1 2 3 4 U.S. ELECTION ASSISTANCE COMMISSION 5 PUBLIC MEETING 6 7 8 Taken at the Marriott Hotel 9 1535 Broadway, New York, New York 10 11 On the date of Thursday, June 30, 2005 12 13 14 15 COMMISSION MEMBERS: 16 Gracia Hillman, Chair 18 19 20 21 Start time: 11:00 o'clock, a.m. 22 Taken before: ELLEN REACH, a court reporter 1 TESTING LABORATORIES' PERSPECTIVES 2 ON THE PROPOSED VVSG 3 4 CHAIR HILLMAN: If we could get 5 assembled, please, so the hearing can begin. 6 We are going to begin this hearing. 7 If I could remind you please to turn 8 off your - or at least silence - your cell 9 phones, pagers and any other electronic devices 10 that might make noise and distract from the 11 proceedings. 12 Thank you very much. 13 If I could please ask everyone to 14 please stand and join me in the Pledge of 15 Allegiance. 16 (Assembly in Pledge of Allegiance) 17 CHAIR HILLMAN: If we might have the 18 roll call, please. 19 MS. THOMPSON: Certainly, Madam 20 Chair. 21 Commissioners, please respond by 22 saying here or present when I call your name. 1 Gracia Hillman, Chair? 2 CHAIR HILLMAN: Here. 3 MS. THOMPSON: Paul DeGregorio, Vice 4 Chairman? 5 VICE CHAIR DeGREGORIO: Here. 6 MS. THOMPSON: Ray Martinez, 7 Commissioner? 8 COMMISSIONER MARTINEZ: Here. 9 MS. THOMPSON: Madam Chair, there 10 are three members present. 11 CHAIR HILLMAN: Thank you. 12 Commissioners, we have before us the 13 agenda for today's public hearing. If there are 14 no adjustments to the agenda, it would be 15 appropriate to move adoption. 16 COMMISSIONER MARTINEZ: Move, Madam 17 Chair. 18 VICE CHAIR DeGREGORIO: Second. 19 CHAIR HILLMAN: Thank you very much. 20 We have our agenda adopted. 21 The purpose of today's hearings is 22 for the Election Assistance Commission to receive 1 testimony and comment on the proposed Voluntary 2 Voting System Guidelines. 3 For those who were with us earlier 4 this morning at our meeting, you did hear several 5 minutes of discussion about the guidelines, our 6 acknowledgement of the work of the Technical 7 Guidelines Development Committee that produced a 8 recommendation and delivered that recommendation 9 to us on May 9th. 10 Our former Interim Executive 11 Director, Carol Paquette, will make a brief 12 presentation that will put this hearing in the 13 context of where the Election Assistance 14 Commission is today with respect to the proposed 15 guidelines that we have now made public. 16 Ms. Paquette. 17 MS. PAQUETTE: Yes, Madam Chair and 18 Commissioners, thank you very much. 19 Section 202 of HAVA directs the 20 Election Assistance Commission to adopt Voluntary 21 Voting System Guidelines and to provide for the 22 testing, certification, decertification and 1 recertification of voting system hardware and 2 software by accredited test labs, among many 3 other duties that are assigned to the Commission. 4 To assist the Commission with the 5 work of developing the guidelines, HAVA also 6 provides for the establishment of a body of 7 subject matter experts, called the Technical 8 Guidelines Development Committee, otherwise known 9 as the TGDC. 10 The composition of this 15 member 11 committee is drawn from a variety of 12 organizations, again, and specified by HAVA. 13 These organizations include the National 14 Association of State Election Directors, the 15 Access Board, the IEEE, the American National 16 Standards Institute, and other representatives as 17 designated in the statute. 18 The TGDC is chaired by the Director 19 of the National Institute for Standards and 20 Technology, NIST, and NIST is charged with 21 providing technical support to the work of the 22 TGDC. 1 In this fiscal year, EAC has 2 provided nearly $3 million to fund this work of 3 the TGDC and NIST. 4 The statute tasks the TGDC with 5 developing recommendations for Voluntary Voting 6 System Guidelines and providing those 7 recommendations for the consideration of the 8 Election Assistance Commission. 9 The TGDC and NIST have been hard at 10 work at this task for nine months, the time 11 period which is allocated by HAVA, and have 12 delivered a very good product to the EAC for us 13 to begin our review on May the 9th. 14 The recommendations provided to the 15 Commission by the TGDC essentially augment the 16 2002 Voting Systems Standards, most particularly 17 in the areas of accessibility, usability and 18 security. 19 This includes new requirements for 20 accessibility, voting system software 21 distribution, system setup validation and the use 22 of wireless communications. 1 Since some states have required the 2 use of voter verified paper audit trails, the 3 recommended guidelines also include performance 4 specifications for this capability. 5 We would note for the Federal 6 certification purposes of voting systems, these 7 are optional. They have been included for use by 8 those states that have added this as an 9 additional requirement for their voting systems. 10 We would also note that paper audit 11 trails is only one of several technical 12 approaches to providing voters an additional 13 means to verify their ballot in addition to the 14 DRE summary screen. And the Commission plans to 15 work with the TGDC and NIST to address the 16 development of similar specifications for audio, 17 video and cryptographic approaches to voter 18 verification. 19 Having received the recommendations 20 on May 9th, the EAC has been, in the intervening 21 weeks, performing its due diligence requirements 22 as the responsible Federal organization for HAVA, 1 to thoroughly review these recommendations. 2 In particular, we have performed a 3 legal analysis of the proposed guidelines and 4 recommended guidelines to ensure that they are 5 fully in compliance with the Help America Vote 6 Act. 7 This has resulted in some changes, 8 particularly in the accessibility provisions. 9 And I would note as an example requirements 2.2.5 10 and 2.2.6, dealing with voter verified paper 11 audit trails, that if the normal procedure is for 12 voters to submit their own ballots, then the 13 accessible voting station needs to provide the 14 same capability for voters who are blind. 15 That is an example of where the 16 Commission changed a 'should' recommendation to a 17 'shall', to be in conformance with Section 301 of 18 HAVA. 19 In addition, the Commission has 20 revised selected portions to reflect the new EAC 21 process to carry out the HAVA mandate for the 22 national certification of voting systems. We 1 have also made some updates throughout the 2 document to include the new terminology that has 3 been specified by HAVA. 4 For example, the testing of voting 5 systems used to be called a 'qualification' 6 process and HAVA now uses the the term 7 'certification' process. 8 We have made these changes in the 9 document because this is a document that will 10 have a substantial life in being used for the 11 testing and certification of voting systems, both 12 at the Federal level and at the state level. 13 That was the principal rationale for the changing 14 of terminology, to put in an appropriate floor 15 for the terminology that will be used as this 16 document is used as a reference for testing 17 purposes in the coming years. 18 We certainly expect to have many 19 more changes coming in over the course of the 20 summer and we will certainly be summarizing at 21 the end of that comment period the further 22 changes that have been made to the document and 1 how the various comments have been handled. 2 The Federal Register Notice on the 3 availability of the guidelines for public comment 4 was published yesterday, June 29th. And that 5 begins the 90 day public comment period. 6 In the Federal Register Notice and 7 also on our web site we provide fairly extensive 8 information on how parties can obtain a copy of 9 the document and also how they can provide 10 comments. 11 The document is a very substantial 12 document. It's about 250 pages. And to 13 facilitate ease in getting the document, we have 14 published it on our web page. It is available in 15 both PDF and HTML formats, so that those parties 16 who use assistive devices will be able to access 17 the document. 18 We have also posted it in a manner 19 so that the entire document can be downloaded or 20 selected sections can be downloaded, again, to 21 make it easier for people to deal with this 22 fairly substantial document. 1 We will also provide copies in hard 2 copy or on CD-ROM. 3 If people wish to send requests by 4 e-mail or by mail to the Commission, we have 5 provided a number of mechanisms for people to 6 submit comments. 7 On our web page we have included an 8 application so that people can log on to the web 9 page and go to the comment screen and provide 10 their comments directly to the web page. 11 Supporting that comment page is a 12 data base where all the comments will be stored 13 and allow for the comments to be posted and for 14 anyone to also look at comments that have been 15 provided to the document. 16 We have established a special e-mail 17 address: votingsystemguidelines@EAC.gov, for 18 those individuals who prefer to send us comments 19 by e-mail. 20 We, of course, will also receive 21 comments by mail. We ask that those be provided, 22 addressed specificaly to Voting System Guidelines 1 Comments, so that they can identified and handled 2 expeditiously. 3 We are providing, in the data base 4 of comments, the ability for the comments 5 received by e-mail and by mail to also be entered 6 into the data base so that we will have one 7 reference source, where comments from all sources 8 will be collected and posted, both for our own 9 review and analysis, as well as the review by 10 other interested parties. 11 I would also note that we are having 12 three hearings in addition to the one just 13 convened. We are holding a second hearing on 14 July 28th at the California Institute of 15 Technology, and a third hearing in August in 16 Denver. 17 And further specifics will be 18 forthcoming as we get that meeting planned. 19 Are there any further questions? 20 CHAIR HILLMAN: Before I turn the 21 mike over to my colleagues, can you just state 22 for the record where the California Institute of 1 Technology is located. I know it's in 2 California, but -- 3 MS. PAQUETTE: It's in Pasadena. 4 CHAIR HILLMAN: Thank you very much. 5 Commissioners, before I turn the 6 mike over to you, I just want to acknowledge that 7 we have with us today members of the Technical 8 Guidelines Development Committee and from the 9 National Institute of Standards and Technology. 10 And I would just like to acknowledge 11 from the TGDC Sharon Turner-Buie, who is a member 12 of the EAC Board of Advisors and Director of 13 Elections for Kansas City, Missiouri. And we 14 also have Brit Williams, who is with the Center 15 for Election Systems - I probably messed that up, 16 Brit, but forgive me - at Kennesaw State 17 University, in Georgia. Thank you for being with 18 us. 19 And from the National Institute of 20 Standards and Technology, which is a part of the 21 U.S. Department of Commerce, we have Craig 22 Burkhardt, who is actually counsel with the 1 Department of Commerce, and has been working with 2 NIST on the development of the guidelines, the 3 proposed guidelines, and Allan Eustis, who is 4 with the National Institute of Standards and 5 Technology. 6 An important part of the comment 7 period is that the EAC Board of Advisors, which 8 is made up of 37 members, and the EAC Standards 9 Board, which is made up of 110 members, will each 10 go through a process to review the proposed 11 guidelines and offer comments to us. 12 And we have with us, in addition to 13 Ms. Turner-Buie, we have Wes Kliner from, I 14 believe, Tennessee, but I thought I heard you 15 moved, so I am going to acknowledge Tennessee. 16 Thank you very much. As a member of the Board of 17 Advisors, thank you very much for being with us. 18 And I think we might have a member 19 of the Standards Board still with us, Mr. 20 Szczesniak - is Ed Szczesniak still here - from 21 the Standards Board. Thank you. 22 This has been a very wonderful labor 1 of love or labor of something that we have been 2 through to produce these guidelines. So if we 3 get emotional about this, it's because we are 4 very, very attached to both the process and the 5 document. 6 And, Mr. Vice Chairman, it would be 7 appropriate if you would like to give some 8 comments to or ask some questions of Ms. 9 Paquette. I know your name, Carol. 10 VICE CHAIR DeGREGORIO: I think 11 labor is an apt description since it was a nine 12 month process to get done. And we know that to 13 have a child it takes nine months. And this was 14 a big child of the EAC to come forth from the 15 Technical Guidelines Development Committee. 16 And I know I made some comments 17 earlier at our meeting this morning, a few 18 minutes ago. 19 But I do think that this has been a 20 very important process that the Help America Vote 21 Act spells out to give the voters of this country 22 a greater trust in our voting systems. And it's 1 the first time that the Federal Government has 2 put real money behind it. 3 Our Chair mentioned that nearly 4 $3 million has been appropriated by the EAC to 5 NIST for this process. And it is the first time 6 we have had some scientists look into this and 7 provide us their guidance. 8 And I think as we go through this 9 over the next 90 days we will see the fruits of 10 their work come forth as we see the human factors 11 and security and other aspects that have been 12 changed since the 2002 standards and that have 13 been enhanced since the 2002 standards were 14 developed, that have improved this product and 15 made it better for the public out there, election 16 officials. 17 And I look forward from hearing from 18 the testing laboratories, from the vendors and 19 from the advocates and from the public, and to 20 hear their feedback in this process as we 21 finalize this document and then move forward with 22 this, because we know that it's going to have a 1 major impact out there. And we want it to be a 2 positive impact. 3 So all the input that we can get to 4 make that happen is important to me as a 5 Commissioner, as has been since we started this 6 process in June 2004. 7 So thank you, Carol, for providing 8 that history of where we are with this and I look 9 forward to the testimony we are going to hear 10 today. 11 CHAIR HILLMAN: Thank you. 12 Commissioner Martinez. 13 COMMISSIONER MARTINEZ: Thank you, 14 Madam Chair. 15 I will be very brief. I do have one 16 quick question for Carol. But just as an 17 introductory statement, I guess I said earlier in 18 our public meeting portion of this day that this 19 is, I think, perhaps the most significant thing 20 that this Commission has done thus far in our 21 short 18 months or so of being in existence. 22 So my thanks to all of our partners 1 in helping to developing this first draft of 2 Voluntary Voting System Guidelines: our partners 3 at NIST and the Department of Commerce and, of 4 course, all of the members who served on a 5 voluntary capacity with the Technical Guidelines 6 Development Committee. We are so grateful and 7 indebted for your service and we thank you for 8 it. 9 I guess for me, there was a poll 10 that was taken a few weeks after the November 11 2004 election, which was not too long ago, by NBC 12 News and the Wall Street Journal, and it said 13 something like a quarter of all American 14 surveyed, that a quarter of those surveyed in 15 that poll did not believe in the accuracy of the 16 voting systems that they were using to cast a 17 ballot in November of 2004. 18 And as a Commissioner - really, as 19 an American - we should just be troubled, I 20 think, by such a high level of lack of confidence 21 in the very machinery and the tools that help our 22 democracy to continue to exist. 1 And that's why, again, I point to 2 this as being among the most significant 3 accomplishments that we have undertaken here at 4 this Commission. 5 So I hope, Madam Chair and Mr. Vice 6 Chairman, that at the end of this process what I 7 hope and I know will be a very transparent and 8 very open and, yes, at times perhaps even an 9 emotional process for us to go through, because 10 there are so many things at stake here, from 11 accessibility for individuals with disabilities, 12 to folks who believe strongly that there ought to 13 be mechanisms in place to ensure better accuracy 14 of ballots that have been cast -- 15 I think this is one of our 16 panelists, welcome. 17 So, Madam Chair, I would simply say 18 that I think we have undertaken a significant 19 project and it's one that I know that we will all 20 feel better about at the end of the process. 21 Carol, if I could, there was a 22 clarification I think that I wanted to make to 1 one of the things that you said. 2 You were giving an example of our 3 legal analysis and you were talking specifically 4 with the functionality of voting systems where 5 the voter has to cast the ballot aside from, for 6 example, on a DRE you can punch a button that 7 says 'cast ballot'. But there are other voting 8 systems, like optical scan systems, for example, 9 where the functionality of casting a ballot 10 requires a voter to do something other than 11 hitting a button on a touch screen. And you 12 refer to that as VVPAT. 13 But I think what you meant to say is 14 that we changed this 'should' to a 'shall' when 15 it comes to the functionality of those type of 16 systems and not necessarily when it comes to 17 VVPAT, because VVPAT contemplates that a voter 18 gets to see something behind a screen, but not 19 walk away with a particular ballot and have to go 20 cast it in another location. 21 So in describing that example, when 22 we made a change from a 'should' to a 'shall', 1 you said that "in the context of VVPAT," but I 2 think what you meant it in was the context of the 3 functionality, for example, an optical scan 4 voting system. 5 MS. PAQUETTE: You are correct, 6 Commissioner. 7 COMMISSIONER MARTINEZ: That's it. 8 Thank you, Madam Chair. 9 CHAIR HILLMAN: Today's hearing is 10 in four sections. We have Panel Number 1, 11 presentations from the testing laboratories. And 12 presenting this morning will be Mr. Joe 13 Hazeltine, Senior Director, Eastern Test 14 Operations with Wyle Laboratories, and Ms. 15 Carolyn Coggins, Director of ITA Operations at 16 SysTest Labs. 17 Following Panel 1, we will take a 18 lunch break. And when we reconvene at about 19 1:30, the second panel will make presentations 20 and these will be people representing vendors and 21 we will have presentations from people 22 representing ES&S, Vote Here, Sequoia Voting 1 Systems and AccuPoll. 2 Panel 3 will be presentations about 3 the specific section of the guidelines or 4 sections of the guidelines that address voter 5 verifiable paper audit trails and perspectives 6 will be presented by four individuals. 7 And then the final section, which I 8 don't know if the mike is - the final section 9 will be public comment period. And we do have 10 people who have registered in advance requesting 11 an opportunity to make a comment. 12 So, can people hear me? No, I 13 didn't think so. I am not sure. Here we go. 14 I am going to repeat that anyhow. I 15 was just explaining the make-up of the four 16 panels for today. 17 So thank you very much to Mr. 18 Hazeltine and Ms. Coggins. 19 And, Mr. Hazeltine, if we could call 20 on you first. 21 MR. HAZELTINE: Thank you, Madam 22 Chair and members of the Commission. 1 I have been asked to speak on the 2 Voluntary Voting System Guidelines, the impact 3 from the -- 4 CHAIR HILLMAN: The mike isn't on 5 here. 6 MR. HAZELTINE: Thank you, Madam 7 Chair and members of the Election Commission. 8 I have been asked to speak on the 9 impact of the Voluntary Voting System Guidelines 10 from a testing laboratory standpoint. And you 11 have a copy of my presentation. I am just going 12 to kind of go through that. 13 I guess, first, in talking about the 14 impact, one is to discuss what's the same in this 15 document as the 2002 Voting System Guidelines. 16 And when we look at Volume I, 17 Sections 2.1 through 2.2.6 and then 2.2.8 to 18 2.6 - which is System Functional Capabilities - 19 have remained the same. 20 Section 3 on hardware, Section 4 on 21 software requirements are the same. 22 Section 5 on telecommunications are 1 the same. 2 Section 6.1 to 6.6.4.3 on security 3 are the same. 4 Section 7 on quality assurance and 5 Section 8 on configuration management 6 requirements are the same. 7 So, the bulk of the document remains 8 the same as before. 9 In Volume II, we see even more - 10 Sections 2, 3, 4, 5, 6 and 7, which would be 11 Description of the Technical Data Package, 12 Functional Testing, Hardware and Software, System 13 Integration and Examination of Configuration 14 Management Practices - have all remained the 15 same. 16 Appendix A, the Description of the 17 National Certification Test Plan is the same, and 18 Appendix B, the Description of the National 19 Certification Test Report. 20 So, again, the bulk of the documents 21 remain close to as it was before. 22 But there been of some changes and 1 they are significant. 2 In Volume I, Section 1 has been 3 redone. Primarily it is new text on the program 4 that clarifies the requirements. 5 In Section 2.2.7 is probably the 6 largest change to the document, where there are 7 many new requirements on Human Factors and 8 compliance with HAVA 301(a)(3) were added. And, 9 certainly, that's an important revision to the 10 document. 11 Section 6.6.4 to Section 6.8.7.5 12 added many new requirements dealing with 13 security, wireless and the verifiable voting 14 paper audit trail. 15 Appendix A was added, which is a 16 very good thing, adding a glossary. 17 Appendix B, adding references, 18 again, a good thing. 19 Appendix C added best practices for 20 election officials and Appendix D added 21 independent dual verification systems. 22 We also added Appendix E, which was 1 the NASED Technical Guide, Number 1, dealing with 2 colorblindness and some of those issues. 3 Volume II had two changes. 4 Section 1, the revised National 5 Certification Testing Guidelines, shifting over 6 from NASED to the Elections Assistance 7 Commission. 8 And then Appendix C, where we added 9 the National Certification Test Design criteria, 10 the revision, in the back of that. 11 So kind of what was changed, in a 12 nutshell. 13 So if we look at what the impact is 14 to a testing labs, or to us, the Volume I Section 15 1 change, we are -- 16 CHAIR HILLMAN: Technology always 17 challenges us, always. 18 MR. HAZELTINE: We are back on. 19 Volume I Section 1, the 20 introduction, there is a new certification that 21 is required for independent testing authorities 22 that would now come through the National 1 Institute of Standards and Technology, through 2 their NVLAP program, so it's a new certification, 3 similar to what we already have. But it will 4 require some additional work. 5 In Section 2.2.7, Volume 1, Human 6 Factors Assessment, accesibility changes were 7 made. And certainly those are major changes with 8 a major impact to the work that we will have to 9 do. 10 In Volume 1, 6.6.4 of the software 11 security, that's relatively small changes, which 12 is mostly administrative. Their impact to us is 13 minor. 14 And 6.7 of Volume I, wireless 15 requirements, would have an impact if that 16 technology were implemented. There are several 17 places where expert staff is needed. There are 18 cryptiographic (sic) issues and other things 19 which would require quite a bit of additional 20 work. 21 In Section 6.8 of Volume I, the 22 Voter Verifiable Paper Audit Trail, I would 1 consider those to be relatively small. They are 2 administrative type changes. For us the impact 3 is not all that difficult from a testing 4 laboratory. 5 Certainly, Volume I, Appendix A, B, 6 C, with the glossary, references and best 7 practices, there is no real impact to us. It 8 certainly adds to the clarity of the document and 9 allows you to understand terminology being 10 addressed much better. 11 Appendix D, where the Independent 12 Verification Systems comes into Volume I, that 13 would be a major impact when implemented because 14 it really requires two systems. So you are 15 basically doing the testing twice, so it would be 16 an impact as far as the work. 17 Appendix E, the NASED Technical 18 Guide is small, mostly administrative changes for 19 us. 20 In Volume II, Section 1, the 21 National Certification Program Guidelines, there 22 were some clarification text added. There is no 1 real impact added to us. That is the same 2 practices the 2002 standard required. 3 Appendix C, the National 4 Certification Test Design Criteria, really no 5 impact to us in as far as conducting a normal 6 test program. If the specimen did fail, there 7 would be some additional testing requirements to 8 work with. But if the system passes, it would 9 pretty much would be a very similar process. 10 I am talking about possible issues 11 where we get into a little more detail of areas, 12 as we went through it where - I wouldn't say 13 concerned - we were just identifying. 14 We will need guidance on the 15 summative usability test required for partially 16 blind, blind, persons with limited motor control, 17 persons with limited command of English, and the 18 general population. 19 I have given some of the references 20 of where that would be in the document. But we 21 will need guidance on how to conduct those tests 22 in the future. 1 For the accessible voting station, 2 guidance with personal assistive devices is too 3 broad. We would need to know exactly what we are 4 talking about there and what devices could be 5 connected and what our requirements would be to 6 evaluate those. 7 I noted that the voter control of 8 contrast must be reset after the vote was cast. 9 They do have the ability to reset contrast. But 10 I didn't see anywhere else in the document where 11 it would reset back to some nominal value for the 12 next voter. 13 Synchronized audio with those 14 scrolling screens, that could be an issue with 15 testing. That would have to be defined exactly 16 what we mean there. 17 The Auto Tactical Interface 18 requirements will require more testing, 19 certainly, because it's a new feature. Certainly 20 a good feature, but a new one. 21 The audio qualification to ANSI 22 C63.19 would require more testing. That's where 1 we have to classify the actual microphones and 2 system ratings. 3 Our recommendation at Wyle would be 4 that we consider EIA/TIA 968, which was the old 5 FCC Part 68 document which regulated telephony 6 (ph) and has much - I don't know - I would say 7 are better rules - that are rules which are well 8 known, have been in place for a number of years, 9 does include hearing aide compatability and does 10 include a process that is somewhat better 11 understood by testing laboratories. 12 The guidance on testing cochlear 13 mounted hearing devices that we needed in here, 14 that could be an issue because I imagine that 15 there are a number of different types. So that's 16 an area of concern. 17 The frequency range of 315 Hz to 18 10 kHz, that requires now equipment of the new 19 testing techniques. Once again I would suggest 20 looking back at the ITA/TIA 968, the old FCC Part 21 68 document where the frequency range was not 22 quite so broad and test equipment already exists 1 for it. 2 Another document would be TIA 470, 3 which would address the acoustic performance as 4 well. 5 The rate of speech control where you 6 are allowed to, actually adjust or advance how 7 fast or how slowly the audio playback is would 8 require more testing and we would need some 9 guidance on that. It says within reasonable 10 levels. Well, reasonable can be interpreted in 11 different ways, so we would need some help on 12 that. 13 Confirmation of the blind paper 14 ballot would require some additional testing, 15 additional skill. We need someone who could read 16 those ballots to make sure the system is 17 performing properly. 18 The actuation force requirement test 19 required in test fixtures so that we can do, I 20 believe that is 22 newtons of force, so we would 21 need some additional devices to run that test. 22 We would need a wheelchair to be 1 provided so that we can evaluate the clearance 2 requirements in Section 2.2.7.4. And there are a 3 number of those. They are good requirements, 4 there are just a number of them. 5 And for the English illiterate voter 6 machines, would require interpreters be present 7 during some parts of the testing. The required 8 languages should be specified. Right now it's 9 quite broad and that would require some 10 additional work on our part. 11 Wireless documentation in accordance 12 to 6.7.2.1.3.1 - there are a lot of numbers 13 there - but requires a "subject area recognized 14 expert." So that would need to be identified and 15 provided to us. 16 If wireless systems currently don't 17 exist, but if they do, that would be be an issue. 18 The wireless system requires that it 19 be able to operate with or without the wireless 20 capability. So that would require us testing it 21 twice, with and without. 22 Encryption verification also 1 requires an expertise in the field and we would 2 need some help there. 3 Again, the Voter Verified Paper 4 Audit Paper Trail testing poses no significant 5 testing issue. It does involve more time, but 6 it's not a technical challenge for us. 7 Am I going too fast? 8 SIGNERS: We can't hear you. 9 CHAIR HILLMAN: If you could move 10 closer to the microphone. 11 MR. HAZELTINE: When I taught I 12 always had two hours of notes and one hour of 13 speech, so I learned how to speak quickly. 