Open Voting at OSCON

fmclain writes "The Open Voting Consortium (OVC) which has already been mentioned here will be demonstrating its open source voting system, which includes a voter verifiable paper trail, at this year's OSCON in Portland. The Mercury News (free reg.) describes this as the touch-screen holy grail. Given Diebold's troubles in California this can't come too soon. The OVC has already demonstrated a working system in Sacramento."

Comments

The Same News We Have Already Heard

[pholower]

This is exactly what the voting system needs in order to become more accurate. A multi-system machine. Or multiple machines to record the votes. The primary machine to do it in real time through a computer, and the secondary machine to validate the votes to prevent fraud.

Of course, that is exactly what I said here as well. But that didn't fly to well with the slashdotters then either.

Re:The Same News We Have Already Heard

[AuMatar]

What I've been saying for a while now is we need:

1)Multiple ways of counting the votes (electronic, paper, OCR). One way must be paper.
2)Different groups doing each count
3)All methods MUST be counted
4)All counts must agree within a small percent error, and the percent error must be less than the margin of the election. If they do not, revote.

Who Gets the Profit?

[Mikkeles]

If it isn't a crony corporation of the government, can this even fly?

Re:Who Gets the Profit?

[muel]

More importantly, who will adopt this 'improved' electronic voting system? Very few. Potential consumers:

1) states/regions that have already spent millions on Diebold machines,

2) states/regions that don't have the budget to overhaul the paper voting systems already in place.

The groups seeking electronic 'improvement' have, for the most part, already tanked their money into Diebold systmes, so you'll be hard-pressed to find a city council / state legislature in those respective areas that will willingly devote more of their budgets to MORE electronic voting machines. Constituents in these communities won't stand for that kind of spending because the information about the faulty machines has been kept too well under wraps to raise popular concern.

Ultimately, however, if these machines can get into even one voting district in this nation in the place of Diebold, then I'll count it a success. However, I doubt the machine producers will feel that giddily about such a small profit margin.

Voter registration fraud

[persaud]

Is at least as important as a paper trail. Since we already have systems that have a paper trail (i.e. paper ballots), computers could be better used to improve the accuracy and reliability of the voter registration process. This would reduce tampering by hostile insiders *and* outsiders.

This is the perfect market open source

[91degrees]

Traditional software companies hate open-source software because no one owns it or collects royalties for it.

sigh... They really don't get it. Unlike Windows XP, or Adobe Photoshop, voting software requires very limited runs, and typically needs to recover its cost on its first sale. There's no need to make revenue on a per copy basis. There is probably only going to be a single customer who will have precise demands. If it was closed source, the amount of work would be the same, and the amount and so that you could charge would be the same.

Companies really need to get over the idea that because code costs money to produce, it must have value. Sometimes it is the case. Often it isn't.

Re:This is the perfect market open source

[cybermace5]

Exactly, it's not like a voting system is going to contain any previously unknown revolutionary concepts in the code. Get input, store input. It should be pretty straightforward, and the only reason people want it open is so that they can make sure there aren't mistakes.

Re:This is the perfect market open source

[laird]

"voting software requires very limited runs, and typically needs to recover its cost on its first sale. There's no need to make revenue on a per copy basis. There is probably only going to be a single customer who will have precise demands. If it was closed source, the amount of work would be the same, and the amount and so that you could charge would be the same."

I agree with this point. I'll also point out that the _real_ value that a voting system vendor provides isn't the software system, it's the complete service. The states don't need someone to provide them software -- the states need someone to provide a complete solution, from training election workers to providing on-site technical support. Compared to the cost of providing tech support to thousands of polling locations, manufacturing and shipping hundreds of thousands of voting machines, getting your product certified in each state, etc., the cost of the software development should be fairly minor.

Re:Questions...

[happyfrogcow]

well at some point you probably have to have real life security. maybe that means a certified, then locked in a vault until voting day computer, or a computer that boots from a trusted source like a certified and locked in a vault until voting day bootable cdrom. i dunno... i just think there has to be human security checks as well as technological security checks. you can't rely only on one.

Re:Whoop de do

[koreth]

Until pressing a button is as secure as writing (or punching) your vote on paper and dropping it in to a box, e-voting won't be mainstream.

So India's 100% electronic general election, underway as I type this, is just a figment of South Asia's collective imagination? How much more "mainstream" than the entire electorate of a democracy three times as populous as the US can e-voting get?

Re:Thank you.

[fmclain]

Thank you Diebold! That's what we intended all along. We appreciate your cooperation.

I hope you intend to open source version 1 as well.

Re:Um.....

[steveha]

If we have a voter verifiable paper trail, that means a vote can be traced back to the person.

It helps to read the article. Go ahead and read it now; I'll wait here.

The computer records the voter's choices, and then prints out a paper ballot, which includes a bar code. If you are not blind, you inspect the ballot with your eyes. If you are blind, you can take the ballot to a bar code reader, and put on headphones, scan the barcode, and listen as it reads back your votes to you.

