Why Voting Technology Must be Open Source

By Timothy M. O'Brien
October 28, 2008 | Comments: 9

Vote_Button.jpg

The following is a video from Video the Vote. From the Video the Vote website.

"Video the Vote is a national initiative to protect voting rights by monitoring the electoral process. We organize citizen journalists--ordinary folks like you and me--to document election problems as they occur. And then we distribute their footage to the mainstream media and online to make sure the full story of Election Day gets told."

Here's a video they published documenting some of the "vote switching" problems which have been reported in West Virginia:

This is a video of Jeff Waybright, Jackson County Clerk in Jackson Country, West Virginia. A state that is very close this year and although it only has a tiny number of electoral votes it could be a deciding race in a close election. In this video, you'll see Jeff demonstrate a machine out of calibration switch a vote for a Democratic candidate to a vote for a write-in candidate. Jeff will then "calibrate" the machine, and continue to have issues with calibration.

The machine should stop working if it isn't calibrated. There should be no possibility that a person can press one candidate and have the vote switch to another. If the possibility exists that the machine could register a wrong vote, it shouldn't be registering a vote at all. It doesn't have to be this way. Voting technology should be open source, both the hardware and the software should be transparent and certified by public interest groups.

If Voting Machinery were Open Source

If voting machines were open source, we could validate that the source had not been patched. We could have a nonpartisan group of software developers certify a build of the voting machine software in a distributed and completely transparent way, and we could distribute certified, signed binaries to voting machines throughout the country.

If voting machines were open source, we could publish an anonymous, verifiable paper trail online which would allow voters to make sure that their single vote was recorded days and weeks after the fact. More importantly we could have a transparent, easily verifiable online audit of who touched a particular machine's software, when, and what version was installed.

If voting machines were open source, we could legally publish the results of a third-party software and hardware audit. You could walk up to the voting machine knowing what happens to your vote. Does it get stored in memory? Does it get stored in a database? If so what database? What version? What happens if there is a blackout?

If voting machines were open source, we could look forward to a series of technical articles by programmers like Brian Behlendorf, Adrian Holovaty, and Greg Stein all certifying various parts of the code. "This year Sunlight paid me to verify that the vote recording mechanism was sound, here is my analysis...." You wouldn't have to take them on faith alone, you could verify the code yourself.

If voting machines were open source, there would have been a long discussion on the developer list last year about making the touchscreen completely fail-safe with no ability to record false votes.

If voting machines were open source, you could watch the same video, find out what software version was installed in Jackson County, WV and run some unit tests against the same binary.

If voting machines were open source, each election cycle would yield improvements in the form of bug reports from poll workers.

Current Reality

Instead, we have series of proprietary solutions with almost no transparency being installed and maintained by a collection of consultants and vendors who answer to election officials who probably couldn't tell you the difference between a SIMM and a CPU.

I couldn't tell you if that Sequoia machine that I use in Illinois has been patched or modified by a third-party prior to use. I couldn't tell you if the machine was calibrated or not, and I'm also certain that the scrolling paper trail is difficult to read and likely useless. Anyone with even the slightest knowledge of software engineering will tell you that a checkbox on a screen and printout is still not a guarantee. I have no idea what happens to my vote data after I leave a polling place. How is it transmitted? Where is it stored?

Demand It for 2010

Maybe, just maybe, we'll have an administration that understands the power of open source. Technical voters need to band together to convince our representatives that open source voting technology is an imperative. Forget the "industry" that has developed around voting machines, this technology belongs to the people.


You might also be interested in:

9 Comments

Wow, that's scary. I bet Google would jump in on the software side with a little prodding, a la Android with less revenue potential. I don't know hardware companies as well, though. Who out there is competent and trustworthy on the hardware side? I'd vote for Bug Labs, but they're tiny. Heck, Apple? Something like this can't be some SourceForge project because most states wouldn't go for it. If it were Apple and Google, that's a different story. The PR would be a dream for them, and they'd make some cash too.

