Even Congressman Rush Holt (a recognized national expert/leader in the movement to improve the security of voting machines in the wake of the 2000 Bush v Gore debacle) showed up to testifyin opposition to the proposed bill. A bill that would have, yet again, delayed compliance until November of 2010 and would have given us more "testing" of Sequoia's "Rube Goldberg" machines that retrofit 21st century printers to work with machines that use technology from the 1980's.
I remember hearing Senate President Sweeney threatening to simply force all New Jerseyans to vote on the modified Sequoia's in 2009 if he didn't get his way. The bill was on a greased path to passage. The "fix" was in.
But something wonderfully weird happened! The senators actually started asking some good questions and engaging in a real debate/discussion on the floor! It seems that some of them actually paid some attention to Holt's testimony and that of the VVPB advocates!
As far as I can tell the only place the video of that vote showed up was here on Blue Jersey and that's only because I used my hand held Casio to catch the last 14 minutes off my computer screen as they were voting.
It would be terrible if the Department of State was allowed to implement Sweeney's threat to use last months dramatic rejection of a flawed and foolish bill as an excuse/pretext to implement the aims of that same dumb ass bill!!!
The criteria for quickly certifying optical scan machines can be legislated into existence. We should have zero obligations to Sequoia. They've had three years to deliver and have failed.
What follows is the full text of Congressman Rush Holt's testimony...
Statement of U.S. Representative Rush Holt To the New Jersey Senate State Government Committee
December 11th, 2008
Committee Meeting: December 11, 2008
With Respect to Consideration of S. 2380
Chairman Scutari, Vice Chair Weinberg, and honorable Members of the New Jersey Senate State Government Committee, thank you for giving me the opportunity to address you today on the matter of S. 2380, a bill to remove the requirement that voting machines produce voter-verified paper records by January 1, 2009, and to replace that requirement with a pilot program for paper records or ballots. I want to commend Assemblyman Reed Gusciora for his leadership and commitment in securing passage of New Jersey's paper record requirement in 2005, and to express my very deep disappointment and concern not only that New Jersey has failed to implement it, but is now considering possibility of abolishing the requirement and its timetable for implementation altogether.
Voting must not be an act of faith, it must be an act of record, and that is why we must implement requirements that make computer-assisted elections independently auditable, an d we must do it without further delay. I will explain my concerns in detail below, but let me dispel some possible misconceptions at the outset. First, as you know, I am a physicist, and so I am not arguing in favor of paper-ballot-based voting out of some fear or lack of understanding of the technology we vote on. Second, the original group of experts who helped me draft my legislation when I first introduced it in Congress in 2003 were computer security experts - among the best and most highly-credentialed computer security experts in the country. Therefore, I would also like to think it is obvious that the driving force behind my legislation is not a lack of understanding of computer security risks, but rather a long experience and familiarity with computers, computer security and computer experts. And finally, as you may recall, I have personally experienced human error in vote counting: in my very first run for the seat I now hold, one of the county clerks in my district ascribed my vote totals to my opponent, and newspapers reported that I had lost the race. In fact, you might even say it runs in my family, because my own father was the apparent victim of the theft of paper ballots when he ran for office. So I am not operating under the assumption that human beings are automatically more reliable than computers, nor that paper ballots are fraud-proof and computer tallies are not. The point is - voting must not be an act of faith, it must be an act of record, and independent audit records (voter verified paper ballots) must be required.
New Jersey enacted such a requirement in 2005. But inexplicably, although more than half of the country has succeeded in implementing such requirements since I first commenced this effort in 2003, New Jersey - once a national leader - is slow to act. Do people in those other states know something we don't? Say what one will about Frank Hague and wandering paper ballots, if it had been pocket-sized memory cards or cartridges we were using back then, that's what would have wandered off. Or software would have been modified, if that is what we were using. Whatever the ballots are recorded on, theft is possible and rigorous chain of custody must be required; this is just as true for memory cards and cartridges as it is for paper ballots.
