Slashdot Mirror
California Sec. of State Wants Open Source E-Voting Systems
Lucas123 writes "California's Secretary of State, Debra Bowen, was among a group of e-voting experts at MIT yesterday who said the nation's electronic voting systems are still not secure and many run on faulty software. Among the suggestions offered to fix the problem: use open source software, stop delivering e-voting machines to polling places weeks in advance of an election, and keep a paper trail for auditing purposes. Bowen also believes that a ubiquitous Internet voting system could not work without the use of a national ID card system."
No need to open source anything or make any other changes... Just slap a sticker with one of those disclaimers on each of the current voting machines that reads "This is not a scientific poll and is completely inaccurate."
Problem solved.
I'm a big tall mofo.
Let us follow them http://news.bbc.co.uk/2/hi/europe/4343374.stm
KDE or Gnome? But since it's California, it'll probably be Enlightenment.
"The average reporter we talk to is 27 years old......They literally know nothing." - Ben Rhodes
Lots of the problems described occur because a voter must actually punch a bunch of buttons in just a few minutes -- matching a (hopefully predetermined) set of things they wanted to vote for. It seems like there's lots of room for error because of the time crunch that everyone feels in this situation.
What if you could actually do the ballot on your computer at home, carefully making sure that the buttons you push are what you intended, and then bring a printout with something like a barcode or other digital encoding of your selections? (This wouldn't have to be tied to your name -- that can still happen in the booth.) Then you bring that barcode to the booth, and it scans it after you walk in, and that "preloads" your selections. Then, you're just down to a verify step, under less pressure.
Seems to even open a new market for various parties to distribute the barcodes of their respective positions... :-/ don't want to make things *that* easy.
Just a thought...
--
Learn electronics! Microcontroller kits for the digital generation.
Next step would be firing the so-called "technology experts" in the popular media, who apparently lack the the tech saavy to google for what "open-source" means.
Coverage of the G1 launch was a beautiful example of their ignorance. Many times I heard the fakers pontificate about the "security concerns" in using open-source software, while not even knowing meaning of the term.
No, don't follow them. Electronic Voting is an inherently flawed idea, let's just stick to pen&paper voting.
Raise your hand: right hand, McCain; left hand, Obama
Google 's brilliant programmers have a flawless (albeit, beta) system that can correctly tally the votes.
Probably.
Schroedinger's Brexit: The UK is both in and out of the EU at the same time!
This was not mentioned in the article. Voters should be issued a voter number and should be able to go online and verify the accuracy of their vote. The election judges should be able to do a printout and be able to go online and verify the paper matches the online data, with the help of a ballot watcher maybe. In cases of more than a 100 percent voter turnout, elections must done again.
I regret that I only have one mod point to give per post.
At Pycon 2007 in Dallas I saw a lightning talk demonstrating how electronic voting could be secure in just 293 lines of code.
However the bottom line is that you shouldn't trust any voting system. What you should have is an auditing system where you can do recounts. The less moving parts or the fewer lines of code you have the easier it will be to audit a system.
"You'll get nothing, and you'll like it!"
Honestly, how hard is it to write voting software?
Comment removed based on user account deletion
Just stick to paper. It works.
Gone!
The main problem with e-voting is that the average citizen can not reasonably check the ballot counting. With paper, even a seven-year-old can check if it is counted correctly. No one can look into the computer, so even if you have paper ballots, they still need to be counted. Nothing won there.
Voting machines could print barcoded papers which can be counted electronically. This would allow fast vote counting without all the problems of the punch cards. Random samples of the paper could be counted manually as a security check.
Whatever happens there must be a paper trail. These are important decisions and any system without bits of paper should be a no-starter.
No sig today...
Every optical-scan voting system should use scantegrity.
This is not that hard, and it sure isn't rocket science.
Strip down a distro to the kernel then ad the following:
Please a driver for something I missed....
The device has only enough ROM to POST and is hard coded to boot from the thumb drive which contains the OS & drivers and voting software with a modified USB connector that is a different shape then standard. This is a mild security feature.
