the Supreme Court didn't have jurisdiction. They were relied upon because Bush knew they would side with him.

There are specific federal laws governing elections, and more specifically the couting of military ballots. The law is, if they aren't in by a certain date, they aren't counted.

Bush and his people (namely his brother and Kathleen Harris) broke federal election laws and counted all military ballots, regardless of when they came in, to the tune of a +800 gain for Bush, pushing him over by 576 votes total.

Obey the existing laws (both state and federal) and Gore wone by nearly 300 votes.

Don't see what the fuss is about

[raitchison]

Even if the worst FUD claims of the anti electronic voting crowd are true electronic voting is no more vulnerable to tampering than paper ballot voting. Where ballots can (and are) lost (or "lost") and there are dozens of opportunities for workers to mess with or change things.

I've voted touchscreen twice and it was great, I got to vote in advance of election day (when it was convenient for me). Though there was a LOT of pressing "next page" for the CA Recall election to sort through the >100 candidates. :)

Like any new system it will no doubt have it's own issues that will need to be worked out. That's the price for progress.

What I'm waiting for is the opportunity to vote online.

Re:Don't see what the fuss is about

[greechneb]

I think what we have here in our county is a nice system. It's similar to a scantron test. you color in your choices, and when you are done, it is fed in. If there is a problem, it spits it back out. At the end of the day, it prints out a total. The judges count the number of ballots to make sure it matches the number of ballots entered. If nothing is out of line, it only takes 30 minutes or so for them to balance at the end of the day, and then take the results in. It cut down the balancing time by nearly half. If there is a recount, the paper ballots still exist. Don't know who makes the machines though... probably diebold.

Re:Don't see what the fuss is about

[truthsearch]

Even if the worst FUD claims of the anti electronic voting crowd are true electronic voting is no more vulnerable to tampering than paper ballot voting.

Wrong. In most states there are multiple eyes on every ballot from the moment they're taken out of the box until they're counted and sealed. An all-electronic vote is not usually reviewed by multiple people. That's why in test runs thousands of fake extra votes are able to be counted. If every electronic vote was scrutinized the same as paper then they'd be roughly equally vulnerable, but they're not and probably never will be.

Re:Don't see what the fuss is about

[pavon]

Even if the worst FUD claims of the anti electronic voting crowd are true electronic voting is no more vulnerable to tampering than paper ballot voting.

That is simply not true. With paper ballot voting the only people who can tamper with the ballots are the election officials, and members of all interested parties, observe the voting and tallying process. With some of these poorly implemented systems, anyone with internet access or access to the voting machine (any voter) could potentially hack and tamper with the voting results. This is not a theoretical concern either - there are proven vulnerabilities with these systems.

The lesser reason why your statement is incorrect is that with paper ballots, if fraud is suspected, you can at least go back and do some post mortem analysis of the election - recount the existing ballots, make sure that the number of people who signed in at a voting location is consistent with the number of existing ballots. With most of these electronic voting systems this is currently not an option (although it could be).

These complaints are not FUD, by any stretch of the imagination.

Re:Don't see what the fuss is about

[Daniel Dvorkin]

electronic voting is no more vulnerable to tampering than paper ballot voting

Please explain to me in detail how one person, exerting no more effort than is necessary to write a line or two of code, could alter several thousand (or million!) paper ballots at once.

Vote From Home

[clinko]

Voting from a PC at home is a bad idea because it unfairly gives people with money a chance to vote easier than without.

I guess this is obvious, but had to be said.

Probably their association with Chicago

[Anonymous Coward]

Mechanical/lever machines are associated with the Kennedy voting fiasco in the 1960 race between Nixon and Kennedy.

Simplifying greatly, the people who tabulated the votes from the lever-operated machines were pro-Kennedy. Vote tabulation was done by opening the machine up, and reading numbers off a little odometer-style readout. When the numbers were written down, the Kennedy numbers were written as higher than the machine recorded, and the Nixon numbers as lower.

However, the Democrats weren't the only people rigging that election. Downstate Republicans did their share of double-voting, including many people from conservative St. Louis hopping across into Illinois to vote in that state, as well as their home state.

