Two Good (but not Great) Sections of the DNC OHIO Report

[Bill Bored]

I've finally started reading the thing.
<http://www.democrats.org/a/2005/06/democracy_at_ri.php>

I skipped to Section VII: Electronic Voting: Accuracy, Accessibility and Fraud and Section VIII: Transparent Aggregation of Voting Results Using the Internet and I have to say they were pretty good!

Dan S. Wallach, Associate Professor, Department of Computer Science, Rice University and Juan M. Jover, Ph.D. should be given some credit.

Does anyone have a problem with either of these two authors or the above sections of the report which they've written?

The one thing I found missing was that elusive random auditing protocol that could detect error or fraud distributed only within a subset of machines or op scanners in a jurisdiction. Perhaps this is because there isn't one! The solution may have to be more of a dynamic one, depending on how many machines there are, the number votes cast on each and the size of the error necessary to reverse the outcome of the election as initially reported. Regrettably, this report does not seem to address this problem and I'm beginning to believe that the only way to solve it (other than 100% hand counted paper ballots of course) is to allow manual recounts any time a discrepancy could change the outcome of an election. I think the new NY law allows this, but I'm seeking legal opinions on this at the moment as laws can be interpreted differently in the courts.

I looked at Section XII, Donna Brazile's Recommendations for Future Action, and this issue wasn't mentioned there either.

Anyone have any other comments on this? I haven't read the so called "whitewash" part about Ohio yet.

[Kip Humphrey]

1. A good start but does nothing about vote switching on vote input devices

Public records of raw vote totals is fine and needed to resolve secret counting but what is to stop voting machine level vote switching as reported during the 2004 election (observed in Ohio, North Carolina, Florida, Minnesota, New Mexico, Arizona, Texas, etc.)?

The solution is Voter's Choice: http://www.democraticunderground.com/discuss/duboard.php?az=show_topic&forum=203&topic_id=369418

[Bill Bored]

3. The report mentioned something like Voter's Choice, but see this:

Kip, I'm not sure how Voter's Choice (e-voting option) prevents the selling of votes but we can discuss this.

I think the PIN numbers should be supplied by the BoEs to ensure that there will be no dupes. I'm sorry this was never fully discussed on DU, but perhaps it's too complicated. That said, see what you think of these ideas:

I can envision a scheme using something akin to public/private key encryption to allow voters to verify their own individual votes without revealing them to anyone else. The latter would encourage vote selling.

Picture this:

Two databases, one public and one private.

The public one would be available to everyone via the Internet, kiosk, printout somewhere such as a newspaper, etc. as you say. It would contain all the votes with no voter IDs except for the PIN number for each ballot which is issued to the voter and printed on the ballot at the time the ballot is cast.

Each voter could then verify her vote in the public database, but there would be no way for her to prove that it's hers since she can't prove that the PIN number actually belongs to her. So, no vote selling is possible.

The private database contains the same information as the public database PLUS the voter's actual name, address, or other form of voter ID. This private database is ONLY accessible via a BoE intranet and only if a voter challenges the recording of her vote as shown in the public database. In that case, the ballot can no longer be secret, but otherwise it shall be secret.

If the voter believes that her vote was recorded incorrectly in the public database, she has the option of going to the BoE, presenting her voter ID, and looking up her ballot in the private database, which would contain her name, her PIN number and her ballot as cast. I like your idea of the checksum on the jpg and all that.

If any discrepancy is found by the voter, the voter can prove to the BoE that the vote was recorded incorrectly in the public database by showing them the record in the private database with her name and the same PIN number associated with it, and of course, her actual voter ID to prove who she is. But again, she can only show ALL this information to the BoE since no one else has access to the private database. Therefore, no one else is able to know for whom the voter voted (to prevent vote selling).

Now let's say the voter says her vote was different than that recorded, even though both databases match. It's her word against the databases. But there is one more way the vote can be verified that every one should be able to agree upon and that's the voter-verified paper ballot, or audit trail. As long as the same PIN number is printed on that piece of paper, and the BoE retains that record permanently, the voter can ask to see her individual VVPAT or VVPB which should put the matter to rest as it will be considered the official vote. I think you also mention this in Voter's Choice.

So what differences are there between this and the e-voting part of your system?

This one has the advantage of not being able to sell votes (except possibly in collusion with BoE insiders who are the only ones with access to the private database).

It has a possible disadvantage of the government (BoE) being able to tell who the citizens have voted for. I would think a law that prohibits access to the private database for purposes other than challenges by voters might solve this one, and there may be other ways to keep the BoEs' hands off of the data such as some form of voter-authenticated access.

So what do you think?

Personally, I think these sort of schemes will be very hard to implement, may still not be trusted by the voters, and do not ensure that enough voters will even take advantage of them. But I do like the idea of this form of random auditing where thousands or millions of voters have the ability to see that their votes were actually counted as cast. It would certainly deter fraud and that's what it's all about.

Peace.

[emcguffie]

5. This sounds like Athan Gibbs' voting machine.

I can't remember the name of it. But he's the guy who died in an accident a few weeks after demonstrating it, I think. Right?

[Time for change]

2. I agree that section VII is fairly good

It does discuss the potential problems with DRE machines and recommend against their use.

But of course getting rid of DRE machines alone will not prevent fraud at the central tabulator level. So I certainly agree with you that manual recounts should be done any time that they could change the results of an election.

Of course the main problem with the DNC report as a whole is what it leaves out and the statement that their analysis provides "strong evidence against widespread fraud". That comes from Section VI and is totally unwarranted in my view.