Jurisdictions across the country upgraded their voting systems between 2004 and 2006 in response to the Help America Vote Act (HAVA), and touch screen deployment increased from 21.5% to 36.6% of counties during that period in part because touch screen machines were specifically recommend by HAVA. However, as reported by the voting systems information clearinghouse Election Data Services in its report on voting equipment usage in 2008, "[a]fter nearly three decades of consistent growth in their use with each election, nearly 10 million fewer registered voters will be using electronic voting equipment in the 2008 general election comp ared to just two years ago. Every county that has changed voting systems since 2006 has moved to optical scan equipment" [my emphasis added]. In 2008, 92.5 million registered voters resided in jurisdictions using optical scan equipment, while only 55 million resided in jurisdictions using touch screen voting equipment (much of which was equipped with voter verified paper audit trail (VVPAT) printers), and the trend is clear. Although the optical scan systems are not perfect - and no voting system is perfect - the fewer problems with optical scan systems and - most important - the ease of verification by each voter and the probability of meaningful audits and recounts make that the system the choice of most states, even states that like New Jersey, at first moved toward electronic, touch screen voting machines.
The national trend is clearly to move away from touch screen voting and towards more reliable, less troublesome optical scan voting systems. There are only seven states left in the country that, on a statewide basis, use paperless voting systems. Regrettably, and despite passage of the Gusciora bill more than three years ago, New Jersey is still one of them.
Based on what other states are doing, New Jersey could certainly have implemented its paper record requirement as of the original implementation date of January 2008. And it could and should do so now, without further delay.
What is the best choice? From a technical standpoint, touch screen20voting machines - whether outfitted with voter verified paper record printers or not - suffer from certain performance problems that optical scan systems do not. For example, since 2004 incident reporting systems have produced an endless stream of reports of calibration problems on touch screen machines - so called "jumping Xs" - in which the voter will touch the screen for one candidate only to find his or her choice register for another. Last month, that reportedly even happened to Oprah Winfrey. This cannot happen with an optical scan system, because voters mark paper ballots, rather than touch a temperamental screen interface that may or may not be properly calibrated. Similarly, when a touch screen machine malfunctions, even if it is equipped with a so-called VVPAT printer, voters cannot vote unless they are given paper ballots. This problem delayed our Governor from voting not long ago, and evidently prevented some others in his district from voting. With an optical scan system, all voters are given paper ballots anyway. If the scanner jams, the completed ballots can simply be placed in a ballot box for later counting.
Optical scan systems are less problematic and more reliable, and they are more cost effective. Today, if the job can be done at least as effectively and reliably but at lower cost, that should be the choice we make. And that is the case with optical scan systems as compared to touch screen systems. That is what other states have determined20 evidently.