14 CHAIR HILLMAN: Mr. Hazeltine, we 15 know you live and breathe this stuff, but some of 16 us are running to catch up. 17 MR. HAZELTINE: I'm sorry. 18 CHAIR HILLMAN: No problem. 19 MR. HAZELTINE: Witness IDV Systems, 20 again, will have twice the amount of testing 21 because there are two devices. I had mentioned 22 that earlier. 1 The bottom line, from our viewpoint, 2 is to be Wyle's viewpoint, we consider the 3 Voluntary Voting System Guidelines to be a 4 significant improvement over the past two 5 documents, the 2002 and 1990 versions, 6 particularly in Human Factors, Accessibility and 7 Security. 8 It's quite well done. There are 9 some issues, but they are resolvable. 10 Prior to implementation, further 11 clarification would be required, but I don't 12 believe it's all that difficult to get that data 13 to us. 14 The current 2005 Voting System 15 Voluntary Guidelines that are out are in need of 16 some corrections and formatting, primarily in 17 correct references and page numbering. Section 18 2.2.7 is probably the one that would need the 19 most work there. 20 Independent testing laboratories do 21 have some cost requirements through the 22 accreditation through NIST which is not an issue. 1 It's just a thing to be noted. 2 We will need some guidance for 3 testing laboratories in several areas, which I 4 have identified in the presentation. 5 There will be some new testing. It 6 will add costs to the certifying of the voting 7 machine. 8 From Wyle's standpoint we are 9 already training our staff to the new document. 10 I know it's a draft, but it's probably going to 11 be fairly close to the final format. We will be 12 ready to implement by the end of the 90 day 13 comment period. We are quite confident on that. 14 And we feel that the mandatory 15 implementation date of 24 months after the 16 comment period is both reasonable and do-able. 17 And I'd like to thank you once again 18 for your time. I have provided my contact 19 information for you on the last sheet and I am 20 certainly willing to answer any and all 21 questions. 22 CHAIR HILLMAN: Thank you. 1 Commissioners, if we could just 2 agree that we will let Ms. Coggins make her 3 presentation and then open it up for questions to 4 both. Thank you. 5 Ms. Coggins. 6 MS. COGGINS: Thank you. Thank for 7 inviting me here today to provide some comments 8 on the proposed Voluntary Voting System 9 Guidelines. 10 First let me say that it is my 11 perception that in view of the HAVA January 2006 12 deadline, there is significant concern in the 13 states and the vendors in how the Human Factor 14 requirements stack up against the current testing 15 of accessibility to the 2002 standards. 16 I think Joe has done a really good 17 job of talking about impact in the long term. I 18 thought I may address my remarks a little to the 19 transition period of providing a little bit of 20 insight as to where the differences are for 21 people are going to have to be making some 22 decisions for January 2006. 1 First I just want to say this 2 release was made available late last week so 3 these comments are based on my preliminary review 4 of the guidelines and should not be considered 5 complete research. 6 Also, there has not been really a 7 peer review on the comments, which is our normal 8 process, in view of the time frame. These are my 9 initial interpretations of the guidelines and 10 shouldn't be considered a formal in-depth 11 analysis. We actually are going to make a 12 recommendation that maybe the EAC consider 13 providing this as a service to the states and the 14 vendors in the interim. 15 Third, in identifying those 16 differences, it does not mean that the systems 17 qualified to the 2002 standards do not contain 18 some of these required features. It means that 19 they weren't required to contain these features 20 when they were tested to the 2002 standards. 21 A vendor may have incorporated some 22 of these features into their systems. In that 1 case, the ITA would have been tested to them. 2 The first group of changes is audio. 3 And in the 2002 standards it did not identify 4 specific decibel volume settings for the initial 5 volume or minimum volume control. The VSS did 6 specify maximum, so things were tested to that. 7 In the 2002 VSS it did not identify 8 specific frequencies over the audible range. So 9 that would be something that would need to be 10 tested. 11 The 2002 VSS did not stipulate the 12 use of human recorded speech rather than 13 synthesized speech. There are 2002 qualified 14 systems that use synthetic speech. 15 And there appears to be a 16 contradiction in this requirement to the 17 underlying principle that you don't dictate a 18 design. 19 And it seems within the discussion 20 on this requirement, it is preferred not to use 21 synthesized speech. But that may be an issue as 22 to systems that are out there. 1 The 2002 VSS did not stipulate that 2 voters control the rate of speed. And now also 3 the voting system guidelines indicate this as a 4 'should' and not a 'shall'. So I'm not clear 5 whether or not the intention is that that will be 6 a requirement or if that's just -- 7 A 'should' is normally optional when 8 you are looking at a requirement. And there are 9 vendors out there who have addressed this issue. 10 The next group of changes deals with 11 controls. The 2002 VSS did not stipulate that 12 mechanically operated controls or keys on the 13 voting system shall be tactilely discernible 14 without activating those keys. I'm not sure that 15 that is a really significant issue as far as 16 current systems out there. It doesn't pop to 17 mind that that seems to be a problem. 18 2002 VSS does not stipulate that the 19 status call logging or toggle control keys, such 20 as shift keys, shall be visually discernible and 21 discernible either through touch or sound. That 22 was not stipulated. Nothing comes to mind as 1 being really an issue with that though. 2 The 2002 VSS did not stipulate that 3 controls shall not require direct bodily contact 4 or for the body to be part of any electrical 5 circuit. Additionally, it did not stipulate that 6 the voting system shall provide a mechanism to 7 enable non-manual input that is functionally 8 equivalent to tactile input. This would be along 9 the lines of CIVNTOP (ph). 10 And, also, this is stated as a 11 'should' and not a 'shall'. But there are 12 systems out there, 2002 qualified, that did 13 incorporate this type of technology. 14 And the same with the next 15 requirement, that the VSS stipulates voters who 16 lack fine motor control or use of their hands 17 perform submission of ballot. 18 Again, that's something that vendors 19 have addressed, some. 20 Also the guidelines, as Joe had 21 said, are changing the physical reach. They are 22 much are more detailed and the numbers appear to 1 have changed. So that's something that may need 2 to be addressed. 3 There are systems out there that 4 have really addressed some of these issues 5 through documentation. I have not done a 6 complete analysis of all the changes that went 7 into effect and the measurements. But in some 8 cases this may actually be -- It's an issue of 9 setup and it may actually be something that can 10 be handled from documentation levels. 11 CHAIR HILLMAN: You will need to 12 speak into the mike directly. You are turning 13 away from it and I think they can't hear you. 14 MS. COOGINS: I'm sorry. I am 15 usually so loud, they usually tell me to quiet 16 down. 17 Visual sight impairments is the next 18 group, the final group. The 2002 VSS did not 19 stipulate a system adjusted by the voter or poll 20 worker, including font size, color and contrast 21 shall automatically reset to the default after 22 each voter. 1 This was actually addressed in a 2 NASED bulletin. So it is my feeling that most 3 systems should have tested that. 4 The NASED bulletins were part of -- 5 The implementation program is this mechanism for 6 having additional bulletins for clarification, 7 were part of the -- 8 CHAIR HILLMAN: Tap it and see what 9 you are getting. 10 MS. COGGINS: Is that any better? 11 CHAIR HILLMAN: Yes. Just have it 12 as close to you as you can. 13 MS. COGGINS: Okay. 14 CHAIR HILLMAN: Okay. There you go. 15 Do you want to start over and repeat 16 everything -- 17 No, I'm only kidding. 18 MS. COGGINS: Do you really want to 19 hear it? 20 So in terms of the NASED Technical 21 Bulletin Number 1, that was a mechanism that was 22 part of the 2002 implementation program that 1 there would be additional bulletins for updates. 2 So that should have been tested. 3 The 2002 VSS did not stipulate that 4 all tests intended for the voter during the 5 voting session be presented in sans serif font. 6 It did state that there should be a clear font. 7 I don't know -- I am under the 8 impression that states do actually stipulate 9 certain fonts on their ballots. So is this 10 perhaps an issue with compliance with state laws 11 or are all states using this font. 12 If this is a common thing, I'm not 13 aware of that. 14 The 2002 VSS did not stipulate 15 minimum figure to ground ambient contrast ratios 16 for text to informational graphics. So that 17 would be something that would have to be tested. 18 And, lastly, the VSS did not 19 stipulate a minimum size for sensitive touch 20 areas. So that would be something that, although 21 I haven't researched the actual sizes, but I 22 don't think that that will wind up being greatly 1 significant to systems that were already 2 qualified. 3 And, as I said, I understand that 4 the guidelines are just out for public comment. 5 But in view of the phone calls that I have 6 received from states - asking me, oh, my God, I 7 have a system; I want to buy it. Am I going to 8 be in trouble if I buy this system because it's 9 2002 qualified? What does this mean? - I would 10 recommend that in order to keep the states and 11 the local jurisdictions informed and the vendors 12 and public, that the EAC at least initiate 13 publishing a formal gap analysis between the 2002 14 voting system standards and the proposed voting 15 system guidelines. 16 In doing so, I believe that will 17 help states ask vendors about specific questions 18 about their voting systems. Also, vendors with 19 2002 VSS qualified systems that support things 20 that are proposed in the guidelines could request 21 the lab to provide independent verification of 22 any supported gap-identified functionality and 1 that this could then go as addendums to their 2 reports. 3 Let's see. There are actually - 4 because of time - there are a number of items and 5 I think Joe went through a number of items. 6 In an overall comment -- 7 Am I off again? 8 CHAIR HILLMAN: Yes. 9 MS. COGGINS: I didn't touch it. 10 CHAIR HILLMAN: He's working on it. 11 MS. COGGINS: I think rather than 12 going through each individual issue that we 13 thought, there is kind of an overall comment that 14 I would like to make. 15 One of the criticisms of the 2002 16 standard was imprecise language. And I will be 17 submitting these comments directly through the 18 comment process. 19 But there are cases in here where we 20 are seeing imprecise language. In some instances 21 we are seeing - I'll give you an example - the 22 audio system should allow voters to control, 1 within reasonable limits, the rate of speech. 2 Reasonable limits is not testable. 3 What's reasonable to you/what's 4 reasonable to me, is a lawsuit. 5 So that's the one thing that I would 6 say. 7 We need those kind of things 8 tightened up, providing us with limits. 9 There is one instance in the Voter 10 Verified Paper Audit Trail, it appears in the 11 guidelines that either the paper or the 12 electronic ballot could be the ballot of record. 13 And in Section 152 of the VSS, which has not 14 change in the guidelines, it actually says that a 15 paper based voting system can't record, store or 16 tabulate selections. 17 So this seems to be a conflict. 18 And I may be misinterpreting what is 19 in the new and changed guidelines. But that just 20 seems to be something that maybe should be looked 21 at. 22 So in terms of just the imprecise 1 language, we would like to see that addressed. 2 Lastly, there is one item that we, 3 as the labs, in the NASED technical committee 4 meeting, had discussed: issuance of a 5 clarification bulletin that addressed some of the 6 issues encountered in interpretation with the 7 code review standards, and specifically the 8 application of the term in module in various 9 requirements. 10 But, also, the VSS gave us very 11 broad language on some security requirements and 12 code review. Well, just basically, security 13 requirements. And it allowed the labs on their 14 own to interpret some coding practices as 15 insecure. 16 The problem is, these practices are 17 not explicity identified and that leads to 18 inconsistent application. A good example of this 19 is the bar coded password. 20 The labs hope that more explicit 21 requirements for security review would be part of 22 this release. And that's something that I think 1 I would like to see in the security requirements 2 of best practices or here are the things in code 3 review that we want to make sure exist. 4 Again, as I said, the imprecise 5 language is really one of my concerns, is that 6 all requirements really have testable criteria. 7 Overall, the standards are good and 8 we look forward to testing to these and we will 9 submit specific comments on these through the 10 comment process. 11 CHAIR HILLMAN: Okay, thank you. 12 While the little problems that we 13 are having with the technology here don't disrupt 14 our meeting, you can imagine it wouldn't be funny 15 if we were encountering these little problems 16 with voting machines on Election Day, pointing 17 clearly to the need for good guidelines, testing 18 and certification and re-testing and 19 recertification as necessary. 20 I thank both of you for your 21 comments this morning. The responsibility for 22 accrediting test labs and the whole certification 1 process is one that the Federal Government, 2 through the Election Assistance Commission, is 3 embracing for the first time. It is a 4 significant and rather awesome responsibility. 5 And it absolutely is important for us to receive 6 your comments against the guidelines as you have 7 noted. 8 The purpose of this hearing, the 9 purpose of the entire 90 day process, is for us 10 to receive this kind of critical input and for us 11 to have an opportunity to explore some of the 12 comments that you have made and we welcome this 13 opportunity to do that. 14 Commissioners, I think we have 15 sufficient time. We have about, in total, 30 16 minutes. And so I guess if we divvy it up 10/10 17 and 10, that does include responses back. So we 18 beg your indulgence to confine your responses to 19 questions, if it appears we are bumping up 20 against the time limit. 21 Thank you. 22 Commissioner, Mr. Vice Chairman. 1 VICE CHAIR DeGREGORIO: Thank you, 2 Madam Chair. And thank you both for the 3 excellent presentation and the written comments 4 that you have provided us. 5 We certainly recognize that this is 6 a first cut for you since these guidelines were 7 just published in the Federal Register and we 8 recognize that over time you will probably add to 9 these comments. 10 But, initially, they are well done 11 and they certainly raise a lot of important 12 issues for us to look at over the next 90 days, 13 because you are the ones that will have to test 14 the equipment to these guidelines. And you have 15 certainly raised some issues here that are 16 important for us to consider. 17 One thing that I have heard from 18 election officials around the country about these 19 new guidelines is the time that may be required, 20 the additional time that may be required to test 21 this equipment against these new guidelines for 22 the equipment to be qualified or certified. 1 And do you see, because you have 2 been in this business for quite a while now, that 3 these guidelines will add additional time to the 4 process to qualify a voting system when it is 5 presented to us and when you have to test it 6 against these guidelines? 7 Mr. Hazeltine, you can certainly -- 8 MR. HAZELTINE: Mr. Vice Chair, yes, 9 I do. Off the top of my head, I am looking at 10 probably one to two weeks. There are additional 11 activities. They are not tremendous, but there 12 are additional activities. 13 VICE CHAIR DeGREGORIO: What is the 14 average time? You say one or two weeks 15 additional, so what -- 16 MR. HAZELTINE: Well, the average 17 time for a system, from start to finish, is 18 normally a three to four month time frame. So we 19 are not talking about an order of magnitude 20 change. 21 VICE CHAIR DeGREGORIO: What about a 22 cost increase to the client? As I understand it, 1 it's going to be the vendors that will pay you 2 all to test their equipment against these 3 guidelines. Is there going to be a significant 4 cost increase as a result of this -- 5 MR. HAZELTINE: I wouldn't say there 6 would be significant. Yes, the activities will 7 go up, so that the time we spend will go up and 8 corresponding new costs. 9 MS. COGGINS: I would say it will 10 add a small amount of time. It's not really 11 significant. Whether they pass or fail is, a lot 12 of times, whether it takes longer or not. 13 So if they come and they have to 14 actually retest, then it takes longer and it 15 costs more. 16 So it's hard to necessarily put a 17 number. You always talk about a number that is 18 successful. But a one week test may require a 19 two month re-engineering effort. So that would 20 be the situation. 21 And you have a very different, a new 22 vendor has a very different experience from a 1 vendor who is in the market and has deployed 2 voting systems and has been doing this a long 3 time. 4 So one person it may take just an 5 extra day for a test. It also depends on what 6 the systems themselves contain. 7 So if we have already tested for 8 some of these things, then there really wouldn't 9 be any additional testing because the VSS 10 requires that we test to their own requirements 11 in certain optional functionality or additional 12 functionality. 13 There I went again. 14 CHAIR HILLMAN: It will be fixed 15 during our break so that we don't have to go 16 through this. So our apologies to the two of 17 you. 18 MS. COGGINS: Not at all. 19 So that's basically it. There are a 20 lot of variables that are involved in it. It 21 depends on the vendor. 22 It will add some time; it will add 1 some expense. And I think also in terms of some 2 people, there will be significant times of 3 engineering. That I think would be more of the 4 time that I see, is engineering the products to 5 reach some of these requirements. 6 But in terms of testing, it probably 7 would be a 5 to 10 percent increase. 8 VICE CHAIR DeGREGORIO: Previous 9 standards have had an implementation period. 10 When the 1990 standards were developed, there was 11 a several year implementation period. And in 12 2002 there was, I believe, a two year 13 implementation period. 14 We have proposed in this draft a 24 15 month implementation period, which would take us 16 to the fall of 2007, before these guidelines 17 would take effect and have to be -- 18 Is that a reasonable approach in 19 your view, Mr. Hazeltine? 20 MR. HAZELTINE: Mr. Vice Chair, as I 21 said earlier, I think that is both reasonable and 22 do-able. 1 There are some things which I think 2 need to be on the fast track, which would be the 3 Sections 2.2.7, the accessibility requirements. 4 MS. COGGIN: The market will not 5 wait for them to be put into place. The market 6 will dictate to us that we have to move faster on 7 these. 8 So while we may have - we will have 9 vendors who will be competing to try and get 10 there before others. So at the lab, we will be 11 trying to get there as quickly as possible. 12 So two years is, it's a good cutoff. 13 Because as we found in 2002, there were 14 situations where people weren't ready to get 15 there. But I don't know that people will 16 actually -- 17 Somebody will be ready to go three 18 months from now, in my interpretation. 19 VICE CHAIR DeGREGORIO: If a state, 20 for instance, chooses to adopt a VVPAT portion of 21 this and have its implementation effective for 22 next year, for next year's election, the 1 equipment they are going to use, because their 2 state law or state regulation requires them to 3 use the VVPAT if they use DRE equipment, so you 4 would be prepared then to test this equipment 5 against the VVPAT section of these guidelines, 6 again, if the states chose to adopt that and have 7 a different implementation date? 8 MS. COGGINS: It depends upon -- 9 In terms of our dealing with the 10 vendor, it's really, we would be working on what 11 their requirements would be. 12 If they are designing their system 13 to a particular state's requirements, then we 14 would need to design test cases for that. 15 The good news would be that, 16 ultimately, we probably are going to be able to 17 re-use that with other vendors. But, initially, 18 getting out there -- 19 You actually are to look -- There's 20 very much custom work that's done in every ITA 21 qualification because no two systems are truly 22 the same. So while you can work off a really 1 high level set of requirements, when it comes 2 into an implementation - and, in fact, we have 3 this discussion sitting in a lot of cases where 4 we are sitting in a room and we are testing 5 something on one voting system and we go, oh, my 6 gosh, this is so much like such and such. Did we 7 test them for that? Oh, no, but they don't do 8 this piece, they don't do write-in's the same. 9 Whew, okay, we are covered. Yes, we tested it to 10 the correct standard in that particular case. 11 So every system actually has a great 12 deal of customization on the test cases. 13 So I think it's actually our normal 14 process that we really have to be responsive to 15 the system. And while the good news will be that 16 we will be able to repeat processes more 17 consistently across various vendors, at this time 18 we actually have to respond to whatever a vendor 19 design is, because that's what the VSS says that 20 we do. 21 MR. HAZELTINE: Relative to the 22 Voter Verified Paper Audit Trail, I see no real 1 issues with it. 2 The major thing that we need - I 3 have already talked to Mr. Wilkey about this - is 4 the compliance matrix. And we have begun 5 discussing having a meeting later this month to 6 do that. 7 Once that's in place, we are ready 8 to go. 9 VICE CHAIR DeGREGORIO: Thank you, 10 Madam Chair. 11 CHAIR HILLMAN: Okay. Thank you. 12 For your entertainment pleasure, we 13 are raising and lowering the blinds, as you can 14 see. Kind of looks like we are now in a 15 submarine and submerging at any moment now. 16 Seriously, for the people on this 17 side of the room, even though the sun is 18 filtered, it was getting kind of warm and then 19 little technical gliches with plants and other 20 things being in the way. 21 So, hopefully, you were entertained 22 for a moment. But we should be squared away now. 1 Thank you. 2 Commissioner Martinez. 3 COMMISSIONER MARTINEZ: Thank you, 4 Madam Chair. 5 My thanks to both of you as well for 6 making the time and the effort to be here. And 7 as the Vice Chair has noted, we were trying to 8 get this out as soon as we could. And, 9 unfortunately, it took us a little bit longer to 10 do our internal due intelligence. So it was just 11 out last week. 12 And we appreciate that you have done 13 a rather hasty, but I think a very excellent, 14 analysis for us to begin to think about with 15 regard to -- 16 There goes my mike. Am I back on? 17 Everybody hear me? 18 -- with regard to the Proposed 19 Voluntary Voting System Guidelines. 20 I want to start with a general 21 question for the both of you. 22 And that is, one of the things that 1 we have to contemplate right now at the EAC, 2 along with our partners at NIST and of course 3 with the TGDC, is the idea of what happens after 4 this initial, after this first draft is adopted 5 in full or in final form. 6 And that is to say that there has 7 been, we have deliberated internally and talked 8 amongst ourselves with NIST and the TGDC about 9 whether this is a Version I versus a Version II 10 coming out at some point in the near future. 11 And I wondered, from your 12 perspective, since you actually have to do the 13 use of the requirements and the actual testing of 14 the voting systems, what life would mean for the 15 testing labs if we were to embrace a chronology 16 that says we go final with VVSG in 90 days with 17 this version and then soon thereafter, let's say 18 six months after that one goes final, we come out 19 with a Version II, another set of, again, revised 20 Voluntary Voting System Guidelines. 21 Is that problematic from your 22 perspective? What opinion would you like to 1 offer us so that we can be better informed in 2 trying to make a decision about the timing of 3 what comes after this, if anything? 4 Mr. Hazeltine. 5 MR. HAZELTINE: Thank you, 6 Commissioner Martinez. 7 From a testing laboratory 8 standpoint, we can be ready in 90 days as to the 9 standards, and if you decide to revise it 90 days 10 later, we can probably be ready for that one as 11 well. From a reality standpoint, test, the 12 vendors will not be -- 13 CHAIR HILLMAN: Your mike. 14 MR. HAZELTINE: Once again, we can 15 be ready from a testing laboratory standpoint. 16 But I do think the vendors would have some 17 problems with that. It would be a continuing 18 ratchet of requirements that the systems need to 19 have. 20 I frankly think that we need to 21 reach some point of stability that people can 22 design to and then if we want to move from 1 forward from that. 2 COMMISSIONER MARTINEZ: Ms. Coggins. 3 MS. COGGINS: I would agree also in 4 terms of, in all quality systems, management of 5 change is one of the important factors. So as a 6 test lab, that's one of our mandates, that we 7 need to be able to do that. 8 I also agree with Joe in that I 9 don't think it's just the vendors. I think it's 10 also the states. 11 There is paralysis if you have 12 got -- Maybe it's better that it comes in six 13 months or maybe it's better that you defer the 14 first one six months. 15 But the basic issue is, people don't 16 know what to do. They don't know whether to buy; 17 they don't know what's going to happen; are they 18 going to lose their HAVA funds? 19 That's information that I am getting 20 from the states where they are calling me and 21 asking me, you know, what does this mean. What 22 does it mean these changes of standards? 1 COMMISSIONER MARTINEZ: Is there a 2 process right now with the current 2002 VSS and 3 the NASED process where if they want to add 4 something to the existing standards, it is done 5 via an amendment of some kind to the 2002 VSS? 6 Or is there a process, if they 7 identify something that needs to be addressed - I 8 think you mentioned something about a bulletin 9 that might be issued by NASED or something like 10 that - so there is a process right now. 11 Can you say something a little bit 12 about that real quick, Mr. Hazeltine. 13 MR. HAZELTINE: Commissioner 14 Martinez, yes. NASED over the years has issued a 15 number of technical guidances that kind of 16 attaches, kind of like an appendix, if you will, 17 to the standard and provide guidance to us and 18 also to the vendors on what the requirements of 19 the system are. 20 That is a kind of a stop gap way of 21 adding things in small pieces. 22 MS. COGGINS: The problem in the 1 process at this point is you have this transition 2 between NASED and the EAC, so we are in 3 paralysis, an example being this bulletin that we 4 talked about about a year ago, can we get this 5 out. 6 And at this point NASED is 7 anticipating giving this over to the EAC. So, 8 yes, the method is there, but the will is not 9 there, in my belief. I'm not trying to speak for 10 anybody else or anything here. 11 But there is just, people are in 12 this flux and they are not quite sure. If I put 13 out a bulletin today, what's the impact? 14 COMMISSIONER MARTINEZ: Right. I 15 guess what I am trying to get at, if I am trying 16 to make a decision about what happens next, is 17 the process working the way it exists now, in 18 that if there is a deficiency that's identified, 19 that you don't have to go through a whole 20 promulgation process again of redoing or revising 21 the entire VSS, but instead you would address 22 those particular deficiencies via a bulletin or 1 via some sort of a more responsive mechanism. 2 And it sounds like that is working right now. 3 So as we contemplate how to 4 structure ourselves - since we are, by law, 5 supposed to take over this process - that is 6 something that we should be informed about, 7 basically. 8 MS. COGGINS: The system is there. 9 It's just not currently working -- It's not, 10 nothing is going on, it really is. But, yes, 11 there is a mechanism. 12 COMMISSIONER MARTINEZ: And one that 13 works, Mr. Hazeltine? 14 MR. HAZELTINE: Yes, I believe it 15 does. 16 COMMISSIONER MARTINEZ: In your 17 testimony, Mr. Hazeltine, you have referenced the 18 fact that in some areas - for example, in the 19 BBSG - the section dealing with wireless 20 technology, that you would need to acquire 21 subject matter expertise. Is that available? Is 22 it even out there? 1 MR. HAZELTINE: Commissioner 2 Martinez, I am not sure and we have not 3 researched that. 4 I am sure there are experts in blue 5 tooth and 80211 or BT (ph) whatever. 6 COMMISSIONER MARTINEZ: So there is 7 personnel that you can get. 8 MR. HAZELTINE: There is. And we 9 have a good bit of expertise in the same thing as 10 well. 11 Frankly, wireless, I'm not sure if 12 the technology is ready for implementation, 13 primarily from a denial service viewpoint. 