The vote can't be traced back to the person, because the person verfies the ballot at the polling place, and then deposits the ballot in the ballot box. Since the voter doesn't write his or her name on the ballot, or any other identifying information, it's exactly the same as current paper-based systems of voting.

Note that if you try to steal the election by tricky programming in the poll computer, the inspection of the ballots reveals your plot. If you try tricky programming of the official ballot-counting computers, you can be found out in a recount with different computers.

This system is way better than a black-box "just trust us" e-voting computer.

steveha

Exactly Appropriate

[thelen]

When it comes to something as critical to the welfare of the public and to our form of government as the assurance of fair elections, open source software should be encouraged vigorously.

Software does not become more secure by hiding the sourcecode, and election results are not made more secure by entrusting the results to a corporation. These facts, compounded by the rampant infiltration of corporate interests in the US government, and, at the same time, the vast amount of public scrutiny sure to be given an open source voting system like this one, make the choice IMO a no brainer.

Paper votes aren't always secure either

[feepcreature]

Until pressing a button is as secure as writing (or punching) your vote on paper and dropping it in to a box, e-voting won't be mainstream. You can't hook up a wire to a box to change all the votes inside can you?

True, you can't change paper votes by wire, but there are lots of traditional methods for interfering with paper votes:

• replace the ballot box with "one I prepared earlier"
• steal the box altogether
• manually stuff lots of extra votes into it
• nobble voters
• register extra voters
• don't register some real voters
• impersonate real voters (especially dead ones still on the register, or sick or apathetic ones)
• etc...

A fair and free vote requires confidence in the mechanism, but also in the count, and the officials, and the register, and lots of other parts of the process.

In some countries, hacking electronic machines might be one of the harder ways to steal an election :-(

Re:Paper votes aren't always secure either

[pangian]

In some countries, hacking electronic machines might be one of the harder ways to steal an election

Its interesting that you mention this. A year ago I was working with groups trying to ensure the integrity of [non-electronic] voting in Nigeria. My boss, who was also working on the project, mentioned that in his conversations with Nigerians many expressed greater confidence in a computerized election.

The argument could be made that this is due to naivite and blind faith in computers... indeed it's telling the many of the greatest skeptics of e-voting in the U.S. are the people who know the technology best. But it is also true that in many places elections are most likely to be stolen at the local level. Computerizing elections may in some ways make it more difficult to alter election outcomes at the polling station level. By requiring a larger conspiracy to affect the outcome they may make some sorts of fraud less likely. Not saying there aren't plenty of problems with e-voting as it is being implemented in the U.S.(in fact I've said the opposite on several occasions), but it is an interesting point.

Or...

[mdfst13]

Or we could just go back to older methods, e.g. paper ballots. These may be harder to count, but they are absolutely reliable.

We do not have time to make the current machines have valid paper trails without sacrificing either security or anonymity, since their printers suck.

Paper Audiot? Even Better...

[Anonymous Coward]

Has anyone ever considered that we can bipass the dangers of electronic voting entirely by using electronic voting machines only as a rough estimate? There is no reason why a complete and total hand count can not be required. Elections do not need instant results (only the Network News channels need "instant results" so that they have something to report). There is no reason why votes cannot be counted by hand with human supervision at each polling place, results combined and processed only after they have been hand verified. Sure, use OCR scanning or something similar to give an instant rough estimate, but don't ever use those totals for a definitive result. Democracy is too important to allow any black box to be involved.

Liberty Day

[otis wildflower]

This is neat and all, but IMHO what this country really needs is a new holiday, the first tuesday in November, so nobody has any excuse not to show up and vote.

Granted, I probably won't vote for the president, but I may vote in my local congressional and state/local races, if I can get home from work in time.

Then again, what this country needs is a TON more holidays ;)

Is anyone else as freaked out about this as I?

[LS]

Any American who truly believes that democracy is highly important to this country should be worried about the trend in voting systems. The ballot box is where the rubber hits the road in a democracy. It should almost be sacred in a democracy. It should be easy to understand it's operation, and it should be implemented completely without involvement from special interests.

I think it's almost ABSURD that a closed-source partisan company is building the ballot boxes. Even if there is no malicious intent, the system is totally open to malicious intent in the future.

This is not a technical issue, it's an idealogical one.

LS

Re:Smart cards!

[zcat_NZ]

Errr.. what?

Any kind of voter-verifiable trail needs to be simple enough that the ordinary person can understand and trust it.

Barcodes, encryption, etc all fail this test, no matter how untamperable they might be. If you want the paper trail to be machine readable, you want a list of names in plain text with a big black machine-readable DOT next to the name they voted for.

A human-readable paper vote, placed into a locked box, and counted under the scrutiny of multiple volunteers is the only system of voting I'll ever trust.