Scrutineering

[lpontiac]

Any layman can look over the shoulder of anybody doing anything with the current system, and know at a glance whether the work is being performed correctly.

I have a degree in computer science, and I can't look at an electronic voting system and see that it is working in the correct manner.

This is why I don't think electronic voting systems can ever replace a manual system.

Re:Scrutineering

[mefus]

That's true: the secrecy of the ballot has been extended to cover the tallying mechanism. Disaster.

That being said, if the system were open source and included some kind of hashing mechanism to verify the binary is a product of the certified code, and the certification rules for the hardware were more strictly obeyed, evoting could be plausible.

None of this is possible with Diebold, or ES&S, or whatever.

If it aint broke....

[Kozar_The_Malignant]

Our county uses optical scan ballots, and they work fine. They are fast to count. There is a physical backup that can be hand counted if need be. They require no fancy equipment and the polling place. And, if you can't figure out how to mark them correctly, you really are too stupid to have your vote counted.

I am totally at a loss to understand this rush to some sort of electronic voting. I regard voting as the one, true sacrament of citizenship. I have no problem with it taking a little bit of time. After lying to pollsters for months, the ritual of going into the booth and casting my secret ballot is very satisfying.

Remember, voters are citizens; all others are residents.

In Alaska

[RealAlaskan]

In Alaska, we vote on bubble sheets. Fill in the oval next to the candidate you want (or dislike least). A machine reads the ballots and counts the votes, giving the instant, error-free [1] readout everyone says they want, and the bubblesheets are still there, to be audited at leisure. It seems like the best of both worlds.

It should be error-free, but, in our local election last week, the machines somehow managed to count 11 more ballots than were cast. That's where the paper ballots come in: they're human readable, and humans are auditing and handcounting them right now.

Anyone else hear about. . .

[smooth wombat]

the new electronic voting system developed by David Chaum? No? If you're reading /. no wonder! Stories of ontopic interest are rarely posted.

Here's the link to the Business 2.0 article talking about his new system which he claims is "the first electronic mechanism that ensures both integrity and privacy."

Just as long as I'm writing the voting software

[NigelJohnstone]

"I've voted touchscreen twice and it was great, "

Playing devils advocate: I could write an easter egg into the software, so that when I come along to the voting booth, I tap my finger on the screen in a few special unmarked places, and that machine then favours my chosen candidate.

Tell a few of my friends and we could easily do that with all the machines in a swing state.

The pre-checks wouldn't pick it up, the random machines taken out for testing wouldn't show the problem (because I wouldn't be activating my easter egg on those machines).

Whoever writes the voting software controls the election without the paper trail.

Its fine to make unauditable voting machines, just as long as I'm writing the software. :)

Make Election Day a Holiday!

[justanyone]

Many other nations make Election day a holiday. We should have election day as a Work & School holiday.It would solve problems:

• with too-few people voting since there's far more time to do it and less hassle;
• evening news coverage couldn't influence the election since most people would have voted by then;
• It would reinforce the idea that democracy requires attention and is important;
• people attempting to vote at the wrong precinct would have time to get to the right one;
• More professionals could volunteer to work at polling stations, which would speed vote counting and allow for disabled people to be assisted by people of both parties;
• We would get another vacation day;
• A better-educated cross section of college students and "slacker-class" (Jon Stewart's term) would vote since the ones too drunk from "no-class-tommorrow" syndrome would have too big a hangover to vote, while the nerdier non-drinkers would vote more reliably.
  • These comprise a very valuable argument:
  • Make Election Day a Work/School Holiday !
    

Validation

[homeslice3]

I think the best solution to the electronic trail is to simply have your e-voting machines two things:

When you're done voting, the record is added to the database. And the computer spits out two copies of the voting record:

One copy goes to the voter, with an outline of how he/she voted.

Another stays with the computer and is used to verify the e-voting tallies if neccessary.

Both printouts are bar coded and have a user/id pin combo that the voter can, after the election verify that his/her vote has been tallied as intended.

This way you get both a paper and electronic audit, and the voter is assured his/vote is counted as intended.