14 COMMISSIONER MARTINEZ: Do you 15 agree? Are there some areas that it's going to 16 be necessary to acquire some additional 17 expertise? 18 MS. COGGINS: Yes, I think that one 19 of our comments about imprecise language is that, 20 it's saying you are becoming a subject matter 21 expert but it's not identifying the 22 qualifications of the subject matter expert. 1 That may be something that is going 2 to be addressed in the accreditation criteria. 3 But we would like to see in the voting guide that 4 it actually references a subject matter expert 5 who is accredited by the lab or whatever is the 6 specific thing. 7 But, yes, that is some of the 8 imprecise language that we would like to see 9 clarified. 10 COMMISSIONER MARTINEZ: I guess 11 going back to something that my colleague, the 12 Vice Chair, was talking about earlier, one of the 13 things that we had certainly emphasized - and I 14 know that the members of the TGDC were in 15 agreement with this - and that is to try to 16 deliver - with the VVSG - to try to deliver a 17 product that could be responsive to the upcoming 18 Federal election cycle next year, if at all 19 possible. 20 And, in particular, what we were 21 dealing with was the fact that - I think at last 22 count - something like 16, maybe 17 states, now 1 have madated, via legislative action or 2 administrative rule, the use of VVPAT technology. 3 And for those states that have moved 4 in that direction - and for perhaps others that 5 will move in that direction between now and the 6 next election cycle - that there would be a means 7 to be able to test -- 8 For example, if I use a DRE in a 9 jurisdiction right now, but the vendor in the 10 area has developed a VVPAT technology that could 11 simply be, essentially, plugged into that 12 existing DRE, that the plug-in, the added device, 13 the VVPAT component, can be tested in an 14 objective repeatable manner, that would happen, 15 obviously, through your test labs. 16 And what I you hear you all saying 17 is that you can be ready to do that once we go 18 final with the VVPAT section of this VVSG. Is 19 that correct, Mr. Hazeltine? 20 MR. HAZELTINE: Commissioner 21 Martinez, that is correct. Of the requirements 22 in the document, that was probably some of the 1 most straightforward and easiest to implement. 2 COMMISSIONER MARTINEZ: Ms. Coggins, 3 do you agree? 4 MS. COGGINS: Yes. In fact, we are 5 already testing that because vendors have already 6 addressed that. 7 COMMISSIONER MARTINEZ: So, 8 irrespective of whether a state decides to pull 9 down that section of the VVSG early, a vendor 10 might come to you after we go final with that and 11 say we want to be among the first to be tested 12 against the new VVPAT requirements, you all are 13 going to be ready to do is so is what I am 14 hearing you say. 15 MR. HAZELTINE: Yes, sir. 16 MS. COGGINS FOGEL: In fact, I think 17 we have been contacted by a vendor already who is 18 anticipating this and has asked us if they can 19 start testing sometime this month. 20 COMMISSIONER MARTINEZ: I will take 21 this opportunity, I think we will talk more about 22 this this afternoon, that we go out of our way in 1 this document, in the VVSGV, to add language - I 2 think it's in the appendix right now, in Appendix 3 D - but to say to the general public and to the 4 reader of the VVSG that VVPAT is but one of many 5 ways to achieve what is termed independent 6 verification by NIST and by the TDGC. That other 7 ways to achieve independent verification - aside 8 from using VVPAT technology - is out there. 9 And at some point perhaps - or 10 perhaps it's usable now - but that there are 11 other means to achieve this. 12 And at some point down the road - 13 perhaps through an amendment process to the VVSG 14 when they go final, perhaps in the next iteration 15 of the VVSG - there would be other testable 16 requirements for other means of achieving 17 independent verification. 18 Are you following what I am saying? 19 MR. HAZELTINE: Yes. If I were to 20 understand what you are saying, I don't think you 21 want to do anything which would stifle 22 technology. 1 COMMISSIONER MARTINEZ: Right. 2 MR. HAZELTINE: Let the vendors be 3 creative and come up with other ways of meeting 4 the requirements. 5 The requirements, I think, are well 6 stated in the document. 7 COMMISSIONER MARTINEZ: Right. But 8 what you said in your testimony is getting to 9 other means, testing the requirements for other 10 means of independent verification will require 11 additional major effort on your part. 12 MR. HAZELTINE: As stated in 13 Appendix D, the dual system where you have got 14 the one system and you go back and you repeat the 15 count on the second system, that is two pieces of 16 hardware; that's twice as much testing. 17 There are other ways to do that. 18 COMMISSIONER MARTINEZ: Right. 19 Any comments, Ms. Coggins? 20 MS. COGGINS: No. In terms of 21 testing, it's going to be nothing really 22 different from what we are currently doing now. 1 COMMISSIONER MARTINEZ: And I am 2 getting a bad stare from the Chair. 3 CHAIR HILLMAN: Excuse me, not a bad 4 stare. I am just letting you know we're a little 5 over time. 6 COMMISSIONER MARTINEZ: I will do 7 one quick, perhaps more of a statement, and you 8 all can do a quick comment, if you want. 9 The gap analysis that you have 10 suggested for the following reason: the current 11 set of voting system standards, the 2002 VSS, was 12 promulgated by our predecessor agency, the 13 Federal Election Commission. It was promulgated 14 over a two or three year period and it went final 15 sometime - I don't remember the exact month in 16 2002 but perhaps one of you can tell me. 17 MR. HAZELTINE: Effective January 1, 18 2004. At that point everything had to comply. 19 COMMISSIONER MARTINEZ: Right. But 20 the FEC Commissioners voted -- 21 MS. COGGINS: It was actually 22 published in April 2002. 1 COMMISSIONER MARTINEZ: That's the 2 date I was looking for. 3 MS. COGGINS: It was released in a 4 meeting in May in Colorado. 5 COMMISSIONER MARTINEZ: The point 6 that I would make, then, to go to your suggestion 7 on the gap analysis, is that, HAVA came along 8 after the 2002 voting system standards went 9 final. 10 So HAVA, like it or not, HAVA has 11 some voting system standards itself in Section 12 301, which are nothing at all voluntary, in fact, 13 they are mandatory on 1/1/06 - or, actually on 14 1/1/06 - upon all jurisdictions covered by HAVA. 15 So the gap that you have, 16 essentially, is that you have current voting 17 system standards that were promulgated before 18 HAVA was even close to its final form. 19 So if you are a vendor, you get 20 tested to the 2002 standards, you still don't 21 have anybody to tell you - other than perhaps 22 your own internal counsel and your own 1 deliberation and your own specifications that you 2 would impose upon yourself as a vendor - whether 3 your system is compliant with what Congress came 4 out with in the Help America Vote Act, in Section 5 301, particulary 301(a)(3), which is the 6 accessiblility language. 7 Would you all agree with that? 8 Mr. Hazeltine, go ahead, or whoever. 9 MR. HAZELTINE: Mr. Martinez, I 10 would agree with that. 11 I believe the common sense way of 12 looking at this, there are voting systems that 13 are out there now which are complying with the 14 majority of this document other than the HAVA 15 requirements. 16 HAVA work station, the voting 17 stations, will probably end up being an 18 independent voting station at the polling place. 19 So it is actually almost a separate piece of 20 hardware. 21 COMMISSIONER MARTINEZ: Ms. Coggins, 22 you were going to say something. 1 MS. COGGINS: I would say that it's 2 not just that -- You have partial compliance at 3 this point, based upon the vendors. You are 4 correct, yes, you had a gap between the time when 5 the standards were there and when HAVA was 6 imposed. 7 And, truly, that gap actually is 8 until today, because this is really the first 9 time those standard are there. So it has been, I 10 don't know what, is it three years or something? 11 COMMISSIONER MARTINEZ: Right. 12 MS. COGGINS: So that's truly what 13 the gap is. 14 COMMISSIONER MARTINEZ: Right. 15 MS. COGGINS: So, yes. And you have 16 had people who have tried to anticipate this and 17 have tried to get there. So in terms of what 18 they are looking - the information - I would 19 absolutely suggest that this information be put 20 out for the benefit of everyone. 21 COMMISSIONER MARTINEZ: Thank you. 22 Thank you, Madam Chair. 1 CHAIR HILLMAN: Okay. 2 Our Executive Director has a 3 question or two that he wants to ask, so I am 4 going to share my time with him. 5 But I am going to exercise my 6 prerogative to go first. So if we run out of 7 time, you don't get to ask your question. 8 I'm only kidding. 9 What does it mean to you that the 10 language has been changed and we are now talking 11 about guidelines instead of standards? What, if 12 anything, in the field does that mean? 13 MR. HAZELTINE: Madam Chair, I am 14 aware of a number of documents. Sometimes they 15 are called guidelines; sometimes they are called 16 standards. 17 They are called -- That's the 18 requirements requirement. So that really the 19 name is no major consequence. 20 MS. COGGINS: I agree. We are being 21 asked to test requirements and the vendor is 22 actually, when they come to us to initiate a 1 contract, they are specifying test us to this. 2 And so that's what we are testing now, that's our 3 interpretation. 4 CHAIR HILLMAN: So the quality has 5 no effect on it? 6 MS. COGGINS: No. 7 CHAIR HILLMAN: I was momentarily 8 interrupted, I believe, when you were responding 9 to the Vice Chairman's question about the length, 10 the anticipated length of time it would take to 11 test under the guidelines. 12 Could you just tell me, ballpark, 13 what that was. 14 MR. HAZELTINE: Madam Chair, my 15 response, right now, our initial review, is maybe 16 one to two weeks. It is not a tremendous -- 17 CHAIR HILLMAN: In total? 18 MR. HAZELTINE: One to two weeks on 19 a program which probably takes normally about 20 three months. 21 CHAIR HILLMAN: So we are talking 22 three to four months. 1 MR. HAZELTINE: It would probably 2 end up adding a half a month to it. 3 MS. COGGINS: One of the things that 4 is a little difference between the last, that, 5 and this initiative, still kind of up in the air 6 where we are going, SysTestLabs is both a 7 hardware and software test lab. 8 So our test cycle may be longer 9 because we have a bigger scope. 10 Currently, Wyle is a hardware ITA 11 and SysTest is a software ITA, the initial 12 software ITA. 13 So that may be -- There is a 14 different scope of testing that goes on. 15 For us the test effort is - it can 16 be two months to - we have had people go 18 17 months. So it just depends on the quality of the 18 system that's brought to mark. 19 THE CHAIR: Did you want to correct 20 something, Mr. Hazeltine? 21 MR. HAZELTINE: Yes, ma'am. 22 We would certify for both. 1 CHAIR HILLMAN: I think I heard you 2 correctly say that there would be an increase in 3 the cost of testing, but it wouldn't be a huge 4 increase. 5 All things are relative. So, is it 6 5%, 10%? 7 I know we are ballparking here. But 8 I'm trying to get an appreciation as to whether 9 the cost of testing is going to increase 25%, 10 50%, 10%? 11 MR. HAZELTINE: Madam Chair, it's 12 not in those ranges. 13 Just using this straight map, you 14 say it's a three month program - and to your 12 15 weeks, you are adding two. 16 So 8 to 10%, somewhere in that 17 range, doing it as an engineer. 18 CHAIR HILLMAN: Even with some of 19 the other things that you said you would have to 20 bring into the testing process because of the new 21 requirements? 22 MR. HAZELTINE: There are new 1 requirements; they require new tests, primarily 2 the functionality of things. 3 Once you have gone through it the 4 first time, you have all the fixtures and the 5 knowledge. So it becomes somewhat repetitive. 6 But the first time through it will 7 always be a bit of a challenge. 8 The accreditation is an additional 9 cost for us. It's not necessarily something 10 which the vendors would be picking up. 11 CHAIR HILLMAN: Ms. Coggins. 12 MS. COGGINS: Would you repeat the 13 question. 14 CHAIR HILLMAN: It was just about 15 the increase in cost to test against the new 16 guidelines. 17 MS. COGGINS: I would not see it as 18 being a significant cost, relative to the overall 19 cost of an effort. 20 It is, again, it's really dependent 21 upon the system that is brought us, independent 22 conditions, as to the length of the time and the 1 cost. 2 CHAIR HILLMAN: My final question is 3 for you, Ms. Coggins, 4 You said in the beginning of your 5 presentation, you were talking about the quick 6 time that you used - and we appreciate it - to 7 review the guidelines and that your normal peer 8 review process could not be engaged. 9 And then you talked about the gap 10 analysis. 11 You said something about providing 12 this as a service to the states. So my question, 13 if you can remember, is, what is the 'this'? 14 MS. COGGINS: The gap analysis. 15 It's basically providing information 16 so that the people understand, here are the 17 likely things you need to be looking for. Ask 18 your vendor, do they have an alternative method, 19 a non-tactile method for entering the voting 20 system. 21 If they have got that, check off the 22 box; we've got one. And is it in their report. 1 That's basically it. 2 The other thing, too, is, states may 3 have the opportunity to request additional 4 documentation on specific issues, if they are not 5 seeing these clearly identified in the 6 qualification reports that have been issued. 7 So that might be something so 8 that -- 9 I know they are all concerned. 10 CHAIR HILLMAN: Thank you. 11 Mr. Wilkey. 12 THE EXECUTIVE DIRECTOR: I will 13 stick to the time limit. I don't want to lose my 14 job before I get my first paycheck. 15 First, I want to make just a quick 16 comment to both of you - and I wish that the 17 other ITA was here also. 18 I have had the distinct honor and 19 pleasure of working with you in a prior life over 20 the past number of years and you have always 21 shown a very high level integrity, of 22 professionalism in everything you have done, as 1 well as all of the members of the NASED Board, 2 who have given hundreds of hours of time or 3 talent, without remuneration, without any kind of 4 assistance whatsoever. 5 I just have one question, and 6 primarily for you, Carolyn. 7 You made some general comments about 8 it needs to do this or it needs to this 9 throughout your document. Is it your intent, 10 when you go back, to give us some suggested 11 language that you would like to see? 12 And, Mr. Hazeltine, you could also 13 answer this. 14 CHAIR HILLMAN: Just a second. 15 Could you please repeat your question, Tom, 16 because the signers are having a hard time 17 hearing you. 18 THE EXECUTIVE DIRECTOR: I think you 19 made a number of general comments that it needs 20 to do this or it needs to do that, without giving 21 some specific language that you think would make 22 it better. 1 I ask you this because I don't think 2 there is anyone who has seen this equipment up 3 close and personal as you have. 4 So I think if you could give us some 5 specific language that you would like to see, it 6 would be very helpful. 7 Is it your intent to do that? 8 MS. COGGINS: I think in terms of, 9 in trying to say that something needs to be more 10 detailed, okay, it may be that the language we 11 can provide is, we need a limit. I can't tell 12 you the limit. I can't say - because that's 13 defined in the voting systems. It doesn't allow 14 me to do that. 15 So we can't actually set the 16 standard, per se, but we can indicate, yes, this 17 language, I need a little more detail on what is 18 the limit here. Or, in some cases, like we have 19 seen, there is issues where it's talking about 20 unacceptable font. It would be helpful if you 21 provided, okay, give us four - I'm sorry, not 22 font, format - give us four acceptable formats 1 and then say you can also do it in publicly 2 acceptable ones. Honing things down a little bit 3 makes it easier not only on us, but makes it 4 easier on the entire community. 5 Now we have two standard methods of 6 file transfers. 7 MR. HAZELTINE: I have no problem 8 providing additional guidance. I think we can 9 give you some additional guidance to work with. 10 CHAIR HILLMAN: Thank you. Thank 11 you all very much. 12 And we will break now for lunch. 13 I will ask if there are any 14 announcements before we do this. 15 MS. THOMPSON: Madam Chair, you just 16 may want to make the announcement again about the 17 restrooms. 18 CHAIR HILLMAN: Just a reminder that 19 the public restroom facilities are on the 8th 20 floor. They are not on this floor. 21 And we will begin again at 1:30. 22 Thank you very much. 1 AFTERNOON SESSION 2 3 PRESENTATION FROM VENDORS 4 REGARDING THE PROPOSED VVSG 5 6 CHAIR HILLMAN: The United States 7 Election Assistance Commission's hearing on the 8 proposed Voluntary Voting System Guidelines will 9 resume. 10 Panel Number 2, presentations from 11 vendors, those companies that manufacture the 12 equipment hardware and software that compose 13 voting systems. 14 We have with us today Mr. John Groh, 15 President of Election Systems and Software 16 International, also known as ES&S; Mr. Jim Adler, 17 CEO Vote Here; Mr. Alfie Charles, Vice President 18 of business development, Sequoia Voting Systems; 19 and Mr. Dennis Vadura, CEO of AccuPoll. 20 We will engage with this panel until 21 about 2:45 and then we will take a short break so 22 that we can set up for Panel Number 3. 1 Gentlemen, thank you for being with 2 us. And without any further ado, I will go in 3 the order that I introduced you. So if we could 4 start with Mr. Groh. 5 I think what we will do is receive 6 your presentations and then we will ask questions 7 following the fourth presentation. Okay. 8 MR. GROH: Thank you, Chairman 9 Hillman. 10 Again, for the record, my name is 11 John Groh. I also serve as Senior Vice President 12 of Election Systems and Software, Inc., which is 13 our American side company. 14 Today I brought along with me, 15 though, a peer of mine, Ken Carbolito, who is 16 Senior Vice President of product development and 17 software development for our company. 18 We want to thank you for inviting us 19 to be here today to present ES&S's views 20 regarding the proposed Voluntary Voting System 21 Guidelines, Version I. 22 I want to acknowledge first the 1 leadership and the hard work that the EAC and 2 your staff has accomplished. It is remarkable to 3 many of us that are in this business that you 4 have been able to make as much progress and so 5 soon. 6 Now let me get into a little bit of 7 the topic today about Version I. But indulge me 8 a little bit to allow me to share a little bit of 9 history on our company, because I think it sets 10 the stage of why we would be someone that could 11 speak to this. 12 For over 35 years, ES&S has been at 13 the forefront of driving election systems 14 innovation and working with election officials to 15 enhance the voting experience for all. 16 We are the largest and most 17 experienced provider of election systems and 18 services, supporting elections across the country 19 and throughout the world. 20 We are very proud of the fact that 21 twice now we have been the first major election 22 systems vendor to certify our entire product line 1 against the Federal Voluntary Voting System 2 Standards, first in 1990 and again in 2002. 3 In our opinion, the adherence to 4 standards and the rigor of the certification 5 process is critical to maintaining the integrity 6 of our elections. Therefore, we will tell you, 7 we embrace this process wholeheartedly. 8 I offer this background because it 9 underscores the fact that ES&S is not new to the 10 standards or the certification process. We 11 understand a dynamic standards process is key to 12 motivating innovation and continued enhancement 13 of the voting technology. 14 That is why ES&S has been 15 enthusiastically providing input during the TGDC 16 and NIST portion in written form and by attending 17 and participating in all meetings throughout the 18 current guidelines development. 19 Let me give you some general 20 observations. Through this lens of our 21 experience, we are pleased to offer some initial 22 thoughts regarding Version I. Because the draft 1 guidelines are extensive and were just issued 2 this week, it will take some time to complete a 3 thorough review. 4 Much like the former presenters 5 today, we too will have a peer review of this and 6 will provide very detailed written commentary on 7 this. 8 In reviewing the draft guidelines, 9 there is nothing more important than giving the 10 process the time and the deliberation it deserves 11 to get it right. And I want to say this again, 12 because we think this is the most important 13 component of this, giving all of us enough time 14 to get this right and deliberate on it correctly. 15 One criteria for us has been 16 paramount, as we look at the process we have been 17 engaged in, and that's feasibility, and that's 18 feasibility for many areas. 19 Technical feasibility because the 20 guidelines must be implemented and have to have 21 any real effect and also in the guidelines, the 22 feasibility of achieving the EAC deadlines of 1 implementation and covering the additional costs 2 associated with substantial upgrades to product. 3 The EAC and NIST and the TGDC 4 rightly determined that an interim step in the 5 guideline development process may be necessary, 6 with the very specific and limited goal of only 7 addressing significant limitations in the 2002 8 standards. 9 All along ES&S understood that the 10 objectives of this interim process was a 11 complement to the 2002 standards, not really 12 meant to replace them. 13 So our first observation about the 14 proposed guideline, particularly given the 15 intention to pursue a second version soon, is a 16 belief that, in sum and content, Version I may 17 have gone beyond the original intent and scope as 18 defined by the EAC. 19 As proposed, the Version I would 20 impose a substantially new expectation on vendors 21 and election administrators alike. That raises 22 some challenging technical issues I will try and 1 address. 2 But it also means that the 3 implementation will take significantly more time, 4 add costs and complexity. 5 On this point we appreciate the 6 Version I effective date provision, clarifying 7 that the guidelines will go into effect 24 months 8 after their final adoption by the EAC. 9 Certainly, given all that must be 10 done to develop the new product and testing 11 authorities, we believe it will take at least the 12 afforded time to accomplish this task as a 13 supplier or a vendor. In fact, when you consider 14 that it took five full years to develop the 15 implementation of the 1990 standards and then 16 three years to do the same in 2002, this proposed 17 timeline for Version I is aggressive. 18 However, and again recognizing how 19 firmly we support an evolving standards process, 20 and how committed we are to continued voting 21 system enhancement, know that we will move 22 forward aggressively in that direction. 1 The effective dates provisions are 2 important because they begin to send messages to 3 states and localities that they can move forward 4 with purchasing new systems to meet the HAVA 5 requirements, especially in Section 301, before 6 these new guidelines are in full force. 7 In other words, what we are 8 recommending is the need for a strong 9 reaffirmation that voting systems certified to 10 the 2002 standards achieve the objective laid out 11 in HAVA, providing a substantial improvement over 12 older voting technologies. 13 Every election official who is 14 working to comply with the January 1, 2006 15 deadline must know that they can purchase 16 certified 2002 systems with the confidence that 17 these systems will fulfill the mission and 18 mandate of HAVA. 19 In addition, they should know that 20 they will be able to continue to use these 21 systems for years to come, that they will not be 22 required to replace HAVA-required technology with 1 new products that meet the proposed Version I 2 until it is necessary. 3 If, in fact, this could be a 4 requirement in the future, they also should know 5 that funding may be provided to cover the cost of 6 a second upgrade of equipment. 7 Without sending a strong message to 8 state and local election administrators along 9 these lines, we believe that some may opt for 10 missing the HAVA deadlines to wait for 11 implementation of Version I, and we think that 12 would be a mistake. 13 Now to the technical issues. The 14 technical context of the Version I, here again 15 with an eye towards feasibility, first, just to 16 reiterate - if the proposed Version I is 17 implemented as it is drafted today, you should 18 know that compliance will or may require 19 comprehensive product development. 20 For those of us who embrace 21 continually evolving standards and who work hard 22 to meet the challenges of updated guidelines, 1 this is a substantial undertaking. At ES&S we 2 have a passion for this innovation so we don't 3 want to say compliance is impossible. But, 4 without question, it will take time and add 5 significant cost and complexity. 6 Take, for example, the setup and 7 validation requirements of Section 6.4, calling 8 for hardware and software verification systems 9 provided through third-party vendors which must 10 perform their function without utilizing the 11 voting system software. This will require a 12 whole new approach to voting system hardware. 13 In addition, to comply, ES&S would 14 have to sacrifice a key security feature of our 15 current voting systems product, the inability to 16 interface with any outside components. 17 We intentionally developed our 18 products without any ports or other connection 19 points, to eliminate the potential for tampering 20 via connected sources. 21 Now under Section 6.4, the only way 22 we can comply would be to incorporate the very 1 connection port we have excluded from our design. 2 Similarly, the voter verifiable 3 paper audit trailer, or VVPAT provision, raises 4 serious concern. 5 To be clear, ES&S applauds EAC's 6 attempt in the proposed Version I guidelines to 7 provide some guidance on this point. 8 Our position on VVPAT has always 9 been that while we have absolute confidence in 10 the reliability, accuracy and the security of our 11 voting systems, we know that many election 12 officials and the public, for that matter, want 13 the added benefit that VVPAT brings. 14 We also understand that this is an 15 important consideration for raising the public's 16 overall confidence in the voting process. 17 For vendors and election 18 administrators who have already started to 19 address the issues of VVPAT, the proposed Version 20 I is overly prescriptive and could very well 21 erode the progress many of us have already made. 22 Specifically we are concerned about 1 the requirement defined in Section 2.2 as it 2 relates to 6.8, which relates to maintaining 3 voter anonymity and providing accessibility for 4 the visually impaired. 5 We support both of those goals. 6 However, to fulfill the specific Version I 7 requirements, including that voter receipts must 8 be shuffled or reorganized in some way, we would 9 have to go back to the drawing board on our VVPAT 10 solution. In fact, we think most, if not all, 11 vendors would be in the same predicament, because 12 many of us - of the available systems today - 13 have used a paper roll. 14 In addition, there is an issue of 15 general inconsistencies between provisions of the 16 proposed guidelines. In considering the Version 17 I as a whole, we have found several instances 18 that we will comment on where the requirements of 19 one section may conflict with the requirements of 20 another. 21 Consider, for example, the following 22 two requirements. 1 In Section 2.2.2.2, it requires that 2 any audio-tactile interface shall allow the voter 3 to have the information provided by the system 4 repeated. But then in Section 5.4.2, it requires 5 that no key or control on a voting station shall 6 have a repeat feature enable. We will need 7 guidance on this. 8 In our more detailed written 9 responses we will provide some additional 10 examples of this same occurrence. Therefore, 11 before finalizing the Version I, some 12 clarifications and revisions in several instances 13 may be necessary. 14 Going to recommendations, taking 15 into consideration feasibility, time 16 requirements, and added complexity associated 17 with implementing Version I, we have three 18 recommendations for the EAC. 19 One, allow election officials to 20 confidently move forward with the current 21 standard; clarify through a safe harbor statement 22 that states and localities that, one, in need to 1 make purchasing decisions today to comply with 2 HAVA, can do so under today's standards and that 3 systems purchased today will not have to be 4 replaced in the future or when the next version 5 comes out. 6 My second recommendation - I would 7 like to indulge a little bit of the EAC's ability 8 for me to amend my proposal or recommendation in 9 a follow-up further recommendation - but we have 10 stated in two, that because of the time and added 11 complexity and cost of implementing Version I, 12 you may wish to consider to combine Version I 13 with upcoming Version II. 14 This would eliminate confusion in 15 the marketplace and allow all of us the time we 16 need to effectively develop and implement very 17 clear and understandable standards. Today, at 18 this moment, I know much more from this morning's 19 sessions that would require me to say I want to 20 amend this because I think I have more clarity on 21 what you mean by the 24, and then having Version 22 II follow after that 24 month period and not in 1 series, as opposed to overlapping in great 2 extent. 