VoteTrustUSA: Humboldt County Announces Transparency Project

[GuvWurld]

I have had my differences with my registrar but she's being bold and doing right by this step. Kudos to Warren for getting the story out. I attend the citizen Advisory meetings mentioned below, and at first we were asked not to discuss this. Earlier this week Carolyn sent out a mass e-mail that I suppose was my cue to blog this story but I just didn't get to it. Anyway, more good news for our movement this wondrous Friday evening...


http://www.votetrustusa.org/index.php?option=com_content&task=view&id=1331&Itemid=113

California: Humboldt County Announces Transparency Project

By Warren Stewart, VoteTrustUSA
June 01, 2006

Humboldt County, California registrar of voters Carolyn Crnich has approached TrueBallot, Inc. with an unprecedented request: to make publicly available scanned images of all ballots cast in the upcoming June 6, 2006 primary election. The registrar’s interest in the project resulted from a recommendation by her Citizens Advisory Committee on Elections, which works with the county clerk to promote accurate, security and public confidence in elections. The goal of making ballot images publicly available is to allow the public to independently verify official vote totals and to detect any discrepancies that results from machine errors or fraud.

Rob Richie, Executive Director of Fair Vote, The Center for Voting and Democracy commented “…we see this proposal as extremely important, and, indeed, potentially historic”.

The project promises several significant short- and long-term benefits. In the short-term, the project would establish a new level of transparency for public elections that would allow any member of the public to independently audit the election. It would enable the public to detect any discrepancies between the official totals and the scanned images if errors or fraud occurred. This would boost public confidence in or justify suspicion about the election process. The project would also create a set of data vastly richer than typical precinct-level results. This data would allow researchers and campaign strategists to study voter-level relationships between voting behavior in different races, and it would allow detailed study of voter errors, the usability of the voting system, and the accuracy of the read-heads on the county’s Diebold optical scan voting equipment.

In the long-term, the project could lead to ballot scanning becoming an ongoing part of election administration in Humboldt County, the spread of ballot scanning to other counties and states, and the use of ballot scanning in the original counting of ballots. Crnich hopes the project will have two additional benefits: encouraging teams of programmers to develop open-source software for interpreting scanned ballot images, and boosting the use of open-source software in public election administration.

For more information visit http://www.TrueBallot.com. Sample scanned images have been posted at http://electionsolutions.com/humboldt.html.

[SnoopDog]

1. This is what we need...

All we have to say is 'prove the vote tally". Since 2004 I have advocated some sort of receipt and published voting record where I can look up my 'vote reference number' and see it in the 'official tally of votes'.

I thank you very much GuvWurld for you patriotism and persistence to voting integrity.

[WillYourVoteBCounted]

2. how will they get the "scanned images"?

We know that the hart intercivic optical scanner scans the
optical scan ballot into it's memory and then counts the scanned
image. Specks of dust or creases in the ballot have
cause huge numbers of undervotes on the hart machines.

How will Humboldt get the scanned image -
with a VoteHere like device?

I think this is great, unless it brings in the "alternative verification systems"
that are intended to replace hand to eye audits of the actual ballots.

The VoteHERE is supposed to be able to work with optical scanners and
DREs.

Just please provide more info on how this is done, so we know it isn't a trojan
horse, (knowing how vendors and others think).

[GuvWurld]

3. I am cautiously optimistic

I don't know exactly what device will capture the images but the idea is to make .tif files available on a CD for anyone who wants to count, by whatever means they want to do the counting.

This is a solid idea that I want to support, and credit Humboldt resident Kevin Collins for the original concept. I can't fathom being seen as if I opposed this. Yet I do wonder what happens if different people or groups report different totals. I don't need to speculate as to how that could happen. It wouldn't matter. We have to have a method in place to handle it.

And it should be no surprise that this scenario is my greatest concern given how much I've written about inherent uncertainty, which is necessarily what we get with unverifiable elections that ensure inconclusive outcomes, results that will not be met with unanimous acceptance. Could this project feed into and reinforce this or will it eliminate it?

[WillYourVoteBCounted]

4. I re-read my post

I didn't mean to say that I am ok with the VoteHERE devices
or any devices that may be used ultimately to replace hand to eye audits of
the paper.

That is the risk of equipment that scans the ballots separately from
the DRE or the optical scanner.

It is a black box on top of a black box.

If the vendors can't stop us from getting the paper, they can
come up with ways to end the hand to eye recounting or hand
to eye auditing - by providing these other systems.

They would like to do away with hand to eye of the optical
scan ballots too...

Always get all of the details...

[GuvWurld]

5. I don't think this will be a second black box

I understand your point and valid concern. You're right, of course, about getting the details, but here we are. If this was being implemented by Ken Blackwell I'd say of course it looks like a fancy alibi, more smoke and mirrors. I won't vouch completely for our registrar, and nobody here will accuse me of being too trusting, but so far I want to support this happening so we can at least evaluate how it works or could work better.

[WillYourVoteBCounted]

12. make sure, please

please make sure.

We took a look at the VoteHERE type schemes, etc where
the device more or less serves like a DVR and video's
the ballots as they go through the scanner.

