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.

Re:The Same News We Have Already Heard

[pangian]

2) Different groups doing each count

This used to be done by having representatives from each party there for the vote counting, corroberating the results. In some countries, in addition to party monitors, independent non-partisan groups check the vote count. In the U.S. however, we have been lulled into trusting the vote count and so as far as I know these efforts haven't been organized recently. Now electronic voting machines that don't produce any sort of auditable trail prevent citizens from exercising this level of oversight should they desire to. There are a few groups talking about non-partisan election monitoring this November. I'm aware of VoteWatch, and perhaps the League of Women Voters and the ACLU will organize monitoring in particularly vulnerable districts. Is anyone aware of other efforts?

Whew...

[Anonymous Coward]

Open Voting at OSCON

For a second there I thought it said Open Voting at SCO.

Paper trail

[7Ghent]

Of course, the most important aspect of this system is that it creates a voter-verifiable paper trail and thus more accountability.

This is all implemented on a state level. Call your local representatives NOW. This is something you personally can get involved in. Chances are, particularly if you live in a backwater state like I do, that your state senators have never heard of open source. It's your responsibility to educate them.

If you wanna make sure your vote doesn't get hacked, get involved!

Re:Paper trail

[persaud]

The most imporant aspect of the voting computer is that it generates paper?

Maybe we should have computers count paper instead of first counting votes and then generating paper.

A real improvement in accountability would be a computer system that audited the *humans* who audit the *process*.

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.

Re:Who Gets the Profit?

[persaud]

Diebold is one animator away from being a cartoon villian. Gerrymandering is at least equally destructive to democracy. From America as a One Party State:

"... We are at risk of becoming an autocracy in three key respects. First, Republican parliamentary gimmickry has emasculated legislative opposition in the House of Representatives (the Senate has other problems). House Majority Leader Tom DeLay of Texas has both intimidated moderate Republicans and reduced the minority party to window dressing, rather like the token opposition parties in Mexico during the six-decade dominance of the PRI.

Second, electoral rules have been rigged to make it increasingly difficult for the incumbent party to be ejected by the voters, absent a Depression-scale disaster, Watergate-class scandal or Teddy Roosevelt-style ruling party split. After two decades of bipartisan collusion in the creation of safe House seats, there are now perhaps just 25 truly contestable House seats in any given election year (and that's before the recent Republican super gerrymandering). What once was a slender and precarious majority -- 229 Republicans to 205 Democrats (including Bernie Sanders of Vermont, an independent who votes with Democrats) -- now looks like a Republican lock. In the Senate, the dynamics are different but equally daunting for Democrats. As the Florida debacle of 2000 showed, the Republicans are also able to hold down the number of opposition votes, with complicity from Republican courts. Reform legislation, the 2002 Help America Vote Act (HAVA), may actually facilitate Republican intimidation of minority voters and reduce Democratic turnout. And the latest money-and-politics regime, nominally a reform, may give the right more of a financial advantage than ever.

Third, the federal courts, which have slowed some executive-branch efforts to destroy liberties, will be a complete rubber stamp if the right wins one more presidential election.

Those federal courts? They have a little something to do with copyright law (see other stories on Slashdot today).

I'm all about some technology

[thebra]

but there are so many ways this can go wrong. I think that we should just stand in a big crowd and raise our hands.

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.

Has OVS attempted to get their system certified?

[David Hume]

The Federal Election Commission has a FAQ About The National Voluntary Voting System Standards. The FAQ indicates that to meet the standards, an election system must satisfy either "FEC's voting system standards" or pass tests "by independent testing authorities (ITAs) designated by the National Association of State Election Directors."

The National Association of State Election Directors has, among other things:

(1) a List of NASED Certified Systems;

(2) an Updated List of NASED Certified Systems; and, most importantly,

(3) an Overview of the Certification Process.