3 Our third recommendation is for you 4 to carefully consider the comments from those who 5 must implement and run the elections: the 6 election administrators that are out there and 7 those who design, develop, build - and the ones 8 you have heard from this morning - test, and 9 deliver this very unique specific technology. 10 So, in conclusion, we thank you 11 again for this opportunity to share our 12 perspective on Version I, its content and 13 challenges around the implementation. 14 We also hope, as you gather input 15 over the 90 days and consider the feedback, that 16 you will call on all of us to assist in any way. 17 Certainly there are aspects of the 18 areas that will require modification and others 19 that raise substantial concern. However, 20 overall, this is a step in the right direction. 21 And, again, we applaud you for it. 22 Again, we will be following up with 1 additional and more specific written feedback 2 almost weekly. In the meantime, please accept 3 these comments as a first contribution to this 4 process. 5 CHAIR HILLMAN: Thank you. 6 Mr. Adler, before you begin, I want 7 to make a clarification here for the benefit of 8 the record. 9 The proposed Voluntary Voting System 10 Guidelines that the Election Assistance 11 Commission has put forth, we do not consider this 12 a Version I. We are not thinking about this in a 13 Version I, Version II. 14 And I understand there was earlier 15 language that may have led people to believe 16 that. 17 But our responsibility was to issue 18 guidelines and the Technical Guidelines 19 Development Committee did put forth 20 recommendations that would augment and update the 21 2002 standards, as well as covering some 22 additional components. 1 We recognize that it was almost 2 impossible, in nine months, for the Technical 3 Guidelines Development Committee and NIST to have 4 done a thorough covering of everything that needs 5 to be addressed. And so we expect and do plan 6 that we will add to the proposed guidelines. 7 But we do not discuss it -- We 8 aren't, at this point, discussing it as a Version 9 I/Version II. We don't want to confuse anybody 10 that we are going through all this effort on a 11 Version I that might be replaced in very short 12 order by a Version II. 13 There will be one version of the 14 guidelines and they will be updated and augmented 15 as the Technical Guidelines Development Committee 16 and NIST has time to do work on, but not in any 17 way to have people think that this will be set 18 aside and totally replaced. 19 MR. GROH: And I thank you for 20 clarifying it. That is one of the major intents 21 of coming to meetings like this, to have this 22 interface. 1 We get clarification where there may 2 be some, as we have interpreted, a wrong 3 misinterpretation. 4 CHAIR HILLMAN: No problem. I just 5 wanted to make sure that we were all -- 6 The other point that I meant to 7 address before the panel presentation began is 8 how we determined what the presentations from the 9 vendors would be today. 10 We were looking for a blend. There 11 are many vendors who make voting systems. 12 And I am just going to call on our 13 General Cousel, Juliet Thompson, to give a brief 14 explanation as to what we were seeking to 15 accomplish by the diversity and representation on 16 this panel. 17 MS. THOMPSON: Thank you, Madam 18 Chair. 19 And, as you aptly pointed out, there 20 are many voting system vendors out there and we 21 expect to hear from all of them during the 22 process of this comment period. 1 But for the purposes of this panel, 2 we applied several factors, one of which was the 3 types of voting systems that they manufacture: 4 to include optical scan; to include touch screen; 5 and to include full face DRE voting systems, as 6 well as components. 7 And the second set of factors would 8 be, of course, to be representative of those 9 vendors that are out there in the marketplace, 10 some large vendors, some small vendors. 11 CHAIR HILLMAN: Thank you. 12 Mr. Adler. 13 MR. ADLER: Thank you, Madam Chair, 14 Vice Chair DeGregorio, Commissioner Martinez. 15 Am I on? 16 THE CHAIR: Yes, 17 MR. ADLER: I am now. 18 THE CHAIR: Just move the mike 19 closer. 20 MR. ADLER: Mr. Wilkey, Ms. 21 Thompson. 22 Thank you for having me here today. 1 My name is Jim Adler and I am President of Vote 2 Here, a company I founded in 1998 with a vision 3 to create for more transparent and auditable 4 elections. We provide independent verification 5 technology for both electronic and paper ballot 6 processing, to prove the voting machines, ballot 7 processing and back-end tabulation systems are 8 performing properly. 9 In the U.S. and abroad, our 10 technology has served more than 12 million 11 voters. I currently coach the IEEE Voter 12 Verifiable Standards Committee. I have 13 testified before the U.S. House Government Reform 14 Subcommittee or TGDC, your TDGC, the National 15 Institutes of Standards and Technology and many, 16 many, many State Legislatures 17 Since you will be hearing a lot 18 about what is wrong with these standards in the 19 coming months, I thought it might be a good 20 opportunity to tell you what's right about them. 21 After being involved with the IEEE 22 voting system standards for the better part of 1 two years, I understand what a daunting process 2 this really is. Given the time constraints and 3 contention around some of these topics, I think 4 getting this body of work completed on time is 5 really a huge accomplishment. 6 It is comprehensive, well-organized 7 and thorough. You, the TGDC, NIST should all be 8 commended on getting to this point. It's a great 9 accomplishment. 10 To respect the time constraints, I 11 won't make any detailed comments today. But, as 12 many others have said, follow up. 13 I would like to touch on two broad 14 topics, primarily with regard to the issue of 15 independent verification, one on classifying what 16 you are calling independent dual verification 17 solutions, and the other on component 18 certification. So let me deal with those in 19 turn. 20 In the April 2005 Voluntary Voting 21 System Guideline submittal to EAC, NIST and TGDC 22 widely settled on defining independent dual 1 verification - and I will use the acronym IDV for 2 that - as a new class of voting system component. 3 In the original submittal to EAC, 4 TGDC defined independent verification, even set 5 requirements for all IDV systems. And then in 6 subsequent sections they defined requirements for 7 a split process IDV systems, witness IDV systems 8 and cryptographic IDV systems and VVPAT IDV 9 systems. 10 In the introduction to the VVPAT 11 requirements, the original TGDC standards say, 12 "VVPAT is a form of independent verification 13 system." This was an excellent classification. 14 It defined a variety of safety 15 equipment, if you will, for voting systems, just 16 like Federal standards define automotive safety 17 equipment - like seat belts, airbag, windshield 18 safety glass, and other safety innovations that 19 improve passenger safety. 20 Specifically, in the IDV sections, 21 in this instance the TGDC did a great job of even 22 establishing detailed requirements on IDV 1 alternatives. 2 This classification was consistent 3 with public comments by the EAC Commissioners, 4 previously and today, on the availability of 5 other IDV methods to accomplish the goals of 6 election, confidence and audit. 7 However, in the proposed VVSG, the 8 classification has been lost. In Section 6 the 9 VVSG now sets VVPAT requirements in a vacuum. 10 The other IDV alternatives available 11 today are put in an appendix in the back of the 12 bus, and the VVPAT standards are provided outside 13 this appendix with no context. 14 The VVSG doesn't say what the VVPAT 15 is or what it intends to accomplish. It just 16 says that the requirements are provided for 17 system certification, testing for those states 18 that have decided to include VVPAT. 19 Understandably, this is a nod to the 20 political activism that has gripped many State 21 Legislatures on this issue, but the standard is 22 needlessly silent on the context. 1 And there is a larger issue here. 2 As we talked about today, these 3 standards will establish guidance now and policy 4 for years to come. Several states are now 5 recognizing the need for IDV but are savvy enough 6 to perceive the unintended consequences of VVPAT. 7 Specifically, they see the erosion 8 of voter privacy through the current reel-to-reel 9 VVPAT designs, the lack of an accessible VVPAT 10 for disabled voters, the dilemma faced when VVPAT 11 ballots are mishandled and don't match the 12 election results, and even question the 13 effectiveness of VVPAT in bolstering voter 14 confidence when in early tests voters are not 15 even checking the VVPAT printout or voters want a 16 receipt they can take home, just like at the ATM. 17 My point is not that VVPAT is all 18 bad. But the harmful thing about recent activism 19 and legislation is that these issues of efficacy, 20 privacy and accessibility have not been carefully 21 considered, which means that VVPAT may very well 22 undermine the same confidence its supporters have 1 aimed to bolster. 2 Any election solution, especially an 3 IDV solution, should be required to demonstrate 4 real measurable effectiveness before it is 5 legislated or established as a standard that 6 stands alone without equal. 7 So I would urge you to revert back 8 to the classification that the technical experts 9 at NIST and the TGDC recommended in their 10 original submission last April. This will allow 11 for best solutions that meet these efficacy, 12 privacy and accessibility requirements needed for 13 all voters, as well as will minimize the costs 14 and administrative burden. 15 Of course the objective that is 16 often raised is that other methods of IDV are far 17 into the future and have not yet been Federally 18 qualified, certified. 19 Well, the fact is, IDV approaches 20 have been researched for the last 25 years and 21 have been put into products over the last seven. 22 The real problem is that IDV 1 solutions can't get independently certified, 2 which brings me to my next comment. 3 The VVSG establishes IDV standards. 4 But current NASED polcy and VVSG Volume II don't 5 allow for certification of independent 6 components. 7 This policy allows for a complete 8 voting system to be certified. And I would ask 9 that you reconsider this policy as Federal 10 testing is transferred to the EAC. 11 The effect of this current 12 system-only certification policy is that 13 components like IDV components can only be 14 certified within an entire voting system, a 15 policy that leads to a classic Catch 22 16 conundrum. 17 With the current policy, voting 18 system vendors won't integrate and certify best 19 of DRE components unless customers demand them. 20 But jurisdictional customers are 21 reluctant to demand them or statutorily can't 22 demand them unless they are Federally certified. 1 The dynamic is especially accute in 2 several states right now that want IDV choices 3 but can't get them because the current process 4 does not allow for component level certification. 5 I think Carolyn Coggins talked about 6 the paralysis that is out there on this issue. 7 This is one of them. 8 So I would like to make a modest 9 proposal for component level Federal testing to 10 help rectify the situation and allow market entry 11 for best-of-breed specialties. 12 Component level Federal testing 13 would be in two phases. The first phase would be 14 component level certification testing, which 15 would require component vendors to deliver a 16 technical data package and a test harness. The 17 test harness would be used to test the component 18 against the component vendor's TDP and the 19 applicable EAC standard. When successful, a 20 component level certification number would be 21 given. 22 The second phase would test the 1 integrated voting system with the pre-certified 2 component. This would test for a successful 3 integration of the certified component into the 4 voting system. 5 And when successful, the voting 6 system would be given a certification number that 7 included the certified component. 8 Assemblance of this process already 9 does exist today when a vendor of a certified 10 voting system submits, say, an audio component 11 for Federal certification. That component is 12 evaluated on its own merits at the component 13 level and then as part of the entire voting 14 system at the system level. 15 The new policy would remove the 16 current barriers to innovation by allowing 17 Federal certification of components and allow 18 interested states and counties to escape their 19 Catch 22 deadlock. It provides jurisdictions 20 choice and assurance that components meet Federal 21 standards, while getting the bulk of Federal 22 testing done early. 1 As you know, jurisdictions face 2 looming immovable deadlines. I know, I guess 3 it's a requirement that every panelist must say 4 that. 5 So anything to expedite testing 6 would help. 7 So in conclusion, I would like to 8 make a comment about the use of technology in 9 elections. Frankly, there is a distinct 10 anti-technology movement that has gripped 11 election reform. Given the current polarized 12 political climate and general technology fatigue, 13 the backlash is understandable. 14 However, in my home state of 15 Washington, we all witnessed the recent 16 Governor's race that dragged on for six months. 17 An election where 90 percent plus of 18 the ballots were cast on paper, it is often said 19 by voting critics that a voting system must prove 20 to the loser that they lost. 21 Well, the paper trail, as tested in 22 Washington, did not meet that standard. 1 After watching this, it is clear 2 that election officials simply just don't have 3 the technological tools to deal with the 4 increasingly close races and heated scrutiny and 5 heightened public attention. 6 It's like we are asking them to 7 accurately weigh a flea on a bathroom scale. 8 It's just not a fair position to put them in. 9 The current VVSG IDV standards are a 10 step toward providing these tools. But 11 expediency is key. 12 In many ways Federal agencies are 13 not geared to build airplanes while they fly 14 them, but this is the unenviable position you 15 find yourselves in. 16 Delays like the 90 day comment 17 period and the 24 month effectivity date are 18 necessary and unavoidable. 19 The comments I have made today will 20 help immediately by providing guidance to 21 jurisdictions that are currently looking for more 22 effective IDV, and by providing a Federal testing 1 policy that can expediently and responsibly 2 certify them. 3 As you deliberate on the testimony 4 you hear today, it is important to keep in mind 5 that your policies will have one of two possible 6 outcomes, especially in the area of independent 7 dual verification. 8 Either IDV competition will be 9 discouraged with jurisdictions being locked into 10 VVPAT, or jurisdications will have choice in 11 their IDV solutions, fostering a climate of 12 innovation. 13 The VVPAT movement, I am afraid, has 14 largely ignored the competing requirements of 15 privacy, accessibility and voter confidence. 16 A climate for innovative IDV 17 solutions just won't be able to get away with 18 ignoring these competing requirements because the 19 competitive market pressures just won't allow it. 20 So, again, I thank you for the 21 opportunity to speak to you today. I certainly 22 will be following up with more detailed comments 1 and, of course, happy to answer any questions. 2 CHAIR HILLMAN: Thank you. 3 Mr. Charles. 4 MR. CHARLES: Good afternoon, Chair 5 and Commissioners. 6 Sequoia Voting Systems has a 100 7 year history of providing election equipment, 8 supplies and support for state and local 9 officials. 10 We print ballots, manufacture 11 optical ballot readers and provide two different 12 types of direct recording electronic voting 13 systems. 14 During the 2004 Presidential 15 election, Sequoia was the largest provider of 16 electronic voting machines in the nation. And we 17 were the first national company to provide a 18 voter verifiable paper record on electronic 19 voting equipment in a major election. 20 We appreciate the opportunity to 21 participate in this hearing today and commend the 22 EAC, the TGDC and NIST on moving quickly under 1 under tight timetables to assemble the latest 2 draft of the amended Voluntary Voting System 3 Guidelines. 4 My comments today will focus on the 5 following key areas: first - the content of the 6 guidelines generally; second - the timeline 7 required for compliance with the guidelines and 8 the expectations of the marketplace; the 9 effective date of the regulations; the Federal 10 testing and certification process; the state 11 testing and certification process; and, finally, 12 the need to evaluate and revise the guidelines 13 over time. 14 While we will reserve detailed 15 comment on the specific components of the 16 guidelines for our written submission, we believe 17 that by and large the draft addresses the types 18 of issues that needed refinement from the 2002 19 standards, particularly in the development of the 20 optional requirements for voter verifiable paper 21 records in the inclusion of human factors and 22 security criteria. 1 We will, however, provide the 2 Commission with a lengthy set of written comments 3 that address areas that we believe warrant 4 correction, clarification or revisions. 5 Our primary concern regarding this 6 draft of the guidelines is the inclusion of 7 requirements that are not necessarily testable by 8 the voting system testing labs because they are 9 either ambiguously worded or because they rely on 10 the manner in which the system is implemented, by 11 local administrators, rather than a design of a 12 system itself. 13 These standards and the systems that 14 result from them are but one piece of the 15 successful conduct of elections. The policies, 16 procedures and people that conduct elections are 17 of equal importance. But to the extent possible, 18 that document should address technology 19 requirements and leave the proper implementation 20 of the system to state and local statutes, 21 procedures and best practices. 22 Our written remarks will attempt to 1 point out those areas which we believe are better 2 suited for best practices guides and local 3 procedure than inclusion in these technology 4 guidelines. 5 Through no fault of the Commission, 6 many observers and officials have operated under 7 the false assumption that the adoption of these 8 voluntary guidelines at the national level will 9 somehow be required before they could comply with 10 the statutory mandates of HAVA. 11 With the 2006 deadline for 12 compliance rapidly approaching, it is important 13 now more than ever for the Commission to help 14 election officials realize that these standards 15 will not be in place and operational prior to the 16 date that equipment purchase decisions will have 17 to be made for compliance with HAVA. 18 Once the guidelines are finalized 19 and ready to be approved later this year, 20 technology providers will develop and implement 21 any required revisions to hardware and software, 22 the testing authorities will need to be certified 1 to test against the new criteria, and the EAC 2 will need to finalize its process for managing 3 certification requests and interpreting the 4 guidelines as testing gets under way. 5 Once Federal and state testing is 6 completed and the products are available, local 7 officials will need to acquire the updated 8 systems, companies will need to manufacture and 9 deliver equipment, software upgrades will need to 10 be deployed, poll worker training manuals and 11 courses will need to be modified and voters will 12 need to be educated about the changes. 13 There simply isn't enough time for 14 that all to happen prior to the 2006 primary 15 elections in many cases. 16 The last time standards were 17 modified, it was a full three years before 18 equipment tested to the new standards was 19 available in the marketplace. While many of the 20 new features and requirements in this draft are 21 already incorporated into many systems, testing 22 against these standards will not be possible for 1 some time. 2 Fortunately for election 3 administrators, time required to meet these new 4 standards should not be confused as an impediment 5 to state or county compliance with HAVA mandates 6 for 2006. There are a large number of voting 7 systems available in the market today, both with 8 and without voter verifiable paper records, that 9 have been tested under both the 1990 on 2002 10 standards and will allow election officials to 11 meet the demands of Section 301 of HAVA. 12 One of the most important decisions 13 facing the Commission will be selecting the 14 appropriate timelines and details associated with 15 the implementing of the new guidelines. 16 The draft discusses a 24 month 17 period after adoption before the guidelines will 18 be in full effect. 19 But it is unclear if that will 20 result in the decertification of all non- 21 compliant systems that have been successfully 22 tested to earlier standards, or if the 1 implementation date will simply prevent future 2 certification of non-compliant systems. 3 The ramifications of the way the 4 effective date is implemented warrant 5 considerable review and discussion with state and 6 local officials. 7 Continued ongoing changes to 8 standards may be important and useful, but they 9 will also require continued funding to modernize 10 equipment at the local level. It's funding that 11 isn't currently contemplated at the Federal level 12 and I don't think most states and counties have 13 anticipated funding on an ongoing basis. 14 The EAC should consider the 15 reasonable likelihood of Federal funding before 16 determining the manner in which this and future 17 standards will need to be implemented. 18 Voting systems are comprised of a 19 series of components that can be tested against 20 existing standards. Components can include the 21 voting machines, the software, ballot activators, 22 et cetera. 1 When qualified components are 2 combined to create a system, the entire 3 end-to-end system is tested before a Federal 4 qualification number is assigned. When any 5 aspect of a component is revised, that component 6 in the entire system must be tested together. 7 If components have been tested to 8 different sets of standards, the complete system 9 is only qualified to a particular version of 10 standards when every component of that system has 11 been tested to that same level. 12 For example, the complete system is 13 only certified to the 2002 standards after all 14 components are qualified to the 2002 standards. 15 If one out of 10 components is qualified to the 16 1990 standards, then that standard still carries 17 the 1990 qualification number. 18 If in setting the effective date of 19 these guidelines the EAC no longer allows 20 modifications to the older systems in use today 21 without bringing out pieces of hardware and 22 software up to new standards, local officials who 1 have a 1990 or 2002 system could be significantly 2 affected when local/state law changes occur. 3 If a county needs to modify one 4 component of their system to comply with the new 5 state law, it could be faced with a possible 6 replacement of an entire voting system that would 7 otherwise be perfectly functional and compliant 8 with prior standards, state needs and HAVA 9 mandates. 10 However, if provisions to individual 11 components can be tested to the latest standards 12 while the complete system retains a certification 13 number reflecting the standards against which it 14 was originally qualified, there shouldn't be any 15 significant fiscal burden imposed on local 16 officials. State law changes will be able to be 17 accommodated more readily. 18 Some states accept Federal 19 guidelines for certification; other states 20 conduct their own testing process; still others 21 combine Federal testing with state reviews. The 22 multiple layers of often duplicative review are 1 costly, time consuming and delay the latest 2 innovations from getting into the market. 3 In developing test policies, we 4 encourage the EAC to invite states to provide 5 Federal testing authorities with a checklist of 6 local requirements so that duplicative 7 examinations can be streamlined and state 8 resources can be saved through concurrent state 9 and Federal testing. 10 Once the guidelines and testing 11 procedures are in place, the EAC will need to 12 evaluate them over time to identify and correct 13 potential weaknesses. We suggest this review 14 should not begin until after the newly qualified 15 systems are deployed and in use for at least two 16 elections. 17 One of the greatest mistakes this 18 Commission could make would be to revise 19 standards too frequently without providing ample 20 time for a particular version of the standards to 21 be put in place and studied. 22 Many election officials will tell 1 you that it takes at least two elections to work 2 through all of the logistical and procedural 3 issues that come with the transition to a new 4 voting system. 5 The quidelines before you today are 6 the result of work that started before even one 7 major election was subjected under the 2002 8 standards. 9 We recognize that with the passage 10 of HAVA, Congress charged you with the duty to 11 provide for the certification and decertification 12 of voting systems and that state policy decisions 13 created a need for VVPAT guidance. 14 But we strongly suggest that once 15 these standards are adopted, they are given time 16 to be tested in the field before a new batch of 17 requirements are developed. 18 The rapidly changing standards 19 process has required companies in our industry to 20 spend an inordinate amount of time revising 21 technology to meet changing guidelines, when that 22 time could have well been spent adding the 1 features and benefits, like VVPAT, that have been 2 demanded by the market in many places. 3 Sequoia Voting Systems believes that 4 when completed, the voluntary systems guidelines 5 will continue to ensure that voting technology is 6 reliable, secure and accurate. 7 To assist the EAC to make the most 8 of this project, we respectfully offer the 9 following recommendations: 10 First, as you are doing today and as 11 you have done throughout the policy decisions 12 this Board has made, solicit comprehensive verbal 13 and written comments from all stakeholders; 14 Second, provide a detailed public 15 timeline to help local officials understand how 16 we get from the draft regulation phase of this 17 process to the evenutal use of VVSG compliant 18 equipment at the polls; 19 Third, rather than waiting 24 months 20 to implement the proposal, allow the regulations 21 to take effect immediately upon adoption or as 22 soon as testing authorities are able to review 1 the systems, provided, however, that individual 2 components in the systems may still be qualified 3 against prior versions of standards, as long as 4 they are appropriately noted as such on testing 5 reports and on published lists of certified 6 equipment; 7 Fourth, facilitate concurrent 8 Federal and state testing through the development 9 of a checklist of state-specific criteria which 10 can be tested by the federally approved 11 laboratories, as necessary, to help streamline 12 the state certification process; 13 Fifth, allow equipment tested under 14 these regulations to be deployed and monitored 15 for at least two elections before initiating a 16 new set of Federal criteria. 17 We appreciate the invitation to 18 provide our thoughts to this hearing and welcome 19 the opportunity to continue to work with the 20 Commission on this project and we plan to submit 21 our detailed written comments on a line-by-line 22 basis within the next couple of weeks. 1 Thank you. 2 CHAIR HILLMAN: Thank you very much. 3 Mr. Vadura, we are at somewhat of a 4 little handicap here because we don't have copies 5 of your -- 6 MR. VADURA: I understand. 7 CHAIR HILLMAN: So if you would just 8 be so kind in making your presentation to bear 9 that in mind. 10 I mean, like he speaks fast but I 11 can read fast, so I was able to keep up with him. 12 So just bear with us as we take notes from your 13 presentation. 14 MR. VADURA: That's fine. Can 15 everybody hear me? 16 Flight schedules last night 17 prevented me from printing them, so I will 18 provide copies to the staff. 19 Good afternoon, Madam Chair and 20 Commissioners. My name is Dennis Vadura. I am 21 co-founder and chairman of AccuPoll and AccuPoll 22 Holdings. We are the only public company in the 1 space that is dedicated solely to voting. 2 And I would like to thank the EAC 3 for inviting AccuPoll and myself to participate 4 in this panel. 5 AccuPoll has been promoting a voter 6 solution that includes a voter verified paper 7 audit trail, or what's now called VVPAT, since 8 AccuPoll's inception. 9 Our view is that VVPAT is to 10 electronic voting what stairs are to tall 11 buildings. It is the essential safety net that 12 allows for emergency action in the event of a 13 disaster. 