Strongly disagree. The open/closed source issue is a red herring. I've never heard a single convincing argument why technology of *any* kind has any value in the voting process - at least, why its meagre and superficial advantages outweigh the profound actual and potential dangers. I can't understand the way it's seen as axiomatic that technology helps the process, and that the only remaining questions relate to what sort of technology, and how it's controlled. Anyone considering the use of technology ought to ask what are the problems that the technology is intended to solve, because there are *none* which are remotely the same magnitude as the problems it introduces.

Forgive a link to myself: Ned Ludd goes to the polls.

Paul,

There are some strong arguments for the use of technology in voting. I ran over a few of them here.

I've looked over Ned Ludd goes to the polls, and I note that your description of the voting process begins when the vote is cast. That's incomplete, as the process begins when the ballot is generated. It's at just that point where many of the advantages to the use of voting machines appear. That said, I agree with you that the generation and marking of voter-reviewable, hand-countable paper ballots is essential.

John,

There are some strong arguments for the use of technology in voting. I ran over a few of them [here].

I see basically two arguments there: one is physical and linguistic accessibility; the other is prevention of vote fraud, by (principally) avoiding a situation where ballots can be limited. I don't see either as a nut which requires the sledgehammer of e-voting.

Both really collapse to ensuring the ability to generate (or have generated) in a timely fashion, enough ballot papers, in appropriate forms, to meet the needs of all of the voters. With available census data, I'm extremely sceptical that this isn't possible using traditional print media, in a way that's sufficiently more expensive than complex voting machines to make it prohibitive. Add to this a working assumption that ballot papers which are marked in a low-tech way, and counted by human eye, can be much simpler - and therefore less expensive to print - than machine-readable papers.

Reacting to concerns about fraud by poll organisers by adding significant layers of opaque (opaque to most voters and poll-workers, that is) complexity to the process seems to be missing the point. It results in a process which is much harder for the - typically older, typically non-technical - poll-workers to properly manage, especially in situations of failure. It also merely changes the methods of fraud; machines can be tampered with, and electrical supplies cut, just as easily as limiting physical ballot papers. But perhaps most importantly, the way to deal with potential fraud at the polling place is to remove control of the process from corrupt and partisan officials, not to engage in a technological arms race with them.

I've looked over Ned Ludd goes to the polls, and I note that your description of the voting process begins when the vote is cast. That's incomplete, as the process begins when the ballot is generated.

Not entirely. I explained that it seems perfectly acceptable to me (though far from necessary) to have technology serve the same function as the pencil. A machine which aids in marking a physical ballot paper, which is then physically transferred by the voter to a physical ballot box (to subsequently counted physically), doesn't present any obvious dangers that I can see, and overcomes the accessibility and vote-fraud issues you raise. Each voter gets to see their ballot, and to cast it by hand.

It's at just that point where many of the advantages to the use of voting machines appear.

And where many of the disadvantages appear. As I tried to describe, each voter's confidence that their ballot has reached the box in a state that's consistent with their voting preferences is crucial.

Me:

John:
It's at just that point where many of the advantages to the use of voting machines appear.

And where many of the disadvantages appear. As I tried to describe, each voter's confidence that their ballot has reached the box in a state that's consistent with their voting preferences is crucial.

John, apologies. That last paragraph of mine comes from a misreading. Your 'at just that point' means the generation of the ballot - and with caveats I agree with you. I'd read what you said as meaning from that point onwards.

But using a machine to help mark a physical ballot doesn't strike me as 'e-voting' in any meaningful way. It is really just using the machine as a glorified pencil, and doesn't create any distance between the voter and their vote.

If you guys have some time, please check out Scantegrity at http://scantegrity.org/

In addition to using open-source software, it has several unique features that ensure vote tally integrity, including the concept of a privacy-preserving receipt that a voter can use to verify their vote was counted and counted properly.

I vote for McCain and Palin. They are a great team. I am afraid of Obama.

Sheri, nobody asked you who you were voting for and that's not what this article is even about. But while you've got me on the topic, I think you're scary for thinking that McCain/Palin are a great team. Surely your reasoning ability is questionable at best.

hey i think that voting should be privet. but on the other hand people are gona know who voted for any ways. so why kept it a secret???

News Topics

Recommended for You

Got a Question?