Wednesday, December 10, 2008
Humboldt At The Tipping Point: Who Dares Defend Diebold?
Here in Humboldt County, CA a local story of national interest broke last Thursday on the websites of the Eureka Times-Standard (archive) and North Coast Journal. The next morning I wrote a letter to the editor that appeared in today's T-S (archive). I'll let this serve as a summary then provide links to much of what's been published already and add some further reasons for optimism at the bottom.Any defenders?
So here's a summary of links from the past several days, then I've got a few more observations.
Letters to the editor
Posted: 12/10/2008 01:15:38 AM PST
First I'd like to congratulate Kevin Collins, Tom Pinto, Mitch Trachtenberg, Parke Bostrom and all the volunteers of the Election Transparency Project.
Their work revealed a discrepancy caused by Humboldt's electronic voting equipment last month.
Over the last few years I've made many different arguments for getting rid of the Diebold (now Premier) equipment used to count votes in Humboldt County. Somehow it wasn't enough that they “count” in secret, can be easily manipulated without detection, and report results impossible in a legitimate election.
Somehow local decision makers weren't deterred from doing business with a company that admitted to illegally installing uncertified software here and elsewhere; that was sued in class action suits filed by company shareholders; and whose then -- CEO said he was “committed to helping Ohio deliver its electoral votes” to Bush in 2004.
Now we learn that Humboldt has finally experienced what is euphemistically called a “glitch.” In reality it was a bug in Diebold's central tabulation program, GEMS. This caused the results of November's election, already certified as accurate by Registrar of Voters Carolyn Crnich, to be proven inaccurate.
Worse still, Diebold knew about the bug at least four years ago and never fixed it. Other counties were made aware of the problem and told how to work around it. Crnich says she never knew, and I believe her.
This raises many questions, most important among them: Who dares defend the continued use of these machines and the county's relationship with Diebold/Premier?
Dave Berman
Eureka
T-S, 12/5/08: Software glitch yields inaccurate election results (archive)
T-S, 12/7/08 Local elections office commended (archive)
T-S Editorial, 12/7/08 - A glitch that should never have been (archive)
Wired - two Kim Zetter articles from 12/8/08:
Serious Error in Diebold Voting Software Caused Lost Ballots in California County
Unique Transparency Program Uncovers Problems with Voting Software
Election Transparency Project volunteers:
Parke Bostrom - http://hum.dreamhosters.com/etp/news/20081204.html (main site)
Mitch Trachtenberg - http://www.mitchtrachtenberg.com/ourvotes.html (main site)
Tom Pinto - http://humtp.com/
John Gideon & Brad Friedman at BradBlog.com, 12/8/08 - 'Humboldt Transparency Project' Reveals Diebold, U.S. Federal E-Voting Scam
The BradBlog piece includes this link to an .mp3 of Crnich with Brad on the Peter B. Collins show on the afternoon of 12/5/08.
* * *The fact that Diebold/Premier did not take the action to recall the systems, actually puts them into a situation where they may very well have violated federal law. The Help America Vote Act of 2002 Title III Section 301(a)(5) mandates an acceptable error rate for voting systems in use in federal elections. That error rate, not counting any error caused by an action of the voter, cannot exceed 0.00001%.
Parke Bostrom's post above describes how "deck zero" became the batch of ballots that were handled properly by the elections department, and yet vanished from the final certified total. He comments further that the audit log for the Diebold GEMS central tabulation software matched the wrongly decreased total:
However, in the case of the Humboldt County vote count, the error rate was 0.31%.
We have asked both the Secretary of State of California and the EAC if they plan to take action by asking the US Attorney Office to investigate this seemingly clear violation of federal law. Neither the CA SoS, nor the EAC has yet replied to our queries on that matter.This means the audit log is not truly a "log" in the classical computer program sense, but is rather a "re-imagining" of what GEMS would like the audit log to be, based on whatever information GEMS happens to remember at the end of the vote counting process.
This demonstrates the system will cover its tracks when reporting an inaccurate result, destroying assurances of built-in memory redundancies and making a mockery of logic and accuracy testing. Not just here, everywhere. Frankly this is just another example of something we've known a long time.
Crnich herself has been very interesting through all of this. In the "Serious Error..." article above, Zetter reports:Crnich told Threat Level the issue has made her question her confidence in the voting system because, even though the company provided officials with a workaround, the problem indicated a fundamental flaw in the company's programming. She said she'd heard a lot of stories from other election officials about problems with voting machines, but never thought they applied to California.