First, fewer items of equipment are needed if optical scan systems are used instead of touch screen systems. The act of voting takes time -- often many minutes per voter. When touch screen machines are used, voters engage in the act of voting on the machine itself. To avoid waits counties provide multiple voting machines. When optical scan systems are used, voters engage in the act of voting in a private marking station using a paper ballot. They only occupy the optical scan machine when feeding the completed ballot into it, which takes seconds. The optical scanner will identify over-voted ballots or ballots with improper marks, and will tally only ballots correctly marked. Even if the machine rejects the ballot, that too only takes seconds and the voter can return to the marking station to correct the ballot. Therefore, while touch screen systems were deployed at the rate of one per every 250 to 850 registered voters in New Jersey in 2004 according to New Jersey's report to the Election Assistance Commission (EAC), jurisdictions using optical scan systems often deployed their optical scanners at the rate of one per thousands of voters. No matter how many voters are assigned to an optical scan election district, it will likely need only one optical scanner and one accessible ballot marking device, for a total of two items of equipment. In 2004, according to the EAC survey, the average number of touch screen machines per polling place was 4.4 in 0AGloucester County, 4.5 in Sussex County, and 5.3 in Salem County; 18 of New Jersey's 21 counties used more than two voting machines on average per polling place. ??In purchase costs alone, converting to an optical scan system would likely cost no more than adding printer retrofits to New Jersey's existing touch screen machines, and probably less. (By the way, get independent estimates; beware of vendor claims.) According to New Jersey's currently-posted Voting Equipment Inventory, there are approximately 11,200 touch screen voting machines in the state. The estimated cost of a printer attachment is $2,000. In 2005, the Office of Legislative Services produced two separate fiscal estimates for the cost of retrofitting the machines, one for $26.4 to $39 million, which assumed that more than 3,000 of the underlying machines would have to be replaced entirely to meet the paper record requirement, and two months later an estimate for $21.4 million, which assumed that only 209 of the underlying machines would have to be replaced entirely to meet the paper record requirement (although almost 3,000 new machines were still needed to meet HAVA disability access requirements). As I understand it, the printer retrofit recommended by the Title 19 Committee only works with the AVC Advantage Model D-10, which is not yet in use in any county. By reference to a recent contract between Sequoia voting systems and a New Jersey County, the price for a headset-equipped Sequoia touch screen machines is $8 ,000. Therefore, arguably, every county will need not only new printers at $2,000 each, but also new touch screen machines at $8,000 each. The total cost could be as high as $112 million.
In contrast, if New Jersey converted to optical scan systems, the total cost would be closer to $36 million. With optical scan systems, only two items of equipment would be needed per election district, and I have been advised that advanced optical scanners and ballot marking devices retail for approximately $5,000 each. New Jersey has approximately 3,600 election districts. Therefore, two items of equipment at a total cost of $10,000 for each of New Jersey's 3,600 election districts would put the cost of converting to an optical scan system at approximately $36 million. In addition, fewer items of equipment translates to costs savings not simply in the purchase, but also in the storage, programming, use, maintenance and transportation.
Finally, the printer retrofit recommended by the Title 19 Committee does not appear to provide a way for disabled voters to verify the accuracy of the paper printout from the paper printout itself, rather than from the internal memory. The most current version of the EAC's Voluntary Voting System Guidelines provides that: "[i]f state statute designates the paper record produced by the VVPAT to be the official ballot or the determinative record on a recount, the accessible voting equipment shall provide features that enable visually impaired voters and voters with an unwritten language to review the paper record. . . . [f]or example, the accessible voting equipment might provide an automated reader that converts the paper record contents into audio output."
In the New Jersey law, as in my federal legislation, if there is a discrepancy between the electronic record and the paper record, the paper record is the vote of record. This is as it should be, because that is the only tangible record of the vote verified by the voter. Therefore, each voter must be able to verify that the contents of the paper printout accurately reflect his or her choices. The New Jersey law specifically requires that the paper records "be made available for inspection and verification by the voter at the time the vote is cast," and that includes disabled voters. ??Advanced ballot marking devices used with optical scan systems meet this requirement, by allowing voters to use audio and other features to mark a paper ballot, and also to convert the contents of the paper record into audio output, to confirm to the voter that the ballot accurately reflects his or her choices. There are no VVPAT printers currently on the market that do this, and nothing I have seen in the descriptions of the printer attachment recommended by the Title 19 Committee reflects that it does that. Therefore, if New Jersey chooses to deploy VVPAT printers instead of converting to an optical scan system, even after spending as much as $100 million or more , New Jersey will find that it is back to square one in terms of meeting HAVA's accessibility requirements.
I would urge you not to proceed with any legislation that removes New Jersey's requirement for paper-based auditable voting, but rather to honor New Jersey's paper record requirement by taking such action as is necessary to deploy optical scan voting systems accompanied by accessible ballot marking devices across the state as expeditiously as possible. Thank you again for your time and consideration. I look forward to continuing to work with you to protect the accuracy, integrity and security of voting systems in New Jersey. --