Has the Open Voting Consortium made any attempt to get their software certified?

hack our way to freedom

[Doc Ruby]

It's too late (200 days left) to manufacture new voting machines for the 2004 election. We're probably stuck with the touchscreens peddled by Diebold, SIAC, Sequoia and ES&S, all of which have had severe field problems (some of which seem deliberate). But we've got enough time to install OSS voteware on these machines. And to test them as deterministic, reliable and accountable by paper trail.

Now you can help, in standard "Open Source Community" fashion. Email stories about this OSS voteware, and the serious problems with the proprietary voteware it replaces, to your local newspaper, TV station, and elected representatives. Keep your tone serious, professional, and no-nonsense about your intolerance of votefixing in the status quo. You have about 75 days left in which to be heard - after that, there's no time to do anything but whine. And soon after that, even whining will be out of the picture.

Re:Questions...

[ignorant_newbie]

>Wouldn't it be very easy for someone to patch
>the software in a bad way and recompile it
>before installation?

yeah, because of course no one's worked out a way to tell that a binary is the one you think it is <cough>checksum<cough>, since this whole open source thing is so new that no one's ever installed it in a security critical place before <cough>nsa<cough>.

you're right, we should run out and install windows right away, since we can trust billg to tell us that our systems are safe.

Re:Questions...

[Kenja]

"Wouldn't it be very easy for someone to patch the software in a bad way and recompile it before installation?"

Wouldn't someone be able to do this with a closed source app as well? Closed source is not the same as no source.

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:Questions...

[ignorant_newbie]

>I assume this has been thought of already,
>but I can't figure out how to prevent
>that kind of danger.

1. design the system to run from a cd ( knoppix ?).
2. have the bios checksum the cd during boot, display te result on the screen
3. the poll-workers verify that the check-sum is correct
4. profit!

Modest proposal: Run it on Diebold's hardware?

[Ungrounded Lightning]

One of the arguments against adopting this in some states is that they've already dumped a bunch of cash on proprietary systems, notably Diebold's.

But Diebods's system appears to be based on a hardware/OS platform that, at its core, is Wintel. No doubt the same is true for many, perhaps even all, of the others. (Even if they're not, Linux and the GNU toolset already has ports to many other processors/platforms, including essentially all commonly available current-generation processors.)

Perhaps it might be possible to port the Open Source voting software to the Diebold and/or other voting machines that have already been purchased?

The bulk of the machines you need are the ones in the booths. Plug an off-the-shelf printer into a Diebold and you're all set there. (No security issues on the printer itself, beyond making sure it's working.)

For the remainder, you only need one (plus maybe a spare) with a working OCR reader, sound card, and modem - for the blind readback and the uplink scan. Put that scanner on the exiting voting machine with the modem (as Diebold does on one of the machines for doing the final uplink to the state's database). Or put it on a cheap desktop, since the touchscreen is not necessary.

(Heck: Put the software for THAT machine on a bootable CD-ROM and you don't even need a special machine. Just borrow one from the school library for election day. Even if some BIOS-based malware managed to get activated and save the data, there's no confidentiality issues with what is on that machine. Any corruption of the data by malware would be detected in a manual recount, just like corruption in any other part of the total system.)

For future instalations you could go with generic touchscreen systems - or stick with the major vendors if their prices come down into the sanity range or if you want to pay a premium for ironclad hardware (like byers of "True Blue" PCs from IBM). The voting machine vendors could even make money as vendors of ruggedized commodity hardware if they don't have to maintain all that proprietary voting software.

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.

Geek Voters

[LordHatrus]

This might be enough to make millions of geeks to leave their CRTs and LCDs, and go vote.... just to check out the new system :)

Re:Um.....

[Mikkeles]

No, because the paper ballot is deposited into the ballot box after the voter receives and verifies it.

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.

Re:The CA problem was garbage in - garbage out

[ivanmarsh]

That's NOT what happened in California.
The software itself didn't work.
Diabold has now admitted it didn't work and apologized for it.
California should sue Diabold.
Cruz Bustemente should sue California and the governator should be recalled.

Diebold *BLOCKED* in California

[KefabiMe]

Article Here

http://www.washingtonpost.com/wp-dyn/articles/A3 4424-2004Apr22.html

Pretty much, California's Voting Systems and Procedures Panel decided by a UNANIMOUS vote of 8 - 0 to block counties from using Diebold machines for the November elections.