14 The key to a legitimate democratic 15 government is a trusted election process. 16 America works hard to promote democratic ideals 17 all over the world and with HAVA we are making 18 sure that we provide trusted and private 19 elections for all of our citizens. 20 AccuPoll is all about promoting 21 trust, accessibility and practicality in the 22 voting process. 1 In this regard we welcome the 2 updated standards for voting systems and would 3 like to offer the following comments with respect 4 to these new proposed standards. 5 We believe that it is possible to 6 create a VVPAT that is both accessible and 7 private. The disability community has fought 8 hard for HAVA and have gained the right under 9 HAVA to vote unassisted and in private. We see 10 no reason that a voter verified paper trail needs 11 to be engineered such that it removes any of 12 these rights. We welcome the inclusion of these 13 HAVA requirements in the revised standards. 14 AccuPoll has recently made changes 15 to its voting system such that we can demonstrate 16 today a practical VVPAT solution that is both 17 accessible, private and, based on our current 18 understanding of the standards, essentially 19 compliant with the proposed standards. 20 We will be demonstrating our 21 solution at the upcoming NASS conference and 22 encourage all that are interested to come and 1 view our approach. 2 The key features of our approach is 3 audio feedback of the VVPAT for every voter, 4 whether they are sighted or not, provisions for 5 privacy of the voting process and of the VVPAT 6 for the voters that require additional assistance 7 as they complete the voting process, and still 8 maintain their privacy. 9 AccuPoll essentially agrees with the 10 disability community that optical scan ballot 11 marking systems are not accessible voting systems 12 and are, therefore, not complaint with HAVA 13 accessibility requirements. 14 Optical scan systems on their own do 15 nothing to help a disabled voter vote in private 16 and provide no easy and accessible confirmation 17 that a voter's intent was correctly read or 18 recorded. 19 Ballot marking systems that 20 complete an optical scan ballot for a disabled 21 voter do nothing to assist a voter that is not 22 able to handle a paper ballot on their own and, 1 therefore, may create a privacy issue for a 2 disabled voter. 3 AccuPoll does, however, have issues 4 with certain aspects of the proposed voting 5 system standards. Given the recent update, like 6 our colleagues here, we will be providing a 7 detailed set of comments regarding specific 8 issues that we see with the current draft of the 9 standards. 10 Most notably, we see some possible 11 ambiguities and a need for additional clarity 12 with respect to certain certification issues and 13 limits as were discussed by Carolyn earlier. 14 In keeping with the practice of the 15 Federal Voting System Standards that set 16 requirements and not implementation choices, the 17 current standards for voting systems must set 18 requirements rather than mandate engineering 19 solutions. In this way the standards would allow 20 the marketplace to develop the most cost 21 effective and practical solution while ensuring 22 compliance with the standards and with HAVA. 1 We are also concerned with too many 2 changes or evolution cycles and amendments to the 3 standards. 4 While it may take the ITA's a short 5 period of time to ramp up and test against new 6 standards, it takes at least six months for 7 vendors to update and certify their systems to be 8 compliant, after which there are renewed state 9 certification requirements before the revised 10 changes can be made available to customers. 11 The state certification process may 12 take in excess of an additional six months. 13 We, therefore, suggest that the 14 revised standards be finalized, issued and not 15 altered for at least two years. This excludes 16 interpretation bulletins that would clarify an 17 interpretation of the requirement, rather than 18 modify or add new requirements. 19 We owe it to our customers and to 20 the ITA's to maintain stability in the 21 requirements so that products can be engineered 22 and marketed on reasonable timelines and costs in 1 what really is anything but a traditional 2 marketplace. 3 In our view, it is better to wait 4 six months and issue a document that is well 5 thought out, rather than issue early and amend 6 later. Thank you. 7 CHAIR HILLMAN: Thank you very much. 8 Commissioners, we are running a 9 little bit late on this panel. So what I propose 10 we do is to make up our time into our break 11 period. 12 We will still plan to have the third 13 panel start its presentation at 3 o'clock and we 14 will need a few minutes to just set the table up 15 for the third panel and that would give us the 16 ten minutes per Commissioner questioning, that 17 will allow you time to engage with the panelists. 18 So, if that works, Commissioner 19 Martinez. 20 COMMISSIONER MARTINEZ: Thank you, 21 Madam Chair. 22 My thanks to all of you for making 1 the time and the effort to be here and to testify 2 and certainly I look forward to receiving more 3 comprehensive, as I know you all will provide, 4 written testimony during the 90 day comment 5 period on the proposed VVSG. 6 John, let me start, if I could, from 7 your perspective. I think one of the things that 8 you have identified is the difficulty in the 9 language in the current draft of the VVSG, which 10 would require randomization of the VVPAT paper, 11 as opposed to, I guess, the current technology of 12 reel-to-reel. 13 Just, if you can talk a little bit 14 on that from your perspective, why that is 15 such a significant change from the vendors' 16 perspective. 17 MR. GROH: Well, again, I will speak 18 for to us and not try and speak for others. But, 19 you know, all of us see each other's technology. 20 And as we approach this and try to 21 target getting a solution out there as we 22 anticipated this coming, we all looked at what 1 would be a best practice from the standpoint of 2 making it easy for the poll worker to use, very 3 easy for the voter to interface with and look at 4 and understand that there were limitations with 5 how to interface with our technology. 6 One of the major concerns we had is 7 how do you handle this at a precinct when you run 8 out of paper or have a paper jam. So I think all 9 of us chose to use a reel type of voting system 10 where paper is going to roll up in a rolling 11 process. 12 If we are required to shuffle it, we 13 will have to basically scrap that and start with 14 something else that will be much different than 15 what we initially perceived and that will take -- 16 that takes some time and a lot of expense to it. 17 COMMISSIONER MARTINEZ: Is the 18 difficulty that you perceive one of technological 19 feasibility or is it more that you designed it in 20 a way that you thought would be more poll worker 21 friendly, if you will? 22 MR. GROH: As Mr. Adler has talked 1 about, there are other methodologies of doing 2 this. So this is one. And understanding we do 3 not adhere to this as the only prescriptive way 4 to do it, but if you are going to offer it as one 5 of the solutions, you also must recognize that 6 poll workers and election administrators need to 7 manage this and handle it. 8 And, as we know, loose pieces of 9 paper have a tendency to get lost. 10 Paper rolls then that are going to 11 have to be cut and also be kept for cancellation 12 on ballots if somebody wants to cancel that 13 ballot or that record on there, when you start 14 having them now cut and commingle, we just saw 15 the inevitable problems with those. 16 COMMISSIONER MARTINEZ: I appreciate 17 that, John. 18 Anybody else on the panel would like 19 to comment? 20 Mr. Adler. 21 MR. ADLER: I think that every 22 election system has to meet simultaneous 1 requirements of privacy and accessibility and 2 effectiveness and election administration. 3 And if you relax any of those 4 requirements, you run into problems. 5 I mean, one of the objectives we had 6 was to simultaneously solve all of these issues. 7 And if you start to relax them, you start to 8 either infringe on people's civil rights or their 9 expectation of privacy or you start to create a 10 system that can't be administered. 11 And, hence the plea for a climate of 12 innovation. 13 COMMISSIONER MARTINEZ: Mr. Charles. 14 MR. CHARLES: We deployed a system 15 in Nevada that had reel-to-reel design. And what 16 we found was that it was much easier for poll 17 workers to administer that approach and more 18 prone to successful use throughout the day. 19 I think that the language in the 20 existing draft is permissive. It may permit you 21 to or -- It doesn't state that you have to cut 22 those ballots and shuffle them. 1 But I think if you added language to 2 it that permitted administrative procedures to 3 ensure the protection of the anonymity of those 4 ballots, that would go a long way to resolving 5 that concern. 6 Because what we see is a 7 self-randomization of those records. When the 8 voter comes into the polls, the voter can use any 9 machine in the polling place. 10 When those records come back, they 11 are sealed and stored and not accessible to the 12 people who may have seen how someone voted. 13 So if you can incorporate 14 administrative procedures into that or at least 15 allow the use of administrative procedures to 16 help assure that randomonization and privacy, I 17 think you accomplish two goals. 18 One, you allow the most funcational 19 and reliable technology to be used. But you also 20 solve the concerns about voter privacy. 21 COMMISSIONER MARTINEZ: Mr. Vadura, 22 any comment? 1 MR. VADURA: Our current solution 2 doesn't use a reel-to-reel, but still maintains 3 the voter privacy issues. 4 And I think the particular 5 requirement requiring the shuffling of the pieces 6 of paper is both probably unnecessary and 7 overburdensome and also probably too costly. 8 There are cheaper solutions to get that done. 9 So, administratively, in the polling 10 place, I believe in another section of the 11 standards it says that you can't, polling 12 officials can't have access to that paper trail 13 during the election day. 14 But if you have a paper jam, you 15 have to have a practical way to clear that paper 16 jam in that kind of scenario. So I think there 17 are some issues with the standards with respect 18 to that particular requirement. 19 MR. GROH: Commissioner Martinez, 20 this morning -- I want to answer a question that 21 you asked this morning and make sure all of us 22 would respond to it. 1 You had asked in the responses, will 2 you make recommendations. From our perspective, 3 as a developer, we will make recommendations in 4 ours for language. So we will try and give you 5 some words or language around that that will help 6 you get your arms around it. 7 COMMISSIONER MARTINEZ: Thank you. 8 The next question, I want to go back 9 to something that I was talking about this 10 morning with the testing labs. And I don't know 11 if everybody, if all four of you were in the 12 audience. I think Jim and John, I think you were 13 both here, so you all heard the exchange. 14 The topic that I brought up was the 15 issue of a Version II, which our Chair talked 16 about a little earlier. 17 And I just want to get - because 18 actually the response from the system, from the 19 testing labs was - it really doesn't impact us 20 greatly. But they thought certainly the vendor 21 perspective and perhaps even the state and local 22 election administrators, et cetera, that it could 1 have a greater impact. 2 Just a quick comment or two about 3 the issue of -- Because I know I am having to 4 contemplate what happens next. 5 Already we are thinking about, all 6 right, so if we go through the next 90 days and 7 we end up with the final product that we adopt as 8 the next iteration or the latest voting system 9 guidelines, then what happens? 10 Obviously, the Help America Vote Act 11 creates this umbrella organization - the 12 Technical Guidelines Development Committee - 13 which I think has brought tremendous expertise to 14 the table. 15 The question becomes how do we 16 utilize that, as well as our partnership with 17 NIST? And what happens next? Do we begin 18 working right away on a full blown version or 19 full blown revision, if you will, of what we have 20 just adopted in final form? Or do we -- 21 That's why I was trying to get 22 myself educated to the current process, and that 1 is, that when we find deficiencies in the current 2 product, that we address them to an amendment 3 process, for example. And is that perhaps easier 4 to deal with from every perspective - vendor, 5 election administrator, et cetera? 6 Any comments? 7 Mr. Charles, you have heard the 8 exchange. Any comments about that whole topic? 9 MR. CHARLES: I think it's important 10 to distinguish between the ease of the testing 11 laboratories in testing the equipment and the 12 difficulty it takes for us to make the 13 modifications and submit them to the test. 14 So there is a much lengthier time 15 for to us to do the work we need to do before we 16 get it to them. 17 I think the first step ought to be, 18 develop a way to judge any gaps in these 19 standards by observing the elections after these 20 standards are adopted in the field. 21 I think develop a criteria for 22 reviewing what happens in elections, determine if 1 there are pieces of these regulations that are 2 overly burdensome or areas that are missing and 3 develop that framework so you can study. 4 And then once you have studied them 5 in practice in an election, then you can apply 6 that to the next round. 7 But I think revising them before we 8 see them in the field is working more on theory 9 than on practical application. And that may not 10 be the best use of the time. 11 COMMISSIONER MARTINEZ: Mr. Groh, 12 any comment on that? 13 MR. GROH: Well, the competitive 14 nature of this market means that many of us are 15 already working on these. 16 But, again - and somebody used the 17 analogy of, we are trying to build a plane while 18 it's running down a runway. 19 And for us, as a vendor community, 20 there is one component I want to remind 21 everybody. We are going to run, from our 22 company's perspective, about 5000 elections 1 between now and next November. And that also 2 keeps us very busy. 3 So running these parallel paths, I 4 don't have the luxury of stopping, working on the 5 elections, fixing small things that are changes 6 in election rules. 7 So we are aggressively pursuing 8 them, but it will take us time. 9 I agree with Ms. Coggins's comment 10 this morning, that the vendors that have gone 11 through this numerous times are at a great 12 advantage because we understand it. 13 But I can tell you the difficulty 14 that we have of submitting something that we feel 15 fits and meets what the requirements are and then 16 when it is tested there are things that are found 17 that we hadn't thought of. And that's part of 18 the process that you go through to flush those 19 out. 20 But that's what takes the 18 months, 21 the 16 months to do, it's the back and forth. 22 And we can't just stop and say let's hold 1 elections for four years, get all this done, and 2 get a final product out there. 3 We are at the same time trying to 4 supply the election sites. 5 COMMISSIONER MARTINEZ: Mr. Adler, 6 Mr. Vadura, any comments?. 7 MR. ADLER: I think there is two 8 buckets here. There is quite a bit of election 9 technology that is well understood and is going 10 to change in a much less frequent rate than say 11 the side of business that, where you are focused, 12 which is on voter verified paper audit trails and 13 independent verification systems, where things 14 are theoretical right now, quite frankly. 15 And these devices are just starting 16 to move into the market and having a mechanism to 17 be able to adjust, where in many respects the 18 states are now laboratories in this field, in 19 this area. And being able to have guidance and 20 being able to draw on NIST experts and your own 21 experience to help states navigate that and have 22 a real mechanism to help the certification 1 process would be very valuable. 2 COMMISSIONER MARTINEZ: Mr. Vadura, 3 any comment? 4 MR. VADURA: Yes. I will speak for 5 us. We just completed our 2002 certification on 6 our complete system with, including a VVPAT. And 7 we got through the Federal certification process, 8 that took some amount of time. 9 We then have to now go through state 10 certifications in various other states, which is 11 fine; that's the way the business works. 12 But the issue is if you make too 13 many changes too frequently, we are in a constant 14 state of certification somewhere at that point. 15 And I don't think that benefits anybody. 16 It doesn't benefit the customers; it 17 doesn't benefit the vendors. 18 COMMISSIONER MARTINEZ: My final 19 question, then, again going back to something I 20 brought up this morning, and that is - maybe it 21 will be more in the form of a comment. 22 If somebody wants to jump in, you 1 can do so very quickly. 2 But the whole notion, there is 3 clearly a gap here. I just want to make sure 4 that you all don't disagree with the exchange 5 that we had this morning that I had with the test 6 labs, and that is, the gap is that the current 7 voting system standards that we operate under - 8 the 2002 VSS - were developed before HAVA was 9 even close to being in its final form. So we all 10 agree with that. 11 To get a system through the testing 12 process and to say we are certified to the 2002 13 VSS does not necessarily mean that you are HAVA 14 compliant, because the 2002 VSS does not test to 15 the language of Section 301 in HAVA. 16 Now I understand that many vendors 17 will take a look at, obviously, what is in 18 Section 301 and build a system to be, from their 19 perspective, compliant with HAVA. But, in fact, 20 there are no voting system standards currently 21 that test to the language of Section 301, 22 specifically to the language of Section 301(a)(3) 1 dealing with accessibility. 2 Any comment about that? And I think 3 my time is out, so any quick comment about that. 4 MR. CHARLES: I think, very quickly, 5 there are systems that meet the plain language of 6 HAVA without being tested as a standard. You can 7 look at those and determine that they are 8 accessible, they do those things, that you can 9 check and change your ballot, all those pieces, 10 under the 1990 standards as well as the 2002. 11 But, you are right, there is not a 12 HAVA certification of a system. 13 MR. GROH: I would concur with that. 14 COMMISSIONER MARTINEZ: Thank you 15 all very, very much. 16 Thank you, Madam Chair. 17 CHAIR HILLMAN: Thank you. 18 Mr. Vice Chairman. 19 VICE CHAIR DeGREGORIO: Thank you, 20 Madam Chair. 21 Actually, to follow up on what 22 Commissioner Martinez just brought up, and Mr. 1 Groh, you specifically, in one of your 2 recommendations, your first one here, you talked 3 about the EAC should clarify through a safe 4 harbor statement that states and localities that 5 want and need to make purchase decisions today to 6 comply with HAVA can do so under today's 7 standards and that systems purchased today will 8 not have to be replaced in the future. 9 Are you suggesting, through some 10 kind of statement that we would issue, some kind 11 of safe harbor statement, that these systems 12 would not have to be compliant with 301(a)? 13 MR. GROH: No, not at all. 14 But, again, we are in that gray area 15 which Commissioner Martinez just pointed out: we 16 are not testing to that. 17 But I think a common sense 18 approach - and there has been a DOJ opinion that 19 goes back to, I think, 2003, on DRE with the 20 VVPAT, of providing and meeting the level of 21 compliance that 301 requires. And I think that 22 that is do-able, but it hasn't been tested to it. 1 VICE CHAIR DeGREGORIO: And I 2 recognize there is no guidelines for this because 3 it is new to HAVA. 4 But I don't see 301(a)(3) as a gray 5 area. It is in the statute and states have to 6 comply with this. 7 MR. GROH: And the gray area to me 8 is it hasn't been tested under the 2002. But it 9 doesn't mean that some of the systems don't have 10 the elements in it that if there were a test, it 11 would pass it. 12 VICE CHAIR DeGREGORIO: Mr. Adler, 13 could you just tell us, your IDV system, your 14 independent dual verification system, how does 15 that work? 16 MR. ADLER: It's a unit that is 17 connected to an electronic voting system and it 18 gives, when the voter votes, the voter goes 19 through their confirmation process and then once 20 they confirm their vote, that vote goes into the 21 unit, and as the VVSG talks to, maintains a 22 separate copy of the vote. 1 The voter then has an opportunity to 2 get a receipt, to probe that ballot to make sure 3 that it did, in fact, reflect their intent. And 4 then they could actually, if the jurisdiction so 5 desires, give them an opportunity to use that 6 receipt to make sure the vote is actually in the 7 count. 8 So the model is very similar to an 9 ATM transaction. You get a receipt. You take 10 out your $20; you make sure the receipt matches 11 the $20. And then you take that receipt and 12 clear it against your statement at the end of the 13 month. 14 The statement at the end of the 15 month in the election context is the transcript 16 that the results are certified against. 17 So it actually provides the ability 18 for the voter to do that. 19 There is also the ability for the 20 election officials to gather these receipts and 21 verify that, in fact, the confidence in the 22 election results are there. 1 And this really goes back to what we 2 discussed at the NIST symposium back in December 3 2003, I believe, where, talking about trust and 4 confidence in election systems, where you can 5 actually prove that the voting system - either on 6 the back end - tabulation data base - or the 7 voting machine itself is not making mistakes. 8 You can prove it. You don't need to 9 assert it; you can actually prove it. 10 VICE CHAIR DeGREGORIO: Is your 11 system a standalone system or meant to be used 12 with another system? 13 MR. ADLER: It's meant to be used 14 with an electronic voting system, a minimal level 15 of interaction. 16 VICE CHAIR DeGREGORIO: Just wanted 17 to make sure I understood how it worked. Because 18 you make great reference to it in your comments 19 here regarding the guidelines that the TGDC came 20 up with. 21 And I want to make sure I understand 22 exactly how you envision this working. 1 MR. ADLER: The appendix, talking 2 about end-to-end cryptographic verification, 3 actually has a comprehensive set of detailed 4 requirements that goes through how the system 5 like this one works. 6 VICE CHAIR DeGREGORIO: Mr. Charles, 7 you, in your comments, talk about recommending to 8 us that instead of waiting 24 months to implement 9 this, that we should do it immediately upon 10 adoption or as soon as testing authorities are 11 able to review the system. But you still want 12 individual components to be qualified under the 13 2002 standards. 14 Are you suggesting that instead of 15 testing a whole system, that a system, part of 16 the system was qualified under 2002 and tested by 17 and ITA to be qualified, if something was added 18 to it, just that component should be tested 19 against these guidelines and that's something 20 that could be done right away instead of waiting 21 the 24 months? 22 MR. CHARLES: It would be faster and 1 less expensive, especially if it relates to older 2 systems where a county or state may have a 1990 3 model system, a 1990 standard system, and they 4 want to add a printer to it. 5 Some of that hardware may not be 6 economically upgradable to the 2005 or 2002 7 standards. But you could add a printer to it and 8 it's functioning for some time and be able to 9 test. 10 Right now you can't test a component 11 and add it to a 1990 system to help a state meet 12 state law without putting the entire system up to 13 2002. 14 That places a pretty significant 15 burden on states or counties that may have a 16 statewide 1990 system and cannot afford to 17 implement the hardware at the necessary level. 18 It's really not just for this 19 version but, in going forward, I don't think that 20 there is a reason, that the different versions of 21 standards cannot co-exist, as long as when you 22 are testing things, you are trying to test them 1 to bring them up and you qualify - and the 2 certification numbers or components of systems - 3 what level that system has passed so that 4 purchasers of equipment will know what level. 5 But that way, if possible, it would 6 be nice to help states avoid replacing entire 7 systems of hardware because the standards may 8 change in the future. 9 VICE CHAIR DeGREGORIO: But when you 10 add new components like that, even if it's just a 11 printer, sometimes doesn't that affect the whole 12 system or could affect the whole system? 13 Sometimes you add a new printer and 14 it's the wrong driver because the driver is ten 15 years old with the old system. And you are 16 getting in there and you are changing the driver 17 on the software on a system that has been 18 certified. 19 MR. CHARLES: But you still test 20 that system end-to-end. So you would confirm 21 that that 2002 standard printer matches up with 22 the 1990 equipment and you could do an end-to-end 1 test from software to election setup to election 2 operations to printer, you can confirm that that 3 works properly within that system. 4 And that's how we moved from the 5 1990 to the 2000 standards, each piece or each 6 component would get upgraded, but they would be 7 tested as a complete system. So as long as you 8 do that complete system test, you would be able 9 to ensure the reliability and durability of that 10 component. 11 VICE CHAIR DeGREGORIO: I know that 12 the three of you provided us with written 13 testimony - and Mr. Vadura you are going to 14 provide us with yours - and I appreciate your 15 testimony. 16 I have been in this business a long 17 time and know many of you for many years because 18 I was a Director of Elections in St. Louis County 19 and have been coming to IAKVIAC (ph) meetings 20 since 1986 and going to the vendors' show and 21 seeing all the new and innovative equipment and 22 meeting new people in the last few years through 1 my work that I did with IFIS (ph) internationally 2 and then in this job. 3 And I have to say, I am amazed every 4 year - and especially in the last two years - to 5 see the new products on display and how you are 6 trying to respond to the marketplace. 7 And, you know, the fact of the 8 matter is that from 2000 to 2004, 25% of the 9 country saw new equipment. 10 From 2004 to 2006 we have this 11 challenge that 30% of the country, though, is 12 using lever machines and punch cards, 13 particularly, are going to be changing over into 14 new systems, so it brings new challenge. These 15 guidelines bring a challenge to you. 16 But the Federal Government, for the 17 first time, is providingly $3 billion to get part 18 of this, part of the equipment part of it done. 19 Are you all up to the challenge and 20 your colleagues - because you work with your 21 colleagues I know. You are friendly - you are 22 competitors but you are friendly competitors, I 1 always felt so. 2 Are you up to this challenge, do you 3 think? Is there enough vendors in the 4 marketplace to meet this demand and to serve not 5 only the needs of election officials - but, let's 6 face it - we are serving the needs of the public 7 out there, and a skeptical public - there is a 8 chunk of them. 9 And the survey data, Commissioner 10 Martinez talked about it this morning, of about 11 one-quarter of the voters that don't quite trust 12 the systems. And there is that skepticism out 13 there. 14 Are you all meeting that challenge 15 and is there anything that in these guidelines 16 that we can add, in addition to what you said 17 here, that might be helpful to you in helping you 18 to meet the challenge? 19 We will start here with Dennis and 20 go this way (indicating). 21 MR. VADURA: I think we are all 22 working really hard to meet those challenges. 1 Having a staple set of standards 2 helps to meet those challenges because you can 3 predict what needs to be done to get it certified 4 in time. 5 There is an effective barrier 6 to entry for new vendors coming into this 7 marketplace, as of today, which is certification. 8 I don't believe anybody can come 9 into this marketplace starting today and get up 10 to speed to deliver anything by 2006. So I think 11 everybody else here would agree with that. 12 So, is there enough vendors? We 13 will find out. I hope so. 14 MR. CHARLES: I think there is. I 15 think that we are not going to sleep much in the 16 next 24 months. I think we are all going to - 17 every county election official, every state 18 official and every company - will be pretty 19 heavily taxed on this. But we have been 20 preparing for it for some time, not just on the 21 manufacturing side, but making sure that there 22 are enough people trained to support the local 1 elections. 2 My biggest concern is for small 3 rural jurisdictions that really haven't invested 4 a lot of time and may not command the market size 5 to get the level of support that they may need. 6 I hope that every election official out there 7 realizes that time is of the essence. The sooner 8 they can make decisions and start, the better 9 they will be and the better that we as an 10 industry will be able to support them and adjust 11 resources accordingly so that they get the 12 support they need. 13 MR. ADLER: I think there is a level 14 of passion for elections on no matter what side 15 of this table you are on. 16 And that translates over to a level 17 of engagement with the public and putting in the 18 sleepless nights, not because you have to, it's 19 in your blood. You just want to and are driven 20 to deliver and make sure every vote counts. 