An additional thumb drive will hold the data, again with a different shape so that the two cannot be confused, and both are encrypted using a two key scheme of some sort, suggestions?
Insert the drive one, power up the machine, it will then POST itself and ask for the data key and will go no farther until it validates the Data Drive. Voting commences and when voting is complete, the machine is shut down, drives are pulled and returned to the registrar for counting.
Hey KID! Yeah you, get the fuck off my lawn!
Create an account for each item on the ballot.
Have voters register their bank accounts when registering to vote.
Only votes from registered bank accounts are accepted.
Only deposits of 1 cent are accepted.
People can vote at ATMs, online banking, or at a teller.
Check the balance at the end of the day.
Everyone has a paper trail.
*Just an example of using a solution for a solved problem for an unsolved problem.
**The system can be implemented without the banks cooperation, but why not have them cooperate - they're nationalized now anyway.
At least in my country, the traditional paper method works well, is easy to implement, and what is most important in a democracy: easy to understand and easy to check for everyone.
No matter if the electronic method is opensource or not, only a very small percentage of voters will actually understand how it works, how it is kept secure and safe from manipulation.
So what immense advantages would electronic voting have to make up for this fundamental problem, that will never change, no matter what the electronic solution will be?
1. Take vote electronically.
2. Assign a randomly generated UUID.
3. Print UUID+vote on internal paper tape for backup.
4. Print UUID+vote on paper receipt for voter to keep.
5. Post UUID+vote on a public web site anyone can view.
Now, anybody can see the tally, do the math themselves, etc. And everyone who cares can look at their own UUID and see if the public tally is accurate.
Our electronic voting machines work pretty good.
http://techaos.blogspot.com/2004/05/indian-evm-compared-with-diebold.html
http://www.eci.gov.in/faq/evm.asp
From TFA:
"Because we don't have any kind of national ID card, we have no method for doing that," Bowen said.
Why the fuck is it that every politician who might come up with one good idea cannot resist tying it to an evil add-on.
"Bowen also believes that a ubiquitous Internet voting system could not work without the use of a national ID card system."
For someone who seems to have a clue, she lost a lot of credibility with that statement. There is absolutely no need for a "national ID card system" to have secure and accurate voting. Voting is handled by the States, not the Federal Government.
Actually, Internet voting must not be allowed unless you can make a connection that can't be eavesdropped between your mind and the voting machine. If third parties can listen in (even if they need your consent to do so) - vote buying will again be possible.
Imagine you manage to get this part right. Now you would only have to trust the voting machine to accurately store your vote, without the benefit of a voter-verifiable paper trail or anything you can possibly show to third parties to demonstrate that you voted one way or another... can't be done.
The voting system shouldn't need to be open-source. In fact I'd go so far as to say that any electronic voting system that requires the voting machines to be open-source is fundamentally flawed and shouldn't be trusted. The system should be designed so that it doesn't matter whether the machines themselves are recording correctly or not, it's still possible to determine whether the counts are correct (ie. match the intent of the voter) or not and ideally to be able to reconstruct the correct count.
Yes, it can be done. Anyone who accounts for money has been doing it for centuries. Any lowly night auditor at a hotel can explain the principles complete with examples.
More info on the Indian EVMs here http://en.wikipedia.org/wiki/Indian_voting_machines
The solution should start by simplifying the elections itself.
Its always going to be confusing when you have to vote for dozens of things at the same time. I mean come one, what do I care who is elected to the school district board when we are talking about presidential elections.
We should have to vote for 3 things max on national elections:
The president.
The senate.
The house.
Thats it... Then we can go back to old school, simple paper ballots with 1 name on it, And one ballot box for each. Then they are all counted separately.
Too much democracy kills democracy.
Computers are NOT going to solve this problem.
Just stick to paper. It works.
Just curious, but why do you Americans vote for so many things all at once? It's like you have federal, state, county, and municipal elections all at the same time. Plus you vote on things like sheriffs, prosecutors, and a gazillion different propositions.
And what's with all of these "propositions"? In Canada, if we have a question put to the general population (we call it a "referendum") then it's pretty big change. They generally only happen once a decade or so.