Crnich losing confidence of course should be music to our ears. She also said a great thing in the interview with Peter B., explaining why she's been willing to work with citizen volunteers. As Humboldt County Clerk/Recorder and Registrar of Voters, Crnich is an elected official and I'm glad she acknowledged a responsibility to listen to constituents.
"I've always sort of listened to those anecdotal incidents with a jaundiced ear because California has some very stringent requirements of election systems that are in use here as well as some very strict security procedures and I didn't think those things affected us here," she said. "But this has sort of put a cloud over any confidence that I had in the Premier equipment that's been in this department since 1995."
In all, the media coverage above practically lionizes Crnich, which I think goes too far. Consider this analogy. Someone builds a fire in the middle of their bedroom and burns down the house. Would this person be praised for the wisdom of having an insurance policy? Using secret corporate vote counting computers, whether by Diebold or any other vendor, is playing with fire.
I've been unable to reach Crnich by phone in the past two days, repeatedly getting voice mail that could not accept more messages.
Also today, The North Coast Journal came out with Hank Sims' "Town Dandy" column called Deck Zero. Sims writes in reference to the known failure of the GEMS central tabulation software:The fact that Diebold/Premier let it stand for over four years, potentially undermining the first principle of American democracy, is an absolute outrage. These people should be shunned. Maybe indicted.
Throw in a little validation from the T-S editorial board...:They were loud, and they were strident in proclaiming that they didn't trust election technologies as much as they trust the ability of actual human beings to count votes.
...and it is starting to sound like we may be at a tipping point here. You might expect me to be frothing about hand-counting paper ballots right about now. You'd be wrong. Thinking as an organizer, I would hope now to establish three things that would be widely agreeable throughout the community:
The recent discovery, thanks to the Humboldt County Election Transparency Project, of a discrepancy in election results due to flawed software reveals that these activists were right to make noise, and right to complain about a company that has been less than responsible in dealing with the problem.
That said, if this is the nature of the opportunity now, I will re-offer to the community the materials I've developed to evaluate hand counting, most notably the forecast tool (spreadsheet) for estimating time, cost and labor needs for hand-counting in the precinct on election night. Back in the summer of 2007, when I first made this public, Sims noted: "Initial twiddling with the numbers suggests that it wouldn't be all that time-consuming or costly -- and wouldn't you rather wait a few days and spend a little more for a trustworthy count?"
I'd like to see more consistency in Sims' election integrity advocacy. And bottom line, I hope he'll push for a thorough examination of Diebold alternatives, as I'm sure Transparency Project volunteers will have other preferences and ideas to contribute to what could become the most envied process and dialog in the country.
This is all another way of saying "what would be better" is an inclusive and engaging community dialog aimed at literally defining "better" than Diebold.
Permalink:
http://wedonotconsent.blogspot.com/2008/12/humboldt-at-tipping-point-who-dares.html
Labels: Brad Friedman, Carolyn Crnich, Diebold, Eureka T-S, Hank Sims, Humboldt Transparency Project, Kevin Collins, Mitch Trachtenberg, North Coast Journal, Parke Bostrom, spreadsheet tool, Tom Pinto
Tuesday, January 01, 2008
Others Criticize Humboldt's So-Called "Transparency Project"
In my last post, I took apart a December 24 Eureka Times-Standard article (archive). As promised, I have submitted to the paper a My Word opinion column that reflects and expands upon what I posted here a few days ago. I think the T-S has a policy of allowing writers a My Word column every six months. In that case, there may be a problem getting my full essay in the paper right now (previous column 8/16/07, (archive)). However, since it is a critique of such egregious problems with their article, I would hope the paper would open itself up to such criticism. If not, I strongly suspect one of Humboldt's other newspapers will be glad to shine the light.
This morning on the T-S website I found a letter to the editor dated 12/30. It is from S. Franke in McKinleyville, and I've posted it in full at the bottom of this post to show who else besides the Voter Confidence Committee is criticizing the so-called Humboldt Transparency Project.
Before I get to that, however, I want to offer one of the ways that the new My Word column expands on my last blog post. I originally pointed out that the T-S article cited Bev Harris of BlackBoxVoting.org as a Project supporter. What I did not realize at the time was that Harris was about to release a statement saying she was misquoted and actually opposes the Project (full statement follows, italics emphasis added to indicate what I quoted in the My Word column):Black Box Voting has been misquoted twice in a week, first in an AP article that said Brunner had consulted us before developing her recommendations (she hadn't, and we disagree with her recommendations) and now saying we are behind the latest voter confidence / read: voter con idea.