21 And engage the skeptics, quite 22 frankly. I know that we have been very open to 1 doing that, during this more difficult period. 2 So I think the process is 3 invigorating. I think it stresses all of us out, 4 we all know that. But it's also, we are doing 5 good work here. 6 And this is not a -- This is a 7 contentious process. I think that's not all a 8 bad thing. But we just have to be honest, 9 respectful and engaged. 10 The work load over the next 24 11 months is going to be horrendous - I will tell 12 you that - no matter where you are. 13 But I think that's not surprising. 14 I am certain it's not a surprise to you, Paul, or 15 anybody else, having done elections for so many 16 years. 17 Having 30% of the country change 18 election equipment is going to be tough. And I 19 agree with Alfie here, for small jurisdictions 20 it's going to be especially difficult. 21 MR. GROH: We look at this from many 22 perspectives. And, again, this is not just a 1 single thread where if we deliver the hardware, 2 that solves the situation. From my company's 3 standpoint, we know we can deliver the technology 4 and get it there. 5 But what we are concerned about is 6 can the election administrators and the poll 7 workers and the voters feasibly take this 8 dramatic change that's going to happen very 9 quickly? And it is not equal across the United 10 States. 11 We have election laws that are 12 different in different places that put an added 13 burden. 14 We heard New York City, one of the 15 election administrators this morning, talk about 16 that, of trying to phase this in. 17 So our concern as a company is we 18 can deliver the best technology. But if we 19 haven't done the next step, which is really 20 shared by us, by the election administrator, it 21 can look like a failure, because it doesn't work 22 on Election Day and it's an administrative or a 1 poll worker or a voter issue. 2 But the first finger that will be 3 pointed will go back to us as the election 4 supplier. So we work very hard on the 5 implementation. 6 So we can do it. We will put in a 7 tremendous number of hours into accomplishing 8 this. I do not see us having a constraint in 9 that we couldn't deliver all that we could get 10 orders for. 11 But it's the project management 12 piece of it. 13 So the other component that you are 14 addressing in HAVA is that voter education, a new 15 way of looking at poll workers that need to be 16 brought on through your college program and that 17 you have that. 18 We applaud all of those, because 19 that's what is going to be required. 20 And then I think that there is a 21 common sense or a reasonable man approach to ask 22 the question, if we deliver all this at one big 1 time and cause a major change-over, can we 2 swallow that. 3 And I guess the analogy I would give 4 is, we all know we can start on an exercise 5 program, but that first day I don't do 50 sit-ups 6 because I won't do a sit-up for about another 7 week. 8 We can't go into this on day one 9 thinking, I am going to enter the exercise 10 program full bore. I have to work my way up into 11 it. 12 VICE CHAIR DeGREGORIO: Thank you 13 all very much. 14 CHAIR HILLMAN: Thank you. 15 In the remaining few minutes, let me 16 just ask a question and then I will give a nod to 17 my colleague, Mr. Wilkey, if he has a question. 18 I have a specific question for you, 19 Mr. Groh. No, maybe it wasn't your question. I 20 think it was for Sequoia. 21 Somebody indicated, made a general 22 reference to - it was Sequoia - ambiguously 1 worded and without any specific reference. 2 And could I just ask, not for you to 3 comment here, but in your follow-up submission, 4 if you would be as specific as possible, because 5 I don't know what to respond to with that. 6 You were all in business and had 7 systems in the market in 2000; is that correct? 8 Did you have something in the 9 market? 10 MR. ADLER: No. 11 CHAIR HILLMAN: You started in '98, 12 but you certainly were in the field in 2000. 13 MR. ADLER: That's right. 14 CHAIR HILLMAN: And, therefore, were 15 following the year's, following the 2000 16 Presidential election, I'm sure, with great 17 interest with respect to the discussions about 18 various voting systems. 19 So I would ask you, it's sort of a 20 two part question. But it's the second part of 21 my question that's probably where you should 22 emphasize. 1 One would be what your companies 2 have done to address some of the deficiencies 3 that are now, that we are trying to address in 4 the guidelines. 5 But the other part of that question 6 is, is there anything in here that's a big 7 surprise to you, given what you have heard and 8 what you have been engaged in over the past four, 9 five years? Is there anything in here that 10 surprises you that you don't think you are ready 11 for or that your company wasn't expecting or 12 preparing for? 13 MR. GROH: I will go ahead and jump 14 into this. 15 No, it hasn't been. And part of why 16 it's not a major surprise to us is because we 17 have been involved throughout the process. 18 Are there little unique things in 19 there? Yes, they are. But they don't rise to 20 the level that I am surprised or that it knocks 21 me out of my chair. 22 All of us that I think have been in 1 this have been trying to work on it before we 2 actually could see the target. We knew there was 3 going to be a target around the corner. I may be 4 off a few degrees and that's going to cause me to 5 have to recheck it. 6 I think our concern, from Election 7 Systems & Software, is that we not be so 8 prescriptive or come up with a specific 9 methodology to solve something that could have 10 other means of solving. Allow us, as the 11 developers of techonology, to provide those. 12 That's been one of the challenges, 13 that ES&S has done as an innovator, we have had 14 our customer base say I want a paper verifiable 15 audit trail. 16 We have attempted to do that in a 17 format that we think is usable and we want to 18 introduce that and have that be tested. 19 We have had our customer base say we 20 want to stay with paper base. Can you provide us 21 with a solution that will allow us to meet 301 22 and allow accessibility but still maintain and 1 utilize a paper base system. We have made that 2 effort to do that. 3 And we are learning and are trying 4 to make adjustments to that. But we are driven, 5 all of us are driven, by the competitive nature 6 of what our customers will pay for, buy and want. 7 CHAIR HILLMAN: Any comments, Mr. 8 Adler? 9 MR. ADLER: I would agree with Mr. 10 Groh, that the process has been collaborative and 11 open and has built upon the work done previously 12 in a very professional way. So there were no 13 surprises really, other than those few degrees 14 that we are addressing through this comment 15 period. 16 With respect to what we have been 17 doing since 2000, I think we are in a paradigm 18 shift in elections. I think that's what 19 generated HAVA, in many respects, and we are 20 grappling with that tectonic shift. 21 And you are leading this industry 22 through that tectonic shift. And those shifts 1 are not often pretty. 2 But there has been a tremendous 3 amount of innovation. There will continue to be. 4 And what I would hope and what I see is that 5 there is a real penchant for encouraging 6 innovation as we move forward. 7 We know this is not the last set of 8 election systems jurisdictions procured and there 9 will be no changes going forward. We know there 10 will be. And this is a roadmap to take us there. 11 MR. CHARLES: Just briefly, I think 12 there isn't anything in this document that is a 13 tremendous surprise and I think that the credit 14 for that goes to the TGDC and NIST in having 15 experts on there that really understand the 16 elections process and the work that they put in. 17 I know that the hours they put in 18 were incredible under a tight timeline. 19 If there was one recommendation that 20 I would have for the next version of this, it 21 would be to invite representatives of the vendor 22 community to participate on the TGDC so that 1 there will be somebody in addition to the testing 2 authorities and the election officials, the 3 designers of the equipment, so that we can have a 4 more direct involvement in the discussions that 5 take place. 6 But even without that, I think they 7 did an excellent job in keeping this to something 8 that was relatively expected by the industry. 9 MR. VADURA: I would argue with 10 that. There is no real surprises in this other 11 than, I think on the next round, I agree with Mr. 12 Charles, let us contribute. Because I think some 13 practicality can be brought to the process from a 14 manufacturing and design standpoint. 15 It will make a better document and 16 something that we can all live with and, frankly, 17 jurisdictions can procure at a cheaper price. 18 MR. GROH: I would like to make one 19 additional point so we can get this on the 20 record. 21 All of us need to remember and 22 recognize that when HAVA was being worked out on 1 the legislative format and it was being costed 2 out or priced out, it was based on the technology 3 that existed then. 4 This technology we have now has made 5 a leap forward, which is positive. 6 But at the time that they scaled 7 this and scoped it, it was not in today's terms. 8 It was in 2002 technology terms. 9 CHAIR HILLMAN: Thank you. 10 EXECUTIVE DIRECTOR WILKEY: I think 11 this is working. 12 Just shifting gears on a subject for 13 one minute. 14 Anyone who has heard me speak over 15 the last twenty years since we started the 16 development of the 1990 standard, I hate to admit 17 I have been around even longer than that, has 18 heard me say on more than one occasion that it's 19 great that we have these standards for hardware 20 and software testing. 21 But what we need to have is the 22 second half of that. That is 50% of the battle. 1 The other 50% is the management 2 guidelines. And I have talked about that until I 3 am blue in the face, as blue as that cap is on 4 that bottle of water. 5 Hopefully during this coming year we 6 will begin a venture with the National 7 Association of the State Election Directors to 8 finally get that moving. And we have set aside 9 some money out of our budget to begin that 10 process, 11 I would like your reaction of how 12 that fits into your plans and how you can help us 13 make that document the best that it can be. 14 MR. VADURA: I think we all have to 15 have operational guidelines as part of our 16 certification documents. And some of what's in 17 there is good stuff and can contribute to this 18 process. 19 And I think any set of formal 20 management guidelines can only help us because it 21 will help to mold further the shape of the 22 system, how it works from an operational 1 standpoint, not just a technology standpoint. 2 MR. CHARLES: I think it's an 3 excellent idea. It's, as you know, a very 4 difficult task to integrate requirements that 5 address multiple types of systems, multiple 6 vendors of those types of systems, as well as the 7 different state laws and requirements. 8 So it's not an easy task, but an 9 important one. And I think we can assist in that 10 process by offering our expertise as project 11 management and people with experience across 12 state lines. So we can help refine those and 13 develop them as they go forward. 14 But it would be very helpful - as 15 with the hardware and software, it would be 16 helpful to have state specific requirements that 17 could be tested so there is a one-stop-shop for 18 approval of voting technology. 19 MR. ADLER: I think that, too, this 20 issue of managing for change, being able to 21 implement new systems and dealing with the real 22 management challenge of conducting elections with 1 temporary workers and under tough scrutiny, that 2 takes real leadership at multiple levels. 3 So anything that could enhance the 4 leadership and management tools that our election 5 directors/election officials have, is a huge, 6 huge benefit to the public. 7 MR. GROH: Election systems and 8 software, if you approach this like you did with 9 TGDC and NIST, we would be more than happy to 10 participate. Because, again, this is that 11 missing link that I explained earlier. 12 I can deliver excellent hardware. 13 Now I am beholden upon the best management 14 practices of the election administrator. And 15 when that fails, I have then failed by virtue of 16 the public exposure to that. And I don't want 17 that. 18 So we are working hard. But as you 19 have eloquently explained, Tom, when this has not 20 be broadly embraced and there is not something 21 that is there that uniformly lays this out so 22 best practices methodology procedures can be held 1 to by County Commissioners holding their election 2 administrators to this, you don't have a way to 3 measure it. 4 The Election Center I know would be 5 a wonderful partner in this, along with the 6 vendor community. 7 We know what it takes to run our 8 technology and what it takes to run and manage it 9 well. So let us at least advise in that. 10 We will come behind it and do the 11 training in that. But it's difficult to do 12 training for people who don't want to be trained. 13 I know so many people that want to 14 go out and pick up a golf club and start hitting 15 a golf ball before they have had any lessons and 16 it just doesn't work. 17 EXECUTIVE DIRECTOR WILKEY: Thank 18 you very much. 19 CHAIR HILLMAN: Thank you very much 20 to the panelists. 21 We are not going to take a break. I 22 am just going to ask the EAC staff to move 1 quickly to reset the table and escort our third 2 set of panelists up to the table. 3 Thank you very much. 4 5 6 * * * * * 7 8 9 PERSPECTIVES ON PROPOSED 10 VVPAT TESTING GUIDELINES 11 12 CHAIR HILLMAN: If we could get 13 reassembled, please. I just need to get 14 everybody. 15 I will remind everybody to make sure 16 that their cell phones, pages and other 17 electronic devices are silenced. We would 18 appreciate that. 19 Our fourth panelist is here and he 20 will be joining us, running a few minutes late. 21 And I don't want to cut into the end of the day 22 by short cutting anybody else's time for 1 presentation. 2 This is our third panel. It is 3 Perspectives on Proposed Voter Verifiable Audit 4 Trail Testing Guidelines. 5 Presenting this afternoon will be 6 Peter Kosinski, Chief State Election Official 7 from the State of New York - welcome; Harvard 8 Lomax, Clark County Registrar of Voters from 9 Nevada - welcome, thank you for traveling; Jim 10 Dickson, who will be joining us in just a few 11 minutes; and Avi Rubin, who is a Professor at 12 Johns Hopkins University in Baltimore. 13 Thank you all very much. 14 It will be, I think, okay. I know 15 that Jim Dickson will be interested in what the 16 other panelists have to say. But Mr. Kosinski, 17 if you would be so kind as to get started with 18 us. 19 And I don't believe we have any 20 written remarks. 21 MR. KOSINSKI: No. 22 CHAIR HILLMAN: I know we invited 1 you very late so this is not -- 2 MR. KOSINSKI: Quite all right. 3 CHAIR HILLMAN: I just want to make 4 sure I'm not missing anything. 5 MR. KOSINSKI: You are not missing a 6 thing. 7 CHAIR HILLMAN: But ask if you would 8 also indulge us to speak slowly so we can take 9 notes. 10 MR. KOSINSKI: I will do my best. 11 If I am speaking too quickly, let me know. 12 CHAIR HILLMAN: And I would ask that 13 people please summarize, particularly if you have 14 submitted written statements, so that we might 15 have ample time for questions. 16 Thank you. 17 MR. KOSINSKI: Thank you, 18 Commissioner. I will try to summarize, even 19 without my written statement, to move the program 20 along. 21 First of all, I'd like to thank you 22 for inviting me today and I'd like to welcome 1 everyone to New York. 2 I know my friend John Ravitz was 3 here this morning to welcome everyone. But I 4 would also like to welcome you on behalf of New 5 York and hope that you can take advantage of some 6 of what we have to offer in the City. 7 I'd also like to congratulate the 8 Commission on their wise choice of their 9 Executive Director. 10 And I certainly would like to note 11 that Tom Wilkey used to be the Director of our 12 State Board of Elections and we feel you made an 13 excellent choice in picking Tom as the Executive 14 Director. 15 And, of course, we worked closely 16 with Tom. I personally worked closely with Tom 17 for many, many years and he is excellent in his 18 job and we have a great deal of faith in his 19 ability to carry forward on this. 20 I'll just spend a couple of minutes 21 speaking a little bit from the administrative 22 standpoint, I think, on how the voter verified 1 receipt affects us, I believe, and a couple of 2 comments regarding this. 3 The State of New York I think 4 decided early on in this process that they would 5 enact and require, as part of any voting system 6 in New York that would be part of the HAVA 7 project, would have a voter verified receipt 8 attached to it. 9 Our legislation, as I am sure you 10 are aware, has just recently passed our State 11 Legislature. But in the very first drafts of the 12 legislation that was proposed in New York, way 13 back two years ago, the voter verified receipt 14 was part of those pieces of legislation in both 15 House of our State Legislature. 16 And I don't think there was a great 17 deal of discussion or certainly dissension 18 regarding that as being a component of any new 19 voting system in New York. And, in fact, the 20 final version of the legislation that was adopted 21 by our State Legislature - and while it hasn't 22 been signed by our Governor, we are hopeful that 1 it will be in the very near future - that that is 2 a part of our legislation. 3 Now, what it does not do, as you 4 might imagine, is give a great deal of detail as 5 to exactly how that is to work. And that is why 6 we are paying such close attention to these 7 guidelines as you are issuing them. 8 We think this is a very, very 9 important component, of course, of any voting 10 system. It does create some issues, though, as 11 we go through the voting process. 12 New York is a state, as I am sure 13 you know, that has had lever machines for its 14 entire voting history. And because of that, 15 paper ballots have been a very small part of 16 voting in New York State. 17 Up until the changes that we are 18 anticipating, the only paper ballots that were 19 really available in New York were absentee 20 voting, the provisional ballot voting, military 21 voting. That kind of voting was all done on 22 paper. But other than that, it was all done on 1 lever machines. 2 And at the end of an election cycle, 3 there would be very, very little paper; maybe 8% 4 of our total votes would be cast on paper. The 5 remaining 92% were cast on the lever machines. 6 And there was no paper that was at issue 7 regarding that particular election, at least as 8 far as the votes cast at the precincts. 9 With the new system, needless to 10 say, this will change. And having the voter 11 verified receipt as part of that system, it will 12 now greatly increase the amount of paper that the 13 state will be dealing with. 14 We think it's important that any 15 system that incorporates this kind of a program 16 have as a component to it that we ensure that the 17 voter actually does verify the votes that are 18 being cast. 19 And I think that, as I read through 20 the standards that are being put out, that it is 21 important to try to build in to any system that 22 would be engaged by the voter, that not only we 1 create the voter verified system, but that we 2 also try as best we can to ensure that that voter 3 actually verifies that ballot. 4 Because we realize that at the end 5 of the process, the official ballot - if there 6 becomes a dispute in an election - will be the 7 paper ballot. And so it needs to have some 8 component that, at the end of the vote, when the 9 voter - prior to leaving that voting booth - is 10 asked to verify that particular ballot, that in 11 fact that voter does that. 12 So that we have some comfort level 13 that, in fact, that voter has looked at that 14 piece of paper, has verified that piece of paper 15 as their actual votes. So if those votes are 16 contested, that we are counting the actual votes 17 of that particular voter. 18 The other issue that, of course, 19 comes up, which I think was touched on - and I 20 think Tom, in fact, touched on it as well - is 21 the management of the system. And the management 22 of the system will largely come about not only 1 during the election itself, but post election. 2 So once you have created this piece 3 of paper in New York State - we can have upwards 4 of seven million to eight million votes in a 5 particular election - when you are creating that 6 kind of paper, that there be a system in place 7 that will allow for the protection of that paper, 8 that will allow for the careful counting of that 9 paper, if that becomes necessary, and that much 10 of New York City's law, in fact, is devoted to 11 that. 12 For example, the amount of audit 13 that must be done to verify that the votes cast 14 on the electronic machine are matching up with 15 the votes to the verified pieces of paper. And 16 that is a very important component of the system, 17 that not only are you creating this piece of 18 paper, not only are you verifying this piece of 19 paper, but that you also have a system in place 20 for ensuring that these pieces of paper are 21 protected and they are also being utilized in a 22 way to make the public aware that the system is 1 being monitored and that the integrity of the 2 system is not being compromised. 3 We would certainly hope, and would 4 anticipate, that any system that would be created 5 would be accessible to the disabled. We know 6 that that's a very, very important concept. That 7 really was a lot of what drove HAVA. It was, I 8 think, a critical element in the development of 9 HAVA itself. 10 And we think it's very, very 11 important that as we create these systems, that 12 we keep that in mind. 13 And while the voter verified 14 receipt, I believe, was a component of this, that 15 really emerged following the enactment of HAVA, 16 it was not necessarily anticipated during the 17 discussions in Washington of the HAVA legislation 18 itself that since this has emerged as such an 19 important part of this particular process, that 20 we make sure that the disabled community is 21 protected; that we do not in any way compromise 22 these systems, so that they are fully accessible 1 to the disabled. 2 Other than that, I think that the 3 creation of this system is something that as we 4 work through it - it will take some time - I know 5 that that's something that has been spoken of 6 earlier here today - that as we work with these 7 systems and see how they are being utilized by 8 the people who are actually in charge of 9 utilizing them - our election inspectors, as well 10 as our voters - we can maybe fine tune these, to 11 make sure that they are accomplishing the goals 12 that we have set out for these particular 13 systems. 14 And we know that that's important, 15 that we get some opportunity to use the system 16 before we make necessarily final adjustments to 17 that system to make sure that it is being 18 adequately monitored and adequately utilized. 19 Other than that, I will just accept 20 questions. 21 CHAIR HILLMAN: Thank you very much. 22 We will hear next from Mr. Lomax. 1 And then following Mr. Lomax we will 2 hear from Jim Dickson, who is Vice President for 3 Governmental Affairs with the American 4 Association for People with Disabilities, and 5 then from Professor Rubin. 6 Mr. Lomax. 7 MR. LOMAX: I go by Larry, so I am 8 Larry Lomax from Clark County, Nevada. 9 And I was asked here because we are 10 one of the few counties that has actually used 11 the VVPAT. We used it for the 2004 primary and 12 general and then for our 2005 municipal 13 elections, primary and general elections. 14 So we do have experience. And I'd 15 like to talk about some of that experience 16 first-hand because, again, that's why you 17 actually asked me, and compare it to the 18 standards as I read them. 19 And I will point out, as I 20 understand, what's important is that the vendors 21 and the technical people understand these 22 standards, not so much that I do. But I see some 1 potential problems, anyway, that I think need to 2 be addressed. 3 First, as a user, I think it's very 4 important that we don't ever forget that the 5 standards have to be practical. And I think 6 there is some ideals that we would all like to 7 attain, but I'm not sure in some cases that the 8 ideals are realistic, or at least they may be 9 overly burdensome upon those of us that have to 10 use the voting machines and implement these 11 things. 12 Second of all, and I just point this 13 out, that Volume I, the very beginning of it 14 says, its purpose is to describe the requirements 15 for electronic components of voting systems. 16 And there's a lot of standards in 17 here that really go way beyond that into the 18 administrative procedures, especially, I saw in 19 the area of security. What they are really 20 requiring is us to perform certain administrative 21 functions that have little to do with electronic 22 components. 1 Third - and I think this has already 2 come up today in two preceding panels - I don't 3 think the standards are sufficiently clear in 4 several cases. And I will identify a few here. 5 And a good example that has already 6 come up, the gentleman from ES&S apparently 7 interprets the standard to mean that the paper 8 trail printer cannot be a spool-to-spool or reel 9 to reel printer, or it must somehow chop the 10 ballots at the end of that. 11 I talked to Alfie from Sequoia. He 12 believes just the opposite, that that is allowed 13 under these standards. 14 So these things have to be clear 15 enough so we all understand exactly what they 16 mean. 17 The standards is guidance. We don't 18 want to have to create guidance to clarify the 19 guidance because that's kind of where we are 20 after HAVA. HAVA was in there; now we are trying 21 to figure out exactly what that meant. 22 Another example I can give you is 1 there is a standard that says the paper record 2 shall be sturdy, clean and of sufficient 3 durability to be used for clarification and other 4 things. 5 Well, I had my staff read these 6 standards and they immediately came to me on that 7 issue because we use thermal paper, which is 8 sufficient for what we need this paper for. But 9 then is it really very durable? But, again, what 10 does a standard such as this mean? 11 So I don't want to find us in a 12 position later on of not being sure exactly what 13 is required and what isn't. 14 A few examples - again, looking at 15 it from the practical level, somebody who has to 16 administer these things - font size, one of the 17 first things mentioned in there. It's not 18 realistic to ask for large font size, I don't 19 believe, anyway. The bigger the font size, the 20 more paper each ballot uses up, the larger the 21 printer is going to have to be, the more 22 difficult the logistical problems are for those 1 of us that have to deal with it. And also the 2 greater the expense. 3 I think the goal would be, at least 4 from my perspective, is a printer where the paper 5 did not have to be replaced on the date of the 6 elections, that has sufficient capacity to deal 7 with all the voters who are going to use that 8 machine. And that means relatively small print. 9 We used a printer that printed in 10 10 point font; that's small. However, the secret to 11 this problem is magnification. We also had a 12 magnifier at every machine. So that somebody who 13 had difficulty reading 10 point font could hold a 14 magnifier up and it would bring it up to 20 point 15 font, which is pretty large. 16 I could honestly say we have had now 17 approximately a half a million voters vote in 18 Clark County using the paper trail printers and 19 this has not been an issue. People comment that 20 it's small, the print is small, but that's it. I 21 haven't had a single person complain or make an 22 issue of the fact that 10 point font is what we 1 are using. 2 Now, these standards require 3 3 millimeters I think, just to make it tough on me. 4 We measured that and that apparently is 12 point 5 font, which is a little bigger and that's fine. 6 But the standards also talk about, 7 you should have two choices of font there for the 8 voter. And I'm not sure in standards what 9 'should' means. I'm not sure what the vendors 10 then have to do. 11 If it says 'should', maybe that's an 12 atta boy. But it seems to me you ought to just 13 set what the minimum requirements are. 14 It looked to me that 6.3 millimeter 15 font in height is 28 point font, and that would 16 require paper this size (indicating), which is 17 out of the question. 18 There is a section in the standards 19 that talk about approve or spoil the paper record 20 and then it talks to having the ability to match 21 the spoiled paper records with the spoiled 22 electronic ballot. There are no spoiled 1 electronic ballots. 2 One of the beauties of electronic 3 voting is, the voter goes through the machine. 4 At any time he or she wants, they can change; 5 correct any mistakes they have made. 6 After they have made all their 7 selections, they get to a review screen, which 8 shows them the selections they have made and 9 shows them all the contests in which they made no 10 selection. They can immediately go back and 11 correct those. 