If everything isn't set up properly, you don't get an accurate
record of the vote.

Additionally, there is no way to assure that what the machine
shows you is what the ballots had.

Computer scientists are advising caution in relying upon anything
less than hand to eye audits of the actual ballots.

I feel very uncomfortable.

it would help if the folks in Humbolt asked exactly how these
"images" would be obtained, and whether any independent computer
scientists (not salesmen, not Ted Selker types) had reviewed the process.

These "images" are yet another digital record, and we need to ask lots
of questions.

I would hate to see these take the place of real hand to eye audits,
which they easily could.

[GuvWurld]

14. Let's collaborate

My group, the Voter Confidence Committee, has been planning an action to coincide with the Primary. We've kept it largely under wraps because at least one main component will be best served to unsuspecting media. That said, it has also allowed us to be flexible and adapt our planning according to developments such as this. So now I actually intend, fully, to address this on Tuesday night in a couple of public ways. Let's just leave it at that for now though obviously this forum will be made aware at the appropriate time.

With this in mind Will, you and everyone else can help me be prepared. Be concise and explicit and tell me what questions I should lay out. Please try to keep in mind that I want to be coming from a place that shows appreciation and supportiveness for our registrar's willingness to work with us on our concerns about transparency. This shouldn't be too limiting, as I am also prepared to ask hard questions as a spokesperson for a watchdog group.

So far I can think of:

How will the ballot images be captured and recorded to CD?

Will proprietary technology be involved?

Are there any mechanisms involved that have even the capability, as with interpreter code, to change the information from the ballot at any stage in the process?

How will the elections department and the county government at large resolve any potential dispute that may arise should different citizens or groups find different totals on their CDs?


Rather than trying to poke holes in this I think we are better off trying to define what success would look like. If we control that frame from the beginning we can set the tone for the expectations developed by the public. We can, in this way, educate people as to what would be a genuine development for election reform, and what would be a mere false alternative.

This is a really important dialog I'm very glad to be having.

[Wilms]

6. Carolyn Crnich is quite the character.

Following your efforts, I've noticed that--while she seems status quo--she kind-of gets that there's a problem.

I'll bet that's your doin', Guv. ;)

[GuvWurld]

8. It takes a village

Thanks for the, um, "vote of confidence," Wilms. My voice is one of many that it would seem Carolyn has at least been unable to ignore. I claim no special credit, however, in influencing her (at least on the basis of the relationship I've developed with her, if you can call it that). Seriously, if this turns into a significant advancement in election auditing, Carolyn will be lauded as a visionary leader. I would reserve that judgment until we see how things will be handled in the event of a discrepancy. I also want to see that history never loses the fact that this idea came from Kevin Collins who is a righteous old hippie with a heart of gold.

[Kip Humphrey]

7. While ballot images could be stuffed in the sense of being switched

on the CD, the coordination with vote tabulator flipping would be high, and, perhaps, too risky to attempt. This project is similar in some ways to mine (Voter's Choice - search DU ERD) in that it places citizens' eyes on the raw input, making vote counting fraud more difficult and risky. I foresee some late-night reprogramming of central tabulator programs forthcoming to isolate Humboldt County from vote flipping algorithms!

(oops: edited for spelling)

[GuvWurld]

9. The many watchful eyes of the public

The many watchful eyes of the public make the strength of this project. I won't speculate as to how the system could still be subverted, but surely that possibility exists. One factor that will be critical to allowing this audit method to succeed will be the timing of the release of the ballot image CDs. For this first run, really just a pilot project (I'm not sure, but it may not be all ballots this time), the CDs will be released a few weeks after the Primary, some amount of time after the official canvass is completed, but before certification of the results. I'm not saying this will happen or how it could happen, but if there were tampering anywhere in the counting/auditing process, the timing of the CD release could make it a very powerful placebo for pacifying public concern. Again, I wouldn't want this to be the case, but in this scenario, as far as election reforms go, this would become another false alternative.

[GuvWurld]

10. This story needs exposure - can we get a fifth R please?

Kick

[Gregorian]

11. Very interesting. I am guessing these are actual images of the ballots.

This is similar to what I did with my last Humboldt county ballot. I simply took a photograph of it, which I still have. With Kerry's name marked.

It would be a lot more difficult to photoshop each image, versus Rove at his master consol electronically flipping thousands at a time.

And it's not that time consuming. If a mechanism were set up so that each ballot took several seconds, this process would be a 40 hour week for one employee, if 50,000 ballots were imaged. Something like that.

And once again, thank you Mr. Gov! I wish I were living back in Humboldt. I'd help find a way to do the imagining.

[WillYourVoteBCounted]

13. see how easy to edit scanned images

"It would enable the public to detect any discrepancies between
the official totals and the scanned images if errors or fraud occurred."

The key word being "scanned images".

I went to the True ballot website, still not sure what
they are doing/selling.

However, give try this below -

Want to see how easy it is to create and edit electronic ballot images without knowing how the computer creates the original image?

http://cgi.cs.duke.edu/~justin/balloteditor/