I'm normally very cynical when it comes to politics, but it's nice to see my state get (somewhat) of a clue.

Re:Diebold *BLOCKED* in California

[KefabiMe]

Sorry for replying to my own post, but I forgot to quote this nugget from the article:

Panel members also recommended that a secretary of state's office report released Wednesday, detailing alleged failings of Diebold in California, be forwarded to the attorney general's office to consider civil and criminal charges against the company.

"I'm disgusted with the actions of this company," said panel member Marc Carrel, an assistant secretary of state.

Re:Diebold *BLOCKED* in California

[Mikkeles]

According to this article at Verified Voting this only applies to Diebold's TSx paperless electronic voting system. Apparently:

"The Voting Systems Panel did not recommend against continued use of the Diebold TS electronic voting machines or use of optical-scan voting machines. The GEMs software is also not affected by this decision."

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.

there is no ONE TRUE METHOD

[feepcreature]

But which concordet method is the right one? Concordet is sometimes ambiguous, which is not ideal in an election.

All electoral methods (indeed most forms of government) represent a tradeoff between different considerations.

For voting methods, criteria of "goodness" might include this list [wot I mostly nicked]:

1. The voting system should always give a result
2. If a voter improves the ranking of a particular option, that option should not be disadvantaged (monotonicity criterion)
3. Removing a candidate should not change the winner of an election unless that candidate is the winner (independence of irrelevant alternatives)
4. Every possible outcome should be achievable
5. Non-dictatorship (i.e. more than one person's vote matters)
6. The number of seats won by any party should be in proportion to votes cast (Proportionality)
7. Simplicity of process, and accessibility to largest range of voters
8. Speed of election and count
9. Reduction of potential for dispute after the fact

In fact, Arrow's impossibility theorem has shown that the first 5 of these cannot be simultaneously met if there are more than . So pick one with disadvantages you can live with.

See this interesting Wikipedia article for further discussion of these ideas...

Re:Condorcet's Method

[Frigid Monkey]

Here is some information on

Condorcets Method

But I am not sure you are sufficiently informed on the nature of our great United States we are a Democratic REPUBLIC and the although I agree Condorcet's method may remove some of the "You wouldn't want to throw your vote away" syndrom that influences so many voters there is still the electoral college and even worse, as the last election demonstrated, it just doesn't matter.

I agree with you on one point: How we count our votes does not equate to voteing reform. We must fundamentally change the methods used in the election of our leaders if this is to be a truely free Democratic Republic. Now try and explain that to the people who have the power to change it. They are the ones the current system put into power eg: the least likely to suppot change.

-FM

Re:Paper trail (Maine Legislature ROCKS)

[Mark_in_Brazil]

I'm not sure if it's because of me (a /.er who's been doing a "Chicken Little" impression about electronic voting for a couple of years now) or because of Ben Cohen's (as in "Ben," the founder of Ben & Jerry's, tho' he's no longer working there) organization, True Majority, which has been sending her e-mails about e-voting, among other issues, but my mom has gotten into this issue. She lives in Maine and sent me an e-mail about an act recently passed in the Maine Legislature entitled An Act to Ensure the Accurate Counting of Votes. Note: navigation is a bit weird on the linked site-- if you go to the text of the Act, the whole text of the bill will not appear on a single page. You will have to use the arrows at the top and bottom of the pages to navigate around through the Act. You can also download a copy in M$ Word format.
Oh yeh-- there's an amendment. To see it, click on the "Amendments" link on the "Bill Text and Other Docs" page, or click here.
This is a sweet little piece of legislation. My favorite parts: it prohibits networking the voting machines, requires the voting machine software to be open source, and requires the voting machines to print paper ballots that are inspected by the voter and then placed into a ballot box. I am deeply impressed with this, and with the sponsor, Maine State Representative Hannah Pingree.
Here's a question: does anybody other than the OVC have a product that meets the criteria specified in the Act?
Responding to the parent post, I'll say that Maine can be considered a "backwater state," and its legislature has produced what appears to me to be a kick-ass piece of legislation on e-voting that explicitly requires open source software. Do big, rich, important states like California have such good legislation? I think not. Score one for the backwater states!