12 There is even a warning notice that 13 comes up at the end that says, you didn't make a 14 choice at every place you could have. And they 15 have to go through all of that before they get to 16 the point where they print their paper record. 17 They print their paper record and 18 and then they review that. They can either 19 accept it or reject it. 20 There is three reasons they might 21 reject it. 22 First, we made a mistake programming 1 the machine. And there we got a disaster and 2 that would be brought to our attention very early 3 election morning and we would have big trouble. 4 But that would be a programming error on our 5 part. And if we have done our job, that's not 6 going to happen. 7 The second reason is, despite all 8 those advantages or opportunities to correct a 9 choice, they didn't catch it until they saw the 10 paper record, in which case, they can reject the 11 paper record. And remember, the electronic 12 record hasn't been recorded get. 13 They can go back and correct the 14 electronic record again, print the paper record 15 again and say okay, it's okay now, and then print 16 and finalize and cast their ballot. 17 The third reason, and this was the 18 most common in my opinion in Clark County, they 19 just want to see what happens if they hit the 20 reject button and see what the printer will do. 21 And in our case it would print VOID 22 and then they could go back and print it again. 1 As I've said, we have had about half 2 a million people vote and very, very, very few 3 people ever void a ballot. 4 Because the thing you have to 5 understand is the printed record will always 6 match what they picked on the voting machine, if 7 we have done our job right. It always matches. 8 I have a lot of people call me up 9 and ask what was the match, how many mistakes 10 were there? There are no mistakes; it always 11 matches. It always matches. 12 So most of the time when they are 13 doing this, they are just playing around. 14 Preserve voter privacy and 15 anonymity, another issue here. Reel to reel 16 printers, from a practical standpoint, in my 17 opinion, are the only reasonable way to go here. 18 In a county my size which is about, 19 we have about 750,000 active registered voters - 20 and we have a two week early voting period - an 21 election will generate in the vicinity of 8000 22 tapes. Now, these aren't the ballots chopped up; 1 these are tapes. 2 Each tape has to be maintained and 3 catalogued in a manner so that we could find any 4 tape for any given machine on any given day if we 5 needed to audit it against electronic results, 6 because in Nevada we are using an auditing 7 process for the tapes, a randomly selected 8 auditing process. 9 If you chop those tapes into little 10 five or six inch strips of ballots that will then 11 curl up because they came off a reel, and I have 12 got a half a million of those I have to track and 13 account for and catalogue in some manner so that 14 I can get them, either to audit, or if a state 15 was using them as ballots, they would have to 16 maintain those, you are really creating some 17 administrative challenges, which are going to 18 lead to problems if you chop those things up. 19 It would also, if you chop them up - 20 as far as I can imagine, anyway - it would 21 require you to manually do any recounts of those 22 things. If you are just dealing with these 1 little curlied up pieces of paper, someone is 2 going to have to hold each one up to do a manual 3 recount. 4 At least with a reel-to-reel, there 5 is a possibility, using bar codes below the 6 ballots, of an entirely independent system going 7 through and reading, doing your recount on a 8 reel-to-reel printer. 9 So I could go into this more, but I 10 really recommend that you make your printers 11 large enough to hold everybody that is going to 12 vote on one day. And that's the way we do it 13 now. 14 They can be sealed up so that 15 workers have no access to them. They attach them 16 to the machines at the beginning of the day; they 17 just take them apart, take them off at the end of 18 the day and they bring them in. But they don't 19 have to mess with that paper; you don't have to 20 worry about storing votes at the voting place. 21 And just, administratively, I think 22 that's really key in my opinion. 1 There is a couple other areas in the 2 area of anonymity which I just think reflect a 3 misunderstanding, again. 4 There's a discussion about 5 protecting the privacy of people who use 6 alternative languages. In the discussion it 7 suggests that we have at least five people who 8 vote in another language on a machine. 9 We have no idea what language they 10 are going to pick when they go into a machine and 11 we certainly are not going to ask them what 12 language they plan on voting. 13 Their identity is protected 14 administratively in the same manner we protect 15 everyone else's. 16 It always amazes me, I hear a lot of 17 concern, for instance, on a reel-to-reel printer, 18 protecting the voter's privacy. And yet no one 19 ever brings up a concern that on an election in 20 Clark County we will get about 60,000 absentee 21 ballots, and those ballots will be recived in an 22 envelope with the voter's name right on it. 1 In fact, the voter has to sign and 2 we make sure that that was the right voter. 3 And, yet, no one is worried about 4 the voter's privacy then. And they don't need to 5 be. We have administrative procedures to make 6 sure those ballots are separated and nobody knows 7 who they are. 8 But we maintain administrative 9 procedures in the county to be sure that we 10 protect the electronic voter's privacy also, by 11 the way in which we do not record the order in 12 which they sign in in a roster book. 13 We have multiple machines in our 14 polling places and the voter picks where they are 15 going go and nobody knows where they are going to 16 go or tracks where they go. So there is no way 17 to go back and reconstruct who used what machine 18 or what order they voted on. 19 I mean, it's very easy to do and 20 everybody's privacy is protected. 21 I will just point out a few 22 additional things because I don't want to take up 1 too much more time. 2 Just an administrative thing. In 3 Section 2.2.7, it does say that if a state is 4 planning to use the paper record as the official 5 ballot, then they have to be sure that a visually 6 impaired person or any other disabled person 7 would be able to review the paper record as the 8 same manner that a somewhat sighted person would 9 be able to do. 10 And on that, that same language 11 should be brought back to the VVPAT Section, 12 because right now it isn't in here. It's only up 13 in front. 14 Another example of something - I 15 guess this is almost laughable in a sense. It 16 says the voting station should be physically 17 secure from intentional damage. Good luck. 18 I don't know what that means or how 19 you would ever enforce that. But if someone is 20 going to come in with a hammer and wants to smash 21 our machine, they are going to be able to do it. 22 So I'm not sure what that standard means. 1 There is another requirement in 2 there - and these are just examples of things I 3 just don't quite understand or if they have been 4 thought through - that says we are supposed to 5 put a seal to seal the connection between the 6 printer and the voting machine. 7 All right, if I have 5000 voting 8 machines that means I got to have - remember, 9 where a printer fits in, in most cases it's just 10 a plug with the two little screws that are 11 screwed in there - you know, the little twisty 12 things like on your own computer at home - 13 Why am I sealing it? 14 First I have got to buy all those 15 seals and then when they install it they have got 16 to seal it and they have got to track it and 17 somehow log it. 18 If the printer becomes disconnected, 19 the voting machine will lock up. And that's one 20 of the standards it requires. It says the voting 21 machine has to alert you if the printer is 22 disconnected. 1 I don't see any - there is no logic, 2 that I can think of, as to why you are wasting 3 your time sealing it. It's an administrative 4 requirement that sounds good, but I'm not sure 5 there is a reason behind it that I could figure 6 out. 7 I'll wrap it up. 8 But, in general, I do think that 9 there is more clarity required in these so that 10 we really understand why these procedures have 11 been developed. 12 And I do think that much of the 13 security, as Mr. Wilkey pointed out, much of the 14 security in an election, in the privacy and 15 anonymity, is a result of administrative 16 procedures and oversight at the polling place. 17 And we have accepted that for years 18 with paper ballots and we seem to have forgotten 19 that. 20 And with the electronic, it is also 21 true. And I think that's an important thing to 22 keep in mind. Thanks. 1 CHAIR HILLMAN: Thank you very much. 2 Mr. Dickson. 3 MR. DICKSON: How much time do I 4 have? 5 CHAIR HILLMAN: Ten minutes. 6 MR. DICKSON: Thank you, Madam 7 Chair, Mr. Commissioners for inviting me to 8 come. 9 You don't have my paper testimony 10 because my paper testimony is in my suitcase 11 which is somewhere lost in Delta Airlines. 12 We will be submitting this short 13 version, plus additional comments. 14 I first want to talk about the 15 context in which Congress passed and the 16 President signed the law that says people with 17 disabilities have the right to vote. 18 CHAIR HILLMAN: Get closer to the 19 mike. 20 MR. DICKSON: I want to put into the 21 context, I want to discuss the context out of 22 which Congress passed a law and the President 1 signed, that says voters with disabilities have 2 the right to vote privately and independently. 3 Twenty-one years ago Congress passed 4 a law that said polling places should be 5 accessible to people in wheelchairs, 21 years 6 ago. 7 As best we can tell, somewhere 8 around 65% of the nation's polling places are not 9 accessible to people in wheelchairs - 21 years, 10 more than half. 11 Missouri just finished - and good 12 for Missouri - a survey to see how many of their 13 polling places were not accessible. The survey 14 reports it's 71%. There is no plan in Missouri 15 for how they are going to fix them, when they are 16 going to fix them by. What good is the report? 17 And this phenomena in Missouri 18 repeats the phenomena of 18 years ago when 19 reports were sent in to the Federal Election 20 Commission that said these are the places that 21 aren't accessible. 22 Twelve years ago Congress recognized 1 that people with disabilities are registered to 2 vote at a rate considerably below the rest of the 3 public. Congress also recognized when it passed 4 motor voter, of the National Voter Registration 5 Act, that many people with disabilities have -- 6 I'm not going to go to the Department of Motor 7 Vehicles. I have no reason. 8 Many of us can't even get there. 9 Many of us don't have the money to pay for the 10 license if we could get there. 11 We are sitting here in New York and 12 the NVRA said twelve years ago that community 13 agencies that offer services to people with 14 disabilities need to do the same thing that motor 15 vehicles do. Twelve years ago. 16 One state - thank God for Kentucky - 17 is obeying the law. One state, it's twelve 18 years. 19 We are sitting here in New York City 20 in New York State, and I think these numbers are 21 right. In New York City, about 56% - I'll get 22 the precise numbers if my memory is off - of 1 people do not have driver's licenses. 2 In upstate New York, well over 90% 3 have driver's licenses. 4 The agencies that conduct services 5 to disabilities in the city, and upstate - 6 particularly state Medicaid - are not obeying the 7 law. 8 So we come to a law now that says we 9 have the right to vote privately and 10 independently. Now, that ought to be pretty 11 here, I least I thought it was clear when 12 Congress voted on it. 13 Yesterday I was at a County Council 14 meeting in Valusha County, Florida. The 15 president of Verified Voting for Valusha, County 16 said on the record, people with disabilities have 17 75% privacy and 75% secrecy. What more do we 18 want? 19 I was not able to ask her what is 20 75% privacy. I wasn't able to ask her what is 21 75% percent of secrecy at this. 22 But that statement reflects an 1 attitude that three Federal laws have tried to 2 remedy, that we are treated like 75% of a 3 citizen. And this is going to stop. 4 And HAVA gives us the tools to stop 5 it. 6 I want to commend the Commission 7 because this version of the standards is 8 considerably improved. But there are still some 9 things that need to be done. 10 In the standard there is I think a 11 structural problem with how you have handled 12 disability that in 28 years of disability policy 13 I have never seen anything structured like this. 14 Usually the disability sections are 15 in one place. Makes it clear; it makes it fair; 16 it's easy for industry. 17 In this version of the standards, we 18 are scattered all over creation and gone. 19 And one result of that is that the 20 standard says that a voter who is blind shall 21 have a right to vote privately and independently. 22 In another section it says that a voter who is 1 paralyzed doesn't have that same right. 2 Now, I hope that that was just a 3 piece of confusion. 4 The ADA makes it crystal clear that 5 when we are talking about disabilities, we are 6 talking about all disabilities. HAVA makes it 7 crystal clear that it's the ADA that defines what 8 disability is. 9 And we cannot have standards for 10 voting systems that allow one group of disabled 11 citizens to vote privately and independently and 12 another group not to do that. 13 It's interesting to me that we just 14 had two examples of this voter verified paper 15 trail Neanderthal nonsense, and it is nonsense. 16 I understand that state legislatures 17 have passed laws. We are comfortable with what 18 Nevada has done, because in Nevada the piece of 19 paper is not the ballot. 20 We got real problems with New York 21 because the piece of paper is the ballot and I 22 know of no way that you are going to make that 1 piece of paper accessible with current 2 technology. 3 I'm going to say something that has 4 been on my mind for a long time that I have not 5 said before. 6 We have got this paper trail thing 7 because there is this small segment of the 8 computer world that's attitude is, you can't 9 trust computers to do anything. I think of them 10 as the cunieform school of computer science. 11 They cooked up an idea of a voter 12 verified paper trail without testing it, without 13 measuring it, without seeing if it will work. 14 It's supposed to give the voter verification that 15 the vote was casted right. That's great. That's 16 a good idea. 17 But let's put something in the 18 polling place that actually will do that. 19 Professor Selker (ph) at MIT just 20 did a study, a scientific test, the first one, 21 that took computers and did this fearmongering 22 nonsense of the ballot is going to change from 1 the machine to the paper. So he had MIT graduate 2 students and ungraduate students vote. 3 And the computer did eat their vote. 4 This isn't real. And the paper printed an error. 5 Okay. So the fearmongers, this is the theory 6 that the fearmongers have said is going to 7 happen. 8 7% of the voters in the test found 9 the error on the paper - 7%. When he was there. 10 And these are MIT graduate and undergraduate 11 students. 12 Now, how effective a form of 13 verification is that going to be for the public 14 at large? 15 Professor Selker also used audio 16 verification for the same voters. Same thing 17 happened, computer ate the vote, the paper 18 printed something different. 85% found the error 19 when it was audio. 20 Now I am tickled with that. I am 21 blind. I live by audio. 22 I am not going out saying we need to 1 pass a law that says let's put audio into the 2 polling place. 3 I think the Commission could do the 4 nation a service, a great service, by making 5 clear that we need standards that are based on 6 science and data, not on theory that generates 7 contracts. 8 HAVA is very clear - private and 9 independent by 2006. And the reason for 2006 10 it's 21 years and we do not have wheelchair 11 accessibility, it's 12 years and we are not being 12 registered to vote under law. 13 There are some -- If a place wants 14 to have the verified piece of paper on the side 15 of the machine, that's fine, as long as it's not 16 the ballot. 17 When it becomes the ballot, it is 18 not accessible. It does not meet HAVA's 19 criteria. It does not meet the ADA. You can't 20 spend Federal money on it. 21 And I hope the Commission will make 22 that clear. 1 Thank you very much. 2 CHAIR HILLMAN: Thank you, Mr. 3 Dickson. 4 And we will now have Dr. Rubin. 5 DR. RUBIN: Thank you, Madam Chair 6 and Commissioners. Good afternoon. 7 My name is Avi Rubin and I am a 8 computer science professor at Johns Hopkins 9 University. My area of expertise is computer 10 security. And I have been doing research in 11 industry and in academia on computer security for 12 14 years now. 13 Since 1997 I have been applying 14 computer security towards studying voting systems 15 in elections and in 2004 I became a poll worker 16 working as an election judge in Baltimore County 17 in the primary and the general election. 18 Election security is part of 19 national security. We look around the world and 20 we see just how important and sensitive the 21 ability to vote with integrity is. 22 And I believe that many states have 1 come a long way since the Presidential election 2 in improving of the security of the voting 3 process. But I also think a lot of work remains 4 to be done. 5 As somebody who specializes in 6 security, I view the world through a special lens 7 of being interested in security. And I imagine 8 the reason I'm here today to talk to you is to 9 talk about security issues, that is what I do. 10 Reading through the proposed 11 guidelines, I have several comments about them 12 and most of them are very detailed and somewhat 13 technical and I'll be putting them in writing and 14 giving them to you. 15 I only had a chance in the last few 16 days to read them very thoroughly and I have 17 about four or five pages of notes to give you on 18 those, which I want to clean up. 19 But I think several important things 20 should come out now. And one of them is a 21 distinction that I would like to make, which is, 22 to distinguish between a voter verified paper 1 audit trail, VVPAT, which is discussed in the 2 guidelines, and a voter verified paper ballot. 3 These things, when we distinguish 4 them, both of them can be viewed as a voter 5 verified paper record, which is something that 6 can be used to audit an election. 7 But paper ballots are very, very 8 different from paper audit trails. And I think 9 if that distinction were made explicit rather 10 than implicit - clearly everyone understands that 11 there is a distinction - but if it were made 12 explicit, it might make the standards read a 13 little more cleanly. 14 And, in particular, I am concerned 15 about the possibility that a voter verified paper 16 ballot, which I view as a more secure system, 17 might not meet the standard because of the 18 wording specifying a voter verified paper audit 19 trail. 20 And so I think looking at the 21 standards again with an eye towards separating 22 out those two terms would be a useful thing to 1 do. 2 I am in the class of computer 3 scientists who believe that a fully electronic 4 system that's made up of software can be made to 5 do whatever the programmer does. 6 I have conducted many experiments 7 with my graduate students; I have written a lot 8 of software myself. And I know that when you 9 program a machine, that machine will do whatever 10 you program it to, with a caveat - a big caveat - 11 that there will always be a lot of bugs that you 12 didn't intend, no matter how good a programmer 13 you are. 14 Microsoft is a very well resourced 15 company, puts a lot of effort into making 16 software and they do a great job. But they 17 release security updates once a month, if not 18 more often, because it's impossible to write 19 software without having bugs in it; that's just a 20 fact of life. 21 So, as a result, many - especially 22 computer scientists - but many other individuals 1 as well, have thought about what can we do about 2 the fact that one of the tools that we have at 3 our disposal, which is electronics and computers, 4 is not entirely reliable from a security point of 5 view. 6 We can still utilize it. Computers 7 are great. I use them every chance I get for 8 everything that I can think of. But I know their 9 limitations. 10 And one of their limitations is that 11 it's very, very difficult for an outside auditor 12 or inspector to look at a large software package 13 and know everything about what it does. And it's 14 very, very easy for someone to program some other 15 functionality. 16 I have papers and experiments that I 17 have done to demonstrate that. 18 So what do we do? We say let's come 19 up with this independent dual verification. 20 When I read this proposed guidelines 21 I loved that section. I think that's the right 22 idea, let's come up with totally independent dual 1 verification. 2 My feeling is that the best one of 3 those options is paper, but not necessarily the 4 reel-to-reel variety. 5 Let me outline some of what I view 6 as the disadvantages of the reel-to-reel VVPAT 7 over the verified paper ballots, which are 8 another form of paper. 9 I think that we have to always keep 10 in mind what the purpose of the paper was. The 11 purpose of the paper is that we don't necessarily 12 trust these machines, so we want to do manual 13 audits. 14 It's important that we not only do 15 manual audits, but random audits. 16 If we do that and they match, we can 17 have confidence that wholesale fraud is not 18 likely to have taken place. With that in mind, I 19 think that the reel-to-reel systems place a 20 tremendous burden on election officials and 21 voters who have to perform those manual counts. 22 That's because I also don't believe 1 that there should be anything that's not humanly 2 readable on the ballot. I don't like bar codes 3 and that's a controversial statement. Even 4 within the verified voting community, there is 5 not uniform agreement. 6 The simple reason that I feel that 7 way is that as a voter I am a plain and simple 8 person and I see a mark that I can't read on my 9 ballot, if I'm suspicious, how do I know that's 10 not my name? It's probably not, right. But we 11 should have a voting system that is completely 12 transparent to the average citizen. 13 If you think about an oppressive 14 government that is out to get you - which is kind 15 of the point of a democracy is to avoid that 16 being possible - then we don't want to have a 17 marking on the ballot that citizens don't know 18 what it is. 19 And if we don't put bar codes and 20 other markings on the ballots in a reel-to-reel 21 system, they are very, very difficult to count 22 manually, as was stated earlier. 1 That's why I think that what we 2 really want to have are voter verified paper 3 ballot systems where the ballot itself is 4 something like what we might find in an absentee 5 ballot, something that can be marked with a 6 machine with accessibility features, can be 7 verified with machines with accessibility 8 features, or visually or through audio. 9 But what I worry about with the 10 reel-to-reel systems is that the path of least 11 resistance for an election worker in a busy 12 hectic election is not going to be to do the 13 manual recount; they'll skimp on it. And I think 14 we should design voting systems where the path of 15 least resistance is the best possible path. 16 Not saying that poll workers won't 17 do it. I'm saying we should make it as easy as 18 possible. 19 I disagree with my fellow panelists 20 about the importance of the privacy in the 21 reel-to-reel. I don't believe in reel-to-reel 22 systems that are chopped up because of the 1 reasons that were mentioned. But I also don't 2 like the idea of the reel keeping the order, 3 preserving the order. 4 If we are suspicious of the 5 government - in the poll site where I worked, we 6 assigned people to the stations - so I think 7 voters would accept a situation, even if it 8 wasn't the right procedure, where poll workers 9 assigned them to the machines. And at that point 10 we have got a list of how those people voted in 11 that paper trail. 12 I believe that reel-to-reel VVPAT on 13 DRE's are the unfortunate product of a 14 misunderstanding of the arguments against 15 paperless voting. But I do believe with a lot of 16 effort and vigilance they can produce a more 17 secure election than without a paper trail at 18 all. 19 I would like to give some guidelines 20 for maximizing the security and auditability of 21 elections regardless of what kind of paper trials 22 that are used. 1 Random audits have to take place 2 where the tallies are compared to electronic 3 totals. These have to be like truly random. I 4 am thinking like ping pong balls coming out of a 5 machine like you have in the lottery that nobody 6 could have known in advance which sites would be 7 picked. 8 If somebody could know in advance 9 which sites will be picked for the audit, that 10 manual recount is losing most of its value. 11 I think that the manual audits 12 should be done by hand, where possible. 13 And the most important point that I 14 can make about this is that we have to have pre- 15 established, in advance, before the election, 16 procedures for what to do in a discrepancy. 17 If it's the case that we get a 18 different tally on the paper than we have in the 19 electronic machine, what to do. 20 And you have to keep in mind that it 21 may be the case that the paper ballots show 22 something that's statistically ridiculous 1 compared to the expectation and the electronic 2 tally matches the expectation. 3 If that were to happen, there has to 4 be a procedure. 5 Also it may be the case that the 6 paper matches the expectation and the election is 7 all for one candidate, which is impossible. 8 Whatever it is that we do, it has to 9 be established in advance what the procedure is 10 for handling it. 11 I think it's important to have 12 comprehensive security reviews of the system in 13 addition to guidance or standard compliance 14 testing. If the ITA's are simply looking at the 15 standards or guidelines document that you 16 produced and matching it up with that system, say 17 yes it meets those, that's not enough. I think 18 you want to have a red team or a security review 19 that's required to go in and do a risk assessment 20 of the system in the face of various different 21 kinds of attacks. And I think that that should 22 be mandated. 1 I know that I am using up a lot of 2 time and you already have my written comments. 3 So I think that I will just wrap up 4 with a couple of things. 5 Let me say that the properties that 6 I would like to see a voting system that utilizes 7 paper have are that it should maximize the 8 probability that the voters will actually verify 9 their vote. And one way to do that is to have 10 them use a marking system that produces a piece 11 of paper that they then have to scan. 12 It's important that the order of the 13 votes in the paper trail be randomized. And the 14 procedures in place for what to do in the case of 15 a discrepancy, as I just said. 16 It's important that every step of 17 the way be accessible to disabled voters, such as 18 blind voters and deaf voters. 19 And, in my opinion, the paper record 20 should be the authoritative vote. 21 Now, let me just talk about two or 22 three things that I found in the standards 1 themselves. 2 One of them that troubled me the 3 most was the requirement that there be real time 4 clocks in the machines. 5 Maybe there is a benefit to audit, 6 but I think that has to be weighed against the 7 security risks of - if I was an attacker and I 8 wanted to write malicious code for a voting 9 machine, the clock would be a great help to me. 10 It would let me know if I was in an election or 11 if I was in a test. 12 The other thing is that I read in 13 there that there should be a time stamp on every 14 record in the machine. Well, that is now an 15 electronic version of the reel-to-reel that 16 preserves the order in which people voted. And I 17 think that needs to be weighed against the 18 possible benefit of the audit. 19 There were other technical things 20 that I found that I didn't think were actually 21 enforceable, like the requirement to know what 22 the values and the registers and the various 1 variables are at setup time. I'm not sure 2 technically how that would be accomplished. 3 And I agree with the statement that 4 five people in the language minimum doesn't do 5 much if you only have three people in a precinct 6 that vote in a particular language. What do you 7 do? 8 So, I will write up all of my 9 specific comments for you that I had mostly with 10 respect to security. And I'll wrap it up with 11 that. 12 CHAIR HILLMAN: Thank you so much. 13 Okay. In keeping with our schedule 14 for today, we will begin the public comment 15 period at 4:30. However, once again, we are not 16 likely to take a break between this panel and the 17 public comment period. 18 So, Commissioners, in keeping with 19 that, we will have to stick to our ten minute 20 allocation of time so that we can do justice to 21 both this panel and those individuals who are 22 signed up for the public comment period. 