I believe that these are major egregious problems with the report because I don't believe that recommendations to not use DREs are going to have much effect on the US public unless they can be convinced at least that there was a strong likelihood that the 2004 election was stolen. If the DNC itself is going to come out and say that it wasn't, then I don't believe that there will be much sentiment in this country for election reform.

[Bill Bored]

4. Unfortunately, it's not the public.

Dale, here in NY, we had a fairly massive campaign against DREs.

I say this because I know the capital received hundreds or thousands of calls and letters asking for DREs to be banned in favor of paper ballots/optical scan (PB/OS). There was finally even a bill in BOTH houses of the legislature that would have mandated PB/OS statewide. It took years to get that point and there were about 2 weeks left in the last legislative session. But those bills were there. As with Boxer, it only took ONE SENATOR to introduce the companion bill in the Senate, and someone finally did it! There were 2 weeks left in the legislative session.

A 10-person committee which was 50/50 Dem/Repub had the opportunity to ban DREs and there was at least one Repub who was in favor of statewide PB/OS. To my mind, this means that there had to be at least one Democrat who didn't want to ban DREs, because in the end they were allowed along with PB/OS. Meanwhile, the Republican head of this committee had his e-mail box full to the point where you couldn't even send him e-mail anymore, the phones were ringing off the hooks, and the staffers knew what people were calling about even before they said so -- they were calling to ban DREs! They knew what the public wanted and THEY DIDN'T GIVE A SHIT!!!

And so, like Florida, we will have no statewide system and the counties are free to decide for themselves whether to have DREs with VVPATs or PB/OS. At least we won't be paperless, and that hasn't been on the table for years, if indeed it ever was.

The point is that those in office don't really care what the public thinks and until some of them lose their jobs over this issue, nothing serious will be done.

This is why I think the best that can be hoped for at the moment are recount laws that, unlike Florida's, will allow hand counting of VVPATs or VVPBs whenever there is the possibility of a change in the outcome of an election.

This is a far cry from "counting every vote", but it might be enough to keep the wrong people from getting elected, as long as the right ones, who may happen to be the LOSERS in the initial counts, are willing to contest results based on the lack of confidence in the machines and the system. As with Kerry in Ohio, it's up to the candidates now and this is where we have to direct OUR attention!

This may sound cold or conservative, but I think it's more important (and definitely more realistic) to guarantee a correct outcome of every election than to count every vote, and unfortunately, I think the courts would agree. It would seem that Kerry and the DNC may have failed us on both counts, but this doesn't mean it ever has to happen again.

I'll read the rest of the Ohio report as soon as I can.

[Time for change]

6. Sounds reasonable to me

I don't care that much about counting votes that have no affect on the election outcome.

[Carolab]

7. Bill: Do you know anything about Michigan's legislation?

I'm trying to find out what was meant at the DNC's executive committee meeting by Dean's co-chair (can't remember his name) when he said that some of the National Lawyers Council members had been helping Michigan with some legislation.

Does Michigan's legislation talk about the random auditing?

I need this information for discussions with folks working on legislation here!

[Bill Bored]

8. I no longer believe in random auditing.

It's good for a start, but it's what comes afterward that really matters. Random auditing is only likely to spot errors present on every machine or scanner, but not on a subset of machines or scanners.

Having said that, I know nothing about the great state of MI.
But you can search their bills here:
http://www.legislature.mi.gov/

I think they're mostly Op Scan which is good because they'd have plenty of paper ballots to audit!

[Carolab]

9. Couldn't find it but look at this voting system from OVC

The Open Voting Consortium is addressing security in a new system. They have demo'd it in California (see below). It has elements of Athan Gibbs' TruVote but uses a bar code on site to verify right after casting rather than verifying them later.

http://www.scoop.co.nz/stories/SC0411/S00068.htm

GRANITE BAY, CALIFORNIA - The Open Voting Consortium (OVC) will introduce legislation state-by-state to ban paperless voting and require that computer source code (the instructions given to computers) used in elections be made public. OVC will also begin to make their low-cost secure public software available for use in public elections.

{snip}

The Open Voting Consortium is a Nonprofit California Corporation dedicated to the development, maintenance, and delivery of open voting systems for use in public elections. OVC architecture will enable,

- An open and secure replacement of faulty election management software

- Electronic voting machine to generate a humanly readable optically scanable ballot

- Quick initial tabulation from precincts integrated with countywide and statewide systems

- A voter verified paper ballot that can be checked even by reading impaired voters

- Paper ballots checked against the electronic record

- The programs, hardware and all aspects of the OVC system are publicly available so that any system integrator can build voting systems, and provide maintenance or training.

- Vendors to compete based on services provided rather than proprietary hardware and software thus providing savings to taxpayers

Contact: Alan Dechert, [e-mail] 916-791-0456

**************
Check them out at:

www.openvotingconsortium.org

LONG BEACH, On June 23th Michelle Smith (pictured above) and other attendees of the Long Beach Open Voting Event tried out the open voting system demo. Michelle and other people at the event used a laptop to select the candidates they wanted to vote for. Attendees of the event also got to see how their printed paper ballot could be verified by scanning the bar code printed on the paper ballot and listening to their choices read back to them on the computer. They also got to try the Reading Impaired Interface, which allows people with reading impairments to vote without assistance. “What I saw was very user friendly and simple,” remarked Michelle Smith regarding the open voting demo. “It was motivating to know that people are working on this issue," Smith added, "After I saw the demo, I called the Secretary of State’s office and asked them to put in an open voting system for California.” The Long Beach demo was one of several live demos given by the Open Voting Consortium during the Southern California Open Voting Tour.

You can try a web demo version of the OVC software HERE.

http://www.openvotingconsortium.org/modules.php?name=Content&pa=showpage&pid=5