Permalink:
The concept of providing ballot images to the public after running them through an intermediary program developed by David Dill (or anyone else!) is absurd and misses the point entirely. What is it about these guys that they just cannot RESIST inserting "An Expert" in between "The People" and "Our Ballots"?
At this point the only place to try ballot images would be in contests like the primaries, just days or weeks away, and only in situations where there is no other alternative but a high volume of centrally counted ballots, like in Los Angeles where there will be nearly two million mail-in ballots to count on Feb. 5, just five weeks from now. And then, the only point in it would be if it was done in public, on video, with a separate non-counting off the shelf scanner, and the CDs were released instantly.
So along comes Dill. "Hey, just stick my handy dandy overlay over the ballot images before you give them to the public, it will manipulate --- strike -- "remove stray marks" -- from the ballot images.
One more concept bites the dust. Amazing how easy it is to take people's eye off the ball, which is CITIZEN oversight, not "Dill's oversight."
http://www.times-standard.com/letters/ci_7841283
(archive)
Dealing with vote accuracy
Letters to the editor
Article Launched: 12/30/2007 01:15:19 AM PST
Regarding your recent story about County Elections offices ("As primary approaches, election offices are in turmoil," Dec. 24), I don't see the significance of scanning and posting voted ballots on the Internet.
First of all, the idea is getting old and with no funding, probably dead. Second, it does not address fraud because the scanned images will only be the sanitized end result of the process. Third, we could not be sure that all the ballots were there and that the software to sort them was accurate.
While playing with the ballots might be fun, it certainly would not add to the accuracy and security of our elections and would be a frivolous waste of money. It will only add another layer of complexity, opportunity for hacking, and falsely assure the public.
I would like to hear more about the new auditing procedures that should have been instituted -- statistical analysis, security measures, etc., as directed by the secretary of state. Most election offices already have internal auditing procedures and trained employees in place.
I don't believe one thing will confirm the accuracy of all the points of possible errors or fraud that can take place. We need to independently check every step of the voting process carefully and not just depend on a vague promise of some future fix.
S. Franke
McKinleyville
http://wedonotconsent.blogspot.com/2008/01/others-criticize-humboldts-so-called.html
Labels: Bev Harris, Black Box Voting, Eureka Times-Standard, Humboldt Transparency Project, S. Franke, Voter Confidence Committee
Saturday, December 22, 2007
A Missed Opportunity
Almost a month ago I wrote a post called Time To Check Your Least, in which I said I wasn't going to beat myself up for not writing enough. Since then I've posted more often than in quite some time. This reinforced the notion explained in that post that commitment to the least you can do philosophy will inherently increase what the least can amount to. I have to take this same outlook in considering yesterday's appearance on the Peter B. Collins show as a missed opportunity.
Click here for the full .mp3 archive of the show. Drag the slider of your media player to the start of the third hour and listen to Peter B. and Brad Friedman question Humboldt Registrar of Voters Carolyn Crnich about the so-called Humboldt Transparency Project. I was not given an opportunity to respond or comment on this, but I would have only repeated the concerns I have raised previously, including in this Eureka Reporter OpEd from July 6, 2007.
In summary, the Transparency Project would allow the public to view pictures (scanned images) of all ballots cast to perform a recount in whatever way one chooses. However, a secondary computer count is promoted as the way to go, although this completely defeats the purpose of checking the official, original count done by secret computer programming; and it relies on post-certification auditing which recent history shows will have no chance at overturning an inaccurate outcome. In addition, to do this ballot image recount without a secondary computer count would be entirely unwieldy and impractical, without even the prospect of proper observation or witnessing typically associated with post election hand-counting or manual auditing.
Not getting to respond to that wasn't really the crux of my missed opportunity. After returning from a commercial break, Peter B. prompted an exchange between the Registrar and me. This is where I blew it yesterday. The Registrar questioned the assumption of six seconds to count each contest per ballot. I noted that this is a variable that can be changed by the user of the hand-count forecast tool, but I then veered off into discussing the New Hampshire origins of that data point. What I should have said is that we have been trying for months to get assumptions from the Registrar herself plugged in to the forecast tool so that the community would have the opportunity to judge the Voter Confidence Committee's (VCC) hand-counting proposal based on the Registrar's official assumptions. Immediately after the segment had ended I realized my missed opportunity and e-mailed Peter B. with one last thought...