--Mark

PS: if you're near a Ben & Jerry's scoop shop, go there next Tuesday, April 27, and take advantage of Free Cone Day!!!

Oh, yes, I will repeat myself over and over...

[hummassa]

Till these topics die.

I live in Brasil. We have had voting machines in the last 12-14 years (yes, twelve to fourteen -- it depends the size of the city you are in). Brazilians here: the first election here in Belo Horizonte to use the machines were the mayoral (and city council, state representation, governor, house and senate) before FHC was elected (as I count it, 2 years + 8 years + 1 1/2 = 11,5 years). I know it, because I was "mesário" (election "table" official? election "clerk"? what is a good English translation?) in the previous election, and in the two subsequent elections). IIRC, there were electronic ballot boxes in Rio and Sao Paulo in the election before that (the only two cities larger than Belo Horizonte).

Our voting machines are mainly of three different (internally) models: (a) the old ones, that use VirtuOS (*) as the OS, (b) the new ones, that use WinCE as the OS, and (c) the newest and deprecated ones that have the second printer to print your vote, show it to you inside a clear acrilic case, and mix it with others inside the machine.

Externally, all of them look roughly the same: a box similar to the old "portable computers" of the eighties, with a 5-6" diagonal LCD and a big numerical keypad in the right side of the screen, that has, besides the 0-9 keys, "confirma" (ok), "erro" (cancel), and "branco" (white).

The electoral process (from the point of view of the voter) begins ... when you get your first job. If you are a mandatory voter (literate person from 18 to 65) you have to go to Electoral Court and register to vote. In the process of registering, you receive the "Título de Eleitor" (voter id), in which you have the number of you voting section. To change jobs, and specially to get a government job, you have to prove you are a registered and regularized voter (you voted in the last election, or regularized your voting situation after it).

In the election day, you scan the newspapers (or the Superior Electoral Court website), search for the address of your section, and go there. No, there is no transit vote, you can only vote at that address. If you can't get there, you'll have to "justify" your absence.

At the section, you will present your voter id to one the "mesários", and if you don't have it on you, you can still vote (you can show other valid id), but will be delayed. The mesário will search for your name in the vote-ticket sheet, and annex it to your id while you vote. You will sign a receipt in a sheet, and proceed to the voting "booth". Another "mesário" will type your voter id # in a remotely connected keypad, setting the machine in the "ready to vote" mode.

The voting "booth" is really only a desk with the voting machine over it, facing nobody else in the room, and sometimes with a cardboard "cover" around it. You will "dial" the numbers of the candidates, in order. when you dial all the digits of one candidate, a star-trek-like chime rings, his/her face will show up in the screen, and if you digited it right, you hit "ok". otherwise, you hit "cancel" and start over. After typing all the candidates, you hit "ok" one last time, the machine chimes again, and goes to "stand by" mode. You have voted. If you don't want to vote for nobody, you can hit "white" instead of the candidate ## (accounted as a "white vote", or "none of the above" -- this is the equivalent of putting your paper ballot in the box without marking anything), or if you really want to protest you can type 9999 or other non-existent-candidate-#, and your vote will be accounted as a "null vote", or "I'm really pissed of" (the equivalent of drawing pictures or writing "improper expletives" in a paper ballot)

Then, you get your id back, your ticket (keep it together with your voter id!!), and you go home. Ah, bars do not open (theoretically) in the election day, so hope you have bought your beer in the day before).

From the point of view of election officials, things are more complicated. The machines

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.

Re:Um.....

[steveha]

0) It's way easier to use than punch cards. Press buttons on a touch screen.

1) It's way easier to verify than punch cards. Just read it.

2) No hanging chads. Inspecting a ballot does not alter the ballot.

3) No hanging chad jokes.

steveha