Slashdot Mirror
Paper Trails Don't Ensure Accurate E-Voting Totals
An anonymous reader writes "In an new report from the Information Technology and Innovation Foundation they say that paper trails increase costs and can actually reduce the chances a voters' choices are accurately counted. Congress is considering a 'Voter Confidence and Increased Accountability Act of 2007,' which would mandate 'voter-verified' paper audit trails."
Just a quick browse of their "ITIF in the news" page and it looks like they are big fans of Real-ID and RFID tagging in general. On network neutrality they appear to be in favor of just leaving it up to the FCC to determine on a case-by-case basis what telecomm companies are abusive and which aren't - no legislation required, and their justification seems to be that some of the proposed legislation has been over-the-top (typical FUD about preventing telecomms from 'innovating').
Who funds these people?
When information is power, privacy is freedom.
Information Week has given itself a black eye by saying nothing at all of any interest or substance about this issue, while hyping a report that it can't even describe adequately. All this means is that a "Black is White, Up is Down" paper will be forthcoming soon from an industry shill. The only news here is that this a self-inflicted reminder not to read Information Week.
... the answer is very simple.
The voter marks the ballot paper with a pencil. The ballots are counted by hand by human beings.
Completely transparent, complete audit trail, safeguards against all the failure modes discovered over the decades, results within hours, recounts within hours if needed.
Oh, and I expect it's cheaper than all this inappropriate mucking around with computers too. Computers aren't the answer to everything. This is one application in which they have no place.
"ITIF wants to spark discussion of how new technology can solve the problems. The report outlines innovations in voting machines that offer "end-to-end verifiability." It explains the cryptography the systems use and says that Congress should pass legislation based on S. 730 and H.R. 2360, which require verifiable audit trails without specifying that paper be used."
1. Not end to end. I can't do cryptography decryption in my head, and the vote verifier at the other end, he can't also do decryptions in his head. So any solution that involved cryptography isn't end to end.
2. One doesn't preclude the other. You can encrypt the electronic vote AND STILL HAVE THE PAPER AUDIT TRAIL to check the machine's cryptographic vote matches the voters intentions.
3. Papertrails, or ballots as we use to call them, have a proven track record of uncovering fraud in voting. To date the fraud in electronic voting is suspect but unproven. It is unlikely that fraud is eliminated in electronic voting, because fraud is *easier* not *harder* to do when votes can be changed so easily and untraceably on mass in a computer. So the lack of uncovering fraud is likely to be a weakness in the auditability of these machines. i.e. we suspect voter fraud because of systematic irregularities in key districts, but nothing can be proved because the lack of paper trail to verify against.
Why does he want unauditable machines? I see from his history that he's a professional technology lobbyists, but I'm curious why the FUD to keep the voting machines unauditable?
"I would have much more confidence in a cryptographic scheme that makes it effectively impossible for a voting machine to cheat. This is not all that difficult to accomplish and the necessary design criteria are widely available in the literature. A paper trail doesn't really help."
There is just one simple, practical, logical rule for machine assisted voting that anyone need remeber:
A machine that prints your choice is at worst a waste of money, a machine that counts your choice is at best a waste of money.
And did you exchange a walk on part in the war for a lead role in a cage? - Pink Floyd.
My opinion is that there is no 'secure' e-voting system.
I also do not see any reason to abandon paper-based voting, which still is not 100% secure, but much more difficult to 'hack' due to transparency by distribution of control.
CC.
TaijiQuan (Huang, 5 loosenings)
I've been interested in the Caltech/MIT Voting Technology Project for awhile now, and they have quite a few papers on voting issues, including discrepancies, statistics, technologies, reliability. One paper in particular (PDF warning) speaks of a study done on different methods of verifying user voting. For the mock election, they randomly inserted incorrect vote records, and thus tested different methods of auditing, to see how often voters noticed the discrepancy. With the paper trail, only 8% acknowledged any problem (i.e., realized that the machine voted wrong). For an auditing system, that's not at all counterproductive as the topical article would have you believe, but it's still pathetically ineffective.
A different kind of auditing system is advocated in the paper: one using audio via headphones to play back the recorded votes to the voter. 77% of voters caught the errors. Of course with every added system, there is inherent risk -- listening devices, accessibility, etc. -- and, of course, audio auditing is relatively untested, but this seems promising. This, however, assumes that the problem is with voters or the machines making an honest mistake and not with the machines maliciously changing votes. Thus, the best course of action would be to have both paper and audio: one to help the voter, the other to verify recounts and prove unreliability.
Of course, no auditing system, no matter the sophistication or rate of helpfulness, will matter if the machines themselves are designed to be corrupted and the vote counts manipulated. Ultimately, it will be far more beneficial to the American people if, rather than trying to force accountability and regulations on corrupted producing companies bought and paid for by corrupt political crooks, the machines are written and produced, or at least heavily tested, by independent committees.... May I suggest academic committees, such as this Caltech/MIT VTP, or similar groups? Their ultimate goal is to certify reliability, and since academics is far less motivated by money, they're far less likely to be corrupted. Or so goes my theory, anyway.
Why the fuck do you Americans need to use goddamn voting machines?
Canada gets away just fine with using paper ballots. When you vote, you use a pencil to put a check in a circle next to the name of the candidate you're voting for. The circle is large and the text is large, to allow those with poor eyesight to get a better view of what's on the ballot, thus reducing mistakes.
What's more, the results for Canadian elections are near-instantaneous. They actually have legislation in place to prevent the media from reporting about the final results in the eastern and central provinces while polling stations are still open in the west! Why the fuck can't the US manage that?
Yeah, the American population is 10 times larger than the Canadian population. But that's irrelevant! Use 10 times as many ballot counters, and the system will scale just as well.
It's a mixed situation here in Europe. Some of our nations use the sensible Canadian method. Others are stupid, and follow the American scheme with doodad voting machines and all that jibberjabber. But really, we should all just use the Canadian method. It's the best, and safest, there is.