No wonder things are so complicated: people have to figure out what a dozen different things mean.
Why is that?
Congress has no intention of approving electronic voting or putting any effort into finding reasonably secure options. Were they to do something like that their jobs suddenly become meaningless in the long run. We would have no need for elected population representatives if the population was able to vote on virtually every subject themselves. Of course then it would become a partisan issue because we would have to give a computer to every single low-income family (government funded through your tax dollars) or we would be discriminating. Republicans wouldn't want that because it would take away their advantage. How do you pass something that no elected official truly wants passed? There is alot more involved to e-voting than just the security concerns. I wonder what circumstances it would take to get it to truly happen.
Why is taxpayer money continuously wasted on voting machines??? Hello! We have something called The Internet! A few PC's at each polling place (linux of course) securely connected to a central server. Each voter gets a printed receipt with a password that they can use to log in if they want to verify their votes. An encrypted printout of each voting session also remains at the polling location as backup and for auditing purposes.
The URL is
http://blogs.computerworld.com/voting_should_be_a_hand_job
I wholeheartedly agree.
People that think an open-source solution will solve this problem are delusional. You can never have a 100% solution no matter how many keys or security mechanisms you dream up. In fact, the more you try to secure it, the more incentive there is to find a way to break it. Don't you people learn anything from all the stories that are posted about cracking DRM? The only thing that is reasonably secure is create a ballot the voter can verify, have a machine or person tabulate the ballots repeatedly and reliably, and create a place for observers to minimize tampering. That's it, that's all you can do.
It is supposed to be impossible for me to show someone how I voted. I can't be given a receipt or anything (it would be too easy to buy votes).
With that as a given, how does a paper trail help? If I as a voter can't be sure that my specific vote is the one on the paper, then it seems like there are still thousands of ways for someone to change it out.
The machine could print fake info first of all, so it would HAVE to be something I see and validate. Since it has to be a public record, it can't be tied to me by any kind of key though--so after it's printed it out for me to see, there is no saying that it doesn't print a second or third line for another candidate at some other time...
If the paper trail disagrees with the digital tally, do you just assume that the paper wasn't messed with or substituted?
I'm guessing people have thought about this more than I have, and I see the mention of "Paper Trail" a lot so I just thought I'd ask.
To provide a more accurate picture to the voting masses, just replace the voting machines with modified slot machines. You have to insert $1 coin and strike 3 Obamas in a row to actually vote for him. All other votes go to McCain. Top part of the machine could display laughing members of Congress, and what are they worth (only the millionaires).
Why not have NIST have another open contest, like for AES or the next hash?
The usual requirements apply, no patents, fully open, etc...
This seems so brain-dead simple to me I can't understand why I haven't seen everyone here saying it?
use open source software, stop delivering e-voting machines to polling places weeks in advance of an election
Voting machine hijackers won't need the machines weeks in advance when they have the source months in advance.
Paper trails are bad. What we need are PAPER BALLOTS. The machine can help by helping voters fill out the damn form correctly and printing the ballot in COMPLETELY human readable form. Then an OCR can read it AFTER the voter has had a chance to make sure they are casting the votes they intended. The ballot helper MUST be completely independant of the OCR and the voter is the only link between the two.
No receipt, no tracking, no paper trail. Just a paper ballot. And of course we keep the ballots just as we normally would.
I'm surprised I have to explain this.
If there needs to be a truly reliable paper trail to audit, why even bother with e-voting? At that point the only difference I can see is whether you push a pencil through paper or push a button, and one is far less trustworthy. Quicker tabulation can't be worth that much to the lazy voting public, can it?
In Punchscan the voter gets to take a receipt of a paper punched vote home that they can then use to verify their votes from a home computer. This receipt does not indicated how the voter voted and when the voter pulls up their vote online to confirm their results, it only reports their receipt. This means no one can tell how someone voted, but the voter can verify that their votes were counted as they believed they were. This makes the entire election able to be mass verified and is completely secure.
Scantegrity is a another system that borrows ideas from the Punchscan system and is for optical scanning voting systems.