There's a trade-off. As I wrote earlier, an ideal system might include the benefits of electronic to guide the voter but produce a ballot that is human- and machine-readable so everyone can be confident of what was recorded. that's what the CalTech-MIT Voting Technology Project recommended. Those are some smart folks.
| || |
___Tech Policy/Security E-letter___ Written by washingtonpost.com's tech policy team, the e-mail version of this weekly feature includes an original news article and links to policy and cyber-security stories from the previous week.
Click Here for Free Sign-up
Read E-letter Archive
What about reports in technical publications that state that: (1) the (Access) data base holding the totals can be accessed without a password; (2) There are no plans to cross-check the cumulative totals with each precinct totals; (3) In the California elections, at least, the manufacturer's (Diebold) representatives had free access to the systems both locally and remotely; and (4) The president of Diebold has stated that he "will do anything in my power to deliver the state of Ohio to the President"?
Dan - Many readers continue to mention Walden O'Dell's now in/famous dual role as a Republican fundraiser and head of a voting machine company. O'Dell and Diebold Election Systems have passed this off as poor wording, unfortunate coincidence, etc. etc., but is there some reason to suspect that there is more to it? What should Diebold do about this? What should local elections officials do? No longer trust Diebold?
Dan Keating: If you read the security studies that have been done, you see that the systems were not secure. As you mention, they allowed direct editing of the tabulation file. They allowed dial-in access to the machine (helpful for technical help, but not secure). The vendors have done a lot to undercut faith in them.
I think everyone agrees that local officials have to take responsibility for the machine security. I think people will pay more attention to testing the machine totals against the tabulations.
Elections historically have been extremely sloppy. In florida in 2000, Volusia double-counted votes. St Petersburg double-counted votes. Palm Beach deleted results from one precinct. Nassau still can't say how many ballots there were. Polk lost several hundred ballots and then found them the day after the election. Those are all hopelessly sloppy procedures. If elections are run that badly, there's no reason to have confidence in them
The Caltech-MIT voting project is located here. I also quoted one of its members at the end of a long article about e-voting in the D.C. area that you can find here.
You reported this week that electronic voting machines are "paperless." Yet I heard an NPR report about election officials in Florida doing a paper recount from the electronic machines they used in the primary. Do electronic machines have paper that CAN be used for a recount? Or is there NO paper in them?
Dan Keating: As i said, the electronic machines record each ballot. That ballot-by-ballot record can be printed out to compare with the running totals the machine produced. The only error it would catch is if the software recorded ballots correctly but counted them wrong. That provides no comfort to people who fear hacking since obviously a hacker could corrupt both parts of the process.
Santa Clara, Calif.:
Would it be possible for an electronic voting machine to provide a PIN to a voter so that after the election a voter could go online to verify that their vote was actually counted in the final tally?
If this was possible, I would be in favor of electronic voting. Without a process to verify my vote I would be opposed. Currently when I vote, I retain a stub with a ballot number when I leave.
Dan Keating: can't happen because then people could be paid for their vote. the payor would make them log in and prove how they voted. that's why receipts can't be given to voters. there's a version that used encrypted codes, but it's pretty complex.
For the purposes of disclosure, I was a lobbyist for HAVA, particularly for disability polling place access and touch screen voting systems (DREs).
The DREs were, by and large, a Democratic sop to certain voting blocks: Voters with disabilities, and voters who don't read English. Can you please talk about the positive benefits of these machines?
Thank you for your time
Dan Keating: I don't think it's true at all that DREs were a sop to certain groups. They have advantages for lots of people if they work correctly. They certainly are helpful for the disabled. And they offer an easier way to do multiple languages. They can prevent lots of kinds of voter error.
A system that captures those advantages and still yields a backup for error or fraud would satisfy most folks.
It seems mindboggling that the idea of internet voting is written off so easily, as if our current system is the hallmark of integrity. I think that internet voting would change our entire system for the better, and greatly raise participation. When will this finally happen?
Dan Keating: The fear is that a successful fraud could work on such a massive scale. One element of our current jurisdiction-by-jurisdiction system is that it's impossible to craft a fraud that would work everywhere. But widespread Internet voting would present closer to a single target for an attack.
Falling Waters W.Va.:
How would the paper trail work? if it has the name and address of the voter it is going to destroy secret voting, If it does not link to a voter it seems to me that it would be worthless as an audit or recount, all the programmer would have to do is put in instructions that say if the vote is for my side print 2, if for the other side print one. If the paper trail is not held in the computer and is given to the voter then during a recount you would have to trust that the Paper is not one of a hundred copies.of that transaction.if it is held in the computer you have just given the winners side a complete record of who voted against him.
Dan Keating: The paper is just a written record -- like a cash register receipt -- showing what choices the voter made. The voter sees what's written. That way, the machine's total at the end of the day can be compared with the paper receipts approved by the voters to show that the votes recorded matched what the voters wanted. There are no names or addresses. In fact, there are some theories that it has to write out-of-order so you can't reconstruct how people voted by knowing what order they went to the machine.
Most rational people seem to agree that without a paper audit trail, computerized voting is extremely questionable and unreliable. This has been known for some time. My question is why major political figures (Republican or Democratic) or the major news media have not made resolving this a national priority?
Dan Keating: there were some things in place, such as software testing and audit trails. but it was not until last year that the movement among computer scientists got momentum saying that the existing systems weren't good enough protection.
until then, the only question was whether manufacturers could build them fast enough to meet demand. so the new debate has changed the dynamics and caused some further thought that appears likely to yield a better result -- just not on the timetable that suits everyone.
It's after 2 p.m. so I'm done.
Thanks for great questions.
I hope everyone concerned has volunteered to work at the polls on election day and make sure everything comes out OK.
Thanks, everyone, for coming online today to check out the Live Online session with Post reporter Dan Keating. I'm sorry we couldn't get to everyone's questions, but please keep checking back with washingtonpost.com as we follow the key developments on electronic voting through Election Day.