Peter B. took a call at that point from Jim Lamport, another regular attendee of the Registrar's monthly Election Advisory Committee meetings here in Humboldt. I have no problem with Lamport's defense of the Registrar, especially since I was not attacking her. I do respectfully disagree with Lamport's assessment that computers are ideally suited for vote counting. Since my phone line into the broadcast was already cut off, I was very glad to hear Brad Friedman challenge this position. Upon concluding that call, Peter B. did me a real solid and read the e-mail I submitted, quickly summarizing that the VCC is in search of official numbers from the Registrar precisely so the public can objectively judge our proposal. This was a save, in my view, salvaging to some extent the opportunity I see as missed.
There was another aspect of the discussion I wish I had handled differently. I noted that New Hampshire counts votes by hand in 45% of its polling places and the Registrar replied that this leaves a 55% majority of polling places being counted by optical scanners. Of course this is true, and I explained it as a function of jurisdiction level choice as opposed to being a state mandate. The opportunity I missed was not asking the Registrar whether Humboldt voters have had an opportunity to make the same choice, which of course we have not.
Earlier in the day, several hours prior to the radio show, the Registrar left me a voice message in response to the two messages I had left her in the two days prior. She mentioned that she and Kelly Sanders in the Humboldt Elections Department were busy gathering up all the data the VCC requested last week in an open letter. When these numbers are provided, yesterday's missed window of opportunity will open once again.
Permalink:
http://wedonotconsent.blogspot.com/2007/12/missed-opportunity.html
Labels: Brad Friedman, Carolyn Crnich, hand-counting paper ballots, Humboldt Transparency Project, Jim Lamport, Least you can do, Peter B. Collins, spreadsheet tool, VCC
Sunday, July 01, 2007
VCC Report Covered in Sunday Eureka Reporter
Online within the past hour or so, Sunday's Eureka Reporter has an article on the recently released Voter Confidence Committee Report on Election Conditions in Humboldt County (.pdf).
As in the past, I would say again that elections beat writer Rebecca S. Bender presents a fair piece with quotes from the VCC press release and one item straight out of the report, balanced with quotes from Registrar of Voters Carolyn Crnich. Bender does err in referring to the 20-page report as a 30-page report, but I suspect this will be quickly forgotten.
Of greater consequence, Crnich questions the relevance of some of the references in the timeline portion of the report. True, not everything in this section refers to equipment used in Humboldt. However, this series of brief paragraphs does include detailed information about the history of machine failures in Humboldt County, the results of hack tests on the same type of equipment used here, results of a statewide audit that found Humboldt was using uncertified software, the determination by the Department of Homeland Security that our central tabulator program, GEMS, is a threat to national security, and the finding that requisite independent testing of our equipment was not done.
You might argue we had enough with just those bits of info. But we felt it was important to place all of this against the backdrop of what was happening nationally, revealing that machines by other manufacturers were also failing or found to be highly vulnerable to tampering, that some elections counted by optical scanners were overturned by hand-counts, and that official reports by the GAO and Congressional committees found that votes had been lost and miscounted. This context is not irrelevant since we want readers to understand why the VCC would take a systemic approach to viewing and improving election conditions. No more secret vote counting!
In the Reporter article, Crnich attempts to marginalize the recommendations in the report by saying that the Humboldt Transparency Project makes most of them moot. With due respect to the Registrar, this is disingenuous. The Transparency Project allows for images of all ballots to appear on CDs so that any community member can do a recount and compare with the official results. The biggest problem with the implementation of this is that the CDs won't be available until after the election has been certified.
It is vital to get it right on election night because the prospects of later getting election results changed are prohibitively small. Just look at the various challenges to questionable Congressional results from last November. The VCC report even points out that Crnich is at greater risk by allowing the post-certification audits than she would we be if such review were possible before certification. The VCC does not oppose the Transparency Project but we do have some reservations about how it is to be deployed.
The Reporter article accurately quotes from the report: "The Transparency Project and hand counting are not mutually exclusive. The Transparency Project would be a welcome backup method of verification for hand counting, but it cannot make any secret vote-counting machines acceptable."
Permalink:
http://wedonotconsent.blogspot.com/2007/07/vcc-report-covered-in-sunday-eureka.html
Labels: Carolyn Crnich, Eureka Reporter, Humboldt County, Humboldt Transparency Project, Rebecca S. Bender, Report on Election Conditions, Voter Confidence Committee