1 And I am going to give myself a 2 treat and start with the line of questioning so I 3 can get all of mine in this time. 4 Professor Rubin, when I first became 5 aware of the Election Assistance Commission 6 having been organized under the Help America Vote 7 Act, I did not hear much, a little bit of 8 conversation about paper trails. And even then, 9 I'm not so sure the term V V P A T had been 10 codified. 11 I think people were sort of 12 struggling with what to call it. Different 13 things mean different things. 14 And I really appreciate in your 15 testimony the way you have sort of identified 16 three different terms for verified paper. 17 What has evolved in the past two 18 years around this discussion, this discussion 19 being verified paper, and what of that has 20 surprised you, the change in the conversation 21 from where we were at this time 2003 to where we 22 are today? 1 DR. RUBIN: You know, it's true. 2 When I started working in elections, I was 3 working with the country of Costa Rica. They 4 asked us to help them design an electronic voting 5 system. 6 And I came to the problem initially 7 from the point of view, well, that's a great 8 application, we can use our computer security 9 knowledge to do that. 10 And every time when you do security 11 work, you always do it in the face of a 12 particular adversary. So let's say the average 13 adversary does this or does this or does that. 14 And every time we tried to design a 15 system that would be fully electronic and secure, 16 we hit a wall. Because we just couldn't figure 17 out what to do if the people who programmed the 18 computers themselves were the adversaries. 19 The idea of a paper trial came along 20 because it's something that's not part of the 21 electronic system. 22 So in order to cheat, it would 1 require much, much, much more effort and what I 2 think is a retail level of fraud, than it would 3 in an electronic system. 4 What I have seen involved is I have 5 seen conferences and workshops - many of which I 6 have attended in the computer science community - 7 where people have tried to come up with solutions 8 that wouldn't require paper. And some of them 9 are very novel - the cryptographic schemes. And 10 I'm a big fan of those schemes. 11 However, along with all the other 12 requirements of accessibility, security, 13 auditability, is one of transparency. And I 14 think that's where some of these schemes fall a 15 little short. 16 What I think has evolved is the 17 understanding that we need to move with the 18 speed - not move too quickly before we understand 19 how to assess the risks that we have. 20 The reason I really enjoyed the dual 21 verification section here is that I think it's 22 exactly the right idea. That if we had -- 1 The more independent ways we have of 2 recording and storing the votes and truly 3 independent - meaning you don't take it into the 4 computer and then store it in audio and then 5 store it in hard drive somewhere - you actually 6 have to capture them separately, like the camera 7 idea does. 8 I think the more we open our minds 9 and look to ways of doing this, the more hope 10 there is to have good auditable elections. 11 I think that paper is the one that 12 we understand the best and the one that average 13 citizens understand the best. 14 If you have got a camera taking a 15 picture of the screen, of your confirmation 16 screen, one, was that camera controlled by the 17 software that's the software that you are worried 18 about or is it control by something else? 19 How do you store these; how do you 20 do the manual recounts? 21 These are all things that need to be 22 explored, not to mention the possible reflection 1 of the voter off of the screen and people's 2 paranoia about having a camera in the voting 3 booth. 4 But when I read that section, I 5 think, now people are thinking about this the 6 right way: how do you verify an audit 7 independently from the electronic tally. 8 CHAIR HILLMAN: Thank you. And I 9 have to say that you must be a pure scientist at 10 heart, because nobody else said that they have 11 enjoyed reading any section of this document. 12 So for you to say you enjoyed the 13 section, makes me feel quite giddy. Thank you 14 very much. 15 Mr. Lomax, how many times has Nevada 16 used its new system with the verified paper? 17 MR. LOMAX: We have used it for four 18 elections, it would be the primary and the 19 general election in 2004 and then we used it for 20 the primary and general municipal elections in 21 2005. 22 CHAIR HILLMAN: And you were the 1 administrator over all four elections, were you? 2 MR. LOMAX: Yes, I was. 3 CHAIR HILLMAN: Of the errors that 4 occurred, what percentage of those would be 5 assigned to human error? 6 MR. LOMAX: I'm not sure what you 7 mean by errors that occurred. 8 CHAIR HILLMAN: Whether, you know, 9 we are talking about jammed paper, something not 10 quite -- 11 MR. LOMAX: Errors related to the 12 printer, is that what we are talking about? 13 CHAIR HILLMAN: Yes, yes. 14 MR. LOMAXL There were no errors in 15 the sense that the paper record did not reflect 16 what the electronic record did, either as 17 presented to us by some voter complaining or when 18 we did our random audits. It always matches. 19 Out of the half a million 20 essentially records that were or ballots were 21 cast and then records that were created, there 22 were less than 25 instances where there was a 1 paper jam in which some portion of that paper 2 record was unreadable because of the way it got 3 caught up. We didn't lose the whole ballot but 4 we lost some portion of it. 5 So that comes out to one in every 6 20,000. 7 And that also was almost always as a 8 result of human error in the sense of the way the 9 people that threaded the paper in the printer. 10 The printer is unbelievably reliable, what we are 11 using. I mean, it works. But you do have to 12 thread the paper in there correctly. 13 So what we found when we had paper 14 problems - and they were very rare to begin 15 with - but they were generally training problems 16 because the paper hadn't been put in there 17 correctly. So it is all human error is what is 18 comes down to. 19 CHAIR HILLMAN: Okay, thank you. 20 Mr. Kosinski, I believe you said 21 that the State of New York has always used a 22 lever machine pretty much. There may be some 1 areas using a paper ballot, but punch cards or 2 optical scans have you used? 3 MR. KOSINSKI: We have used punch 4 cards and optical scans for absentee voting, 5 central count voting only. 6 CHAIR HILLMAN: So for the voters of 7 New York, an experience with either an optical 8 scan or a touch screen with verified paper would 9 be a first time experience that they would be 10 voting on something at a polling place in a 11 voting booth that would produce a piece of paper, 12 whether marking a ballot or an optical scan or 13 verification on an electronic machine; is that 14 correct? 15 MR. KOSINSKI: Yes, yes, that's 16 correct. 17 CHAIR HILLMAN: Have the draft 18 guidelines that you have seen and today's 19 discussion helped you in your thinking about what 20 New York is going to be confronting with the 21 transition? 22 And I'm not going to talk about the 1 delays or 2006 or '07, or just whatever length of 2 time it takes for New York to get to the point 3 where it replaces the lever machine. 4 MR. KOSINSKI: I think it helps. I 5 think, however, a lot of what we are faced with 6 in New York is what you have probably identified, 7 which is this is something new to our voters; it 8 is also something new to our election workers, 9 you know. 10 So that without having any 11 experience in this area, it's very hard to know 12 really how these guidelines will help us until we 13 actually start to use it and see how people react 14 to it. 15 The guidelines may assist us in 16 developing a system that will work, technically 17 work. But I think then you are faced with a 18 system that has to work with the people that you 19 are going to use it on. 20 And whether that is inspectors that 21 have to use it or it's voters that have to use it 22 for the very first time, that's obviously an 1 unknown factor. 2 And I think that the primary thing 3 that we are looking at in that area is the 4 training and the education of the voters prior to 5 the systems being introduced into our state. 6 CHAIR HILLMAN: Thank you. 7 Mr. Dickson, you will indulge me. I 8 really don't have any questions for you. 9 I had the opportunity to hear your 10 presentations and, as usual, you are clear and 11 right to the point. And there is not much 12 confusion in my mind over what you had to say. 13 Thank you. 14 Mr. Vice Chairman. 15 VICE CHAIR DeGREGORIO: Thank you, 16 Madam Chair. I think we have a very diverse 17 group of folks in front of us. We have Mr. 18 Kosinski, who is going to be new to VVPAT and his 19 state hasn't used it yet; Larry Lomax has had 20 half a million voters use it; we have Mr. Rubin 21 who advocates for it; and Mr. Dickson who is 22 opposed to it. 1 So we have had a very diverse group 2 of opinions. 3 Let me ask Mr. Lomax, though, some 4 technical questions about this and about your 5 system. 6 First let's talk about the font 7 size, because you suggest that a 10 point font 8 size - using Times New Roman as the font, I 9 assume - your statement here is in 12, I'm pretty 10 sure it's 12. If I have to look at a 10 I have 11 to make sure that I've got these on and the 12 bottom part of my glasses on. It still seems 13 pretty small to me. 14 Is there any evidence, when you have 15 such a small font size, that people will just 16 ignore the VVPAT and not even look at it because 17 it's too small, whereas if the font size was 18 larger, if it was 15 or 20, then they might even 19 look at this and pay attention to it and compare, 20 whereas if it's a small font size, they might -- 21 Is there any evidence or anything 22 that you have looked at -- 1 MR. LOMAX: I have no data that 2 would reflect that. 3 I do know that in Clark County it 4 had been advertised, so to speak, that the voters 5 did not look at the paper trail. 6 There was a CD made - after we 7 caught somebody filming up there for a couple of 8 hours and they got thrown out - they did their 9 own analysis and that was their determination. 10 But I would point out that in Clark 11 County we have been doing DRE voting since 1996. 12 So adding the paper trail to the touch screen 13 machine was not nearly as revolutionary as it's 14 going to be, for instance, in New York, when they 15 go from a paper system to this whole setup, the 16 electronic voting and the printer. 17 So our voters were very comfortable 18 with the electronic machines to begin with. This 19 paper trail issue was not a big deal in Clark 20 County. 21 So I think a lot of the voters, in 22 fact, in a lot of polling places we had a lot of 1 our older machines without the printer and our 2 new ones with the printer, the majority of voters 3 chose to use the old machine, just because they 4 were comfortable with that and they didn't have 5 the printer. 6 They had their choice which machine 7 they wanted to use. 8 So I think we had a comfort level 9 that makes us, perhaps, a bad example to relate 10 to how many people look at the printer. 11 I don't doubt, though, that the 12 smaller font might discourage some people from 13 looking. 14 VICE CHAIR DeGREGORIO: Because I 15 was in your warehouse in February '04 and saw 16 your two different systems. And what you just 17 said here confirmed what I thought, you have two 18 systems there and voters were given a choice, if 19 they wanted to use the one with the verification 20 of the paper and ones that did not have that. 21 What percent used the verification? 22 MR. LOMAX: As I said before, we do 1 early voting for two weeks. During the early 2 voting period we only use the machines that have 3 the paper trail printer. 4 So during that time period 271,000 5 people voted using the printer. 6 On Election Day we had at least two 7 of the paper trail machines in every polling 8 places. And in some polling places we had two of 9 those and 20 of the others. 10 But on Election Day, about 70,000 11 people chose to use the paper trail machine and 12 that makes about 150,000 that chose to use the 13 other one. 14 VICE CHAIR DeGREGORIO: You allow 15 observers in your polling stations, party 16 observers or somebody who can sit there all day? 17 MR. LOMAX: In the Presidential 18 election, we had up to 14 poll watchers from the 19 various parties. 20 VICE CHAIR DeGREGORIO: So someone 21 could come in and make a record of every single 22 person who voted and the time they voted and 1 the -- 2 MR. LOMAX: You can say that's true, 3 but that's absolutely not true. 4 First of all, if you come into one 5 of our polling places, you don't know who the 6 people are as they go up. 7 We vote multiple precincts in a 8 polling place. Just because you are standing in 9 there doesn't mean you have the foggiest idea who 10 all these people are going off to vote. They are 11 going off to all the different voting machines on 12 their own. 13 VICE CHAIR DeGREGORIO: Well, some 14 states have the requirement that the political 15 party observers or the candidate observers can 16 sit behind the poll workers and actually have to 17 be told the name of the person that is coming to 18 vote. 19 It may not be true in your state, 20 but some states have that requirement. 21 I just wanted to see if Nevada had 22 that requirement. 1 Let make get to a couple of things. 2 You make a very good suggestion here 3 on this reel-to-reel issue where you have two in 4 one polling place and people go use either one 5 and there is no way to track the sequence of 6 voting and who voted. 7 And that is, you know, a reasonable 8 solution to allow a reel-to-reel that prevents 9 anyone from - and allows, I think, the privacy of 10 the voter - that you can't determine someone's 11 vote by looking at these pieces of paper 12 afterwards because you have two and you are using 13 both of them. 14 It seems like a very reasonable 15 accommodation. 16 But if you did have one, just one at 17 a polling station, and it was reel-to-reel and it 18 was a state that required or allowed some 19 observer or someone to sit there and take a name 20 of the person and then in some kind of a 21 discovery in a recount or election contest, we 22 can actually go in then and determine who voted 1 for whom. 2 MR. LOMAX: I would never argue 3 that. 4 VICE CHAIR DeGREGORIO: That could 5 be problematic. 6 MR. LOMAX: I mean, you have to have 7 appropriate administrative procedures. 8 As I said before with the absentee 9 ballots, the ballot comes with the voter's name 10 on it. 11 You have to have procedures that you 12 break it apart so you can't track it. 13 You can overcome any of these 14 systems, depending on what you want to do. 15 VICE CHAIR DeGREGORIO: Thank you. 16 Let me ask the Professor here, who 17 does not believe in a reel-to-reel system - has 18 made that clear - but believes in paper. 19 And I appreciate your discussion 20 here about trying to make a distinction between 21 the VVPAT and what is a ballot and what is a 22 non-ballot, and that is helpful. 1 Are you suggesting here, Dr. Rubin, 2 though, an optical scan type of system would fit 3 the definition of the VVPB, the voter verified 4 paper ballot, marked by the voter, so that 5 definition would - an optical scan system would 6 qualify to that? 7 DR. RUBIN: That's right. However, 8 it doesn't mean that the optical scan ballot is 9 necessarily filled out by hand. 10 VICE CHAIR DeGREGORIO: Right. 11 DR. RUBIN: But the nice thing is 12 that whatever machine is used isn't part of the 13 trusted computing base of the system, so it 14 doesn't have to be trusted. 15 VICE CHAIR DeGREGORIO: Do you have 16 any comments on font sizes that's in these 17 guidelines when it comes to the paper trail? 18 Have you looked at this issue? Have your 19 students done any studies or anything like that? 20 DR. RUBIN: No, we have not looked 21 at that issue and the human factors. 22 VICE CHAIR DeGREGORIO: Jim, I want 1 to second our Chair's comments. You are a very 2 articulate advocate, spokesperson. 3 MR. DICKSON: I want to, if I 4 could -- 5 CHAIR HILLMAN: He wants us to ask 6 him a question, so, Jim, please. 7 VICE CHAIR DeGREGORIO: I am going 8 to allow you the opportunity to comment on that. 9 MR. DICKSON: I really appreciate 10 the effort that Larry and his colleague took with 11 the magnifying glass and I'm sure that was 12 helpful for some. 13 But, again, we have got to really 14 look at fact and reality. The reality is that 15 sight loss and those that use/need magnification, 16 there isn't a single magnifying glass that's 17 going to work for everybody. There are too many 18 variables about the light, where the damage in 19 the eye is. 20 So it is great to have it there, but 21 we need to really understand that it will fit a 22 few, but it's not going to fit most. 1 And in terms of the size of the 2 population that we are talking about, for every 3 one person who you can tell is blind because we 4 use dogs or canes or are essentially totally 5 blind, there are nine whose vision is such that 6 they can't read print. So it's a lot of people. 7 CHAIR HILLMAN: Thank you. 8 Commissioner Martinez. 9 COMMISSIONER MARTINEZ: Thank you, 10 Madam Chair. 11 I will be brief in my questions. 12 I guess I want to start by 13 reiterating, at least for the purpose of 14 discussion, just what this panel is not. 15 This panel is not a discussion about 16 the security or lack thereof of DRE systems. I 17 think that discussion has to continue. I have 18 certainly had discussions with Professor Avi 19 Rubin about continuing to engage - who is 20 actually, you have been, I think, very accessible 21 to the EAC in talking about that particular 22 issue. 1 But, again, we are not here to 2 discuss the security or lack thereof of 3 electronic voting systems. 4 We are not here, either, to discuss 5 the pro's and con's as to whether VVPAT is an 6 appropriate method to increase voter confidence 7 or to increase security. 8 The position that the EAC has taken 9 is that that is a decision left solely to the 10 discretion of state and local governments. 11 What we felt was appropriate was, 12 given the fact that at least 16 jurisdictions - 13 and I guess I am learning, Mr. Kosinski, that in 14 this legislation that is pending here in New 15 York - that perhaps 17 states now, will be 16 requiring - either through legislative action or 17 administrative rule - that those states use VVPAT 18 by a date certain, most of which are 1/1/06. 19 Though some states, like Colorado, 20 which recently passed a VVPAT requirement, says 21 they don't have to implement it until 1/1/08. 22 And so, from my perspective, that is 1 a decision appropriately left to state and local 2 jurisdictions. And they, in turn, will have to 3 be held accountable to the citizens that they 4 serve. 5 Having said that, having stated from 6 my perspective what this panel is not, I want to 7 start, if I could, Mr. Lomax, with just a general 8 perspective. 9 You are, among other things that you 10 do, you are obviously the administrator there in, 11 is it Clark County? 12 MR. LOMAX: Yes. 13 COMMISSIONER MARTINEZ: -- in 14 Nevada, of the elections. 15 And I wonder if you can give us the 16 perspective of before you had VVPAT and where you 17 stand now, in terms of just a personal 18 perspective about what purpose it serves for you 19 as a local election administrator. 20 MR. LOMAX: Prior to having the 21 VVPAT, every election there was a small - as it 22 is nationally - small but very vocal minority 1 that complained to us, sent me e-mails, telephone 2 calls about the lack or of a way to verify 3 electronic voting. 4 None of this occurred prior to 5 Florida, but a lot of it occurred after Florida. 6 Since we have the VVPAT, all of that 7 has gone away. I didn't have any complaints in 8 the last election. 9 So I would still suggest that the 10 great majority of people in Clark County really 11 didn't care one way or the other. 12 But for that small minority of 13 people who passionately did care, it's made a big 14 difference. They have been very comfortable and 15 we really had no issues related to voting 16 machines in the last election. 17 COMMISSIONER MARTINEZ: Is there a 18 state law in Nevada, a requirement that you audit 19 a certain number of the paper records to the 20 electronic ballot, because that's what it is, I 21 assume? 22 MR. LOMAX: The Secretary of State 1 has put out guidelines in what we call the 2 Administrative Code. And what he requires, that 3 1% of the machines after the election are 4 randomly selected. And what we do then is we had 5 teams of people manually recount the paper 6 records and we match those results against 7 electronically recorded results from those 8 selected machines. 9 COMMISSIONER MARTINEZ: Does the 10 paper that is produced with respect to the VVPAT 11 have any legal standing? 12 I know it's not the official ballot, 13 but is it the official record for recount 14 purposes in Nevada? 15 MR. LOMAX: No, it is not. Right 16 now our recount procedures are all based upon the 17 electronic results. 18 COMMISSIONER MARTINEZ: Professor 19 Rubin, the Vice Chair alluded to this a few 20 minutes ago, but Mr. Lomax has talked about some 21 of the administrative protocols that they 22 implement to ensure that the reel-to-reel aspect 1 of this VVPAT technology that they use in Clark 2 County accounts for trying to maintain the 3 privacy of the voter with the use of at least two 4 machines. And I guess they don't have a sign-in 5 requirement; they don't keep a record of exactly 6 where each voter signed in, et cetera. 7 Does that help to alleviate, from 8 your perspective, some of your concerns about 9 reel-to-reel technology? 10 DR. RUBIN: To some degree. But I'm 11 still concerned about the possibility, not of 12 being able to match every voter with their vote, 13 but of being able to target a particular person. 14 If you are an observer and you want 15 to know how did that person vote and you get to 16 participate in a recount and you watch when they 17 came in and how many people went before them, you 18 will be able to tell how they voted. 19 I think the privacy you want is 20 having a sequential record of the order of which 21 people voted is a bad idea, because I think we 22 can have voting systems that are just as good in 1 every possible way that don't do that. 2 COMMISSIONER MARTINEZ: Thank you. 3 Mr. Dickson, I do have a question 4 for you, so you can have another shot at giving 5 us what I think is a very eloquent perspective. 6 I guess, from my - and, Jim, you and 7 I have had many discussions over the course of my 8 short time on this Commission, and certainly over 9 the past few months - and I think that the 10 dilemma for me, obviously, is that I certainly do 11 recognize the language that has been written very 12 intentionally by Congress. 13 And we have to keep in mind that 14 HAVA was passed overwhelmingly in both Houses of 15 Congress and signed quickly by President Bush 16 back on October 29th of 2002. 17 There is language, clearly, that 18 represents significant civil rights strides with 19 persons with disabilities when it comes to being 20 able to vote. 21 So I certainly recognize and give 22 due credit to the eloquence that you bring to the 1 table at all times in reminding us of that 2 particular language. 3 I guess my question is also, it 4 plays along those lines. And that is, as 5 somebody who has to help determine an 6 interpretation of an important Federal statute, 7 like the Help America Vote Act, is it 8 appropriate, from your perspective, for us to 9 temper our decision or be influenced by the 10 technology, the capabilities that currently 11 exist, from a technology perspective? 12 MR. DICKSON: I think you have to. 13 We have a deadline that is the law and it was put 14 there consciously. We can't-- 15 There is a notion abroad in the land 16 that, well, we can just wait until something 17 better comes. 18 And I think it would be very helpful 19 if the Commission would say, very clearly, you 20 know, you got to buy it by the first of the year, 21 it has to be accessible. You can't wait for 22 something that's better. 1 I also think that there is an 2 underlying assumption by those who say we got 3 something better coming down the road, and that 4 is that there is going to be money to buy that 5 something better. 6 Well, I have been working for 25 7 years to try to get government to put more money 8 into elections. We have got one pot of money. I 9 hope there will be additional money coming. I 10 will certainly do - and the disability community 11 will do - everything we can to get the additional 12 money to election officials. 13 But you got to buy what's available 14 now is absolutely clear. And we can't be waiting 15 for something in the future. 16 David Dill (ph), he is the guy who 17 really lifted the paper trail up, big time. When 18 I first talked to him 2 1/2, a little, about, 19 over 2 1/2 years ago, he said that right around 20 the corner there was going to be something that 21 was going to make paper accessible. 2 1/2 years 22 have gone by and it's not here. 1 I have heard - and this is true for 2 most people with disabilities - I have heard 3 medical scientists say to me, right around the 4 corner is going to be a cure for blindness. I 5 have been listening to that for 35 years. 6 I not only can't see what's right 7 around the corner, I still can't see the corner. 8 COMMISSIONER MARTINEZ: I appreciate 9 that. 10 I guess two additional comments from 11 my perspective. 12 The first is that you had, Mr. 13 Dickson, talked about in your testimony the idea 14 of if VVPAT is required by jurisdiction, that, 15 for example, you complimented the Nevada 16 situation because it's not the official ballot. 17 And I do want to point out that we 18 did add language in the VVPAT requirement in the 19 proposed VVSG - and I'm sorry for all the 20 acronyms, but I think we by now are following 21 along - that says that in a jurisdiction where it 22 is being used as the official ballot, that full 1 accessibility shall be required. 2 And I am sure you are probably 3 familiar with that language. 4 MR. DICKSON: Yes. 5 COMMISSIONER MARTINEZ: And I guess, 6 going back to the reason that I asked the 7 question about technology, because there is 8 this - I don't know what the right word is - this 9 dichotomy, I guess, if you will, between - in the 10 VVSG - between - there is, the way it was laid 11 out, the various disabilities are dealt with in 12 different sections, and I think you pointed that 13 out that it was in different places and not all 14 in one particular spot. 15 I think that, from my perspective, 16 what I have learned in going through the process 17 and in sitting through the various meetings that 18 occurred by the TGDC - and I know that you were - 19 if not physically, then certainly via the 20 Internet technology - a part of many of those 21 sessions - a realization, if you will, from the 22 TGDC - from the Technical Guidelines Development 1 Committee - and the various experts that they 2 drew from at NIST and other places, that 3 technology, perhaps, is not where it ought to be 4 when it comes to ensuring that we can meet what 5 is, I think, very important language of privacy 6 and independence, for all persons with 7 disabilities. 8 And, therefore, there was this 9 difference in language between those that had 10 visual impairment versus those that had other 11 disabilities, like limited dexterity. 12 I just, again, I just wonder from 13 your perspective the appropriateness of that type 14 of an analysis and whether the availability of 15 technology is an appropriate consideration, as we 16 make those types of tough choices. 17 MR. DICKSON: I think it's 18 inappropriate to have one standard for one type 19 of disability and another standard for another 20 type of disability, particularly when we are 21 talking about sizable populations. 22 The dexterity issue isn't just for 1 people who are quadriplegic. We are talking 2 about voters who have Parkinsons and have such 3 tremor that they can't handle the paper - take it 4 out or put it into a slot. 5 We are talking about people who have 6 severe hand/eye coordination problems. 7 There are many people who have 8 perfectly good use of their arm, but their hand, 9 they can't pick up the paper. So I think it's 10 got to be consistent. 11 COMMISSIONER MARTINEZ: I appreciate 12 that. 13 Madam Chair, I apologize. I have 14 gone over my time. Thank you. 15 CHAIR HILLMAN: Thank you very much. 16 We are just about 4:30. But, Mr. 17 Wilkey, you had a quick question. 18 EXECUTIVE DIRECTOR WILKEY: No. I 19 am going to make this easy because two great 20 minds, the two questions and comments that I had, 21 Commissioner Martinez adequately covered both of 22 them. 1 I'll just comment that as a HAVA 2 baby, so to speak, as someone who really believes 3 in what this law says and did, I understand, as 4 Commissioner Martinez so eloquently said, that 5 one of the most important aspects of this law is 6 to ensure that every voter gets to vote 7 independent and privately. 8 And we need to try to reach that 9 goal. That's it. 10 CHAIR HILLMAN: Thank you. 11 12 * * * * * 13 14 PUBLIC COMMENT PERIOD 15 16 CHAIR HILLMAN: We have nine people 17 who have signed up for the public comment section 18 of our hearing. 19 I'll read those names and ask the 20 individuals to be prepared to come to the 21 microphone in the order that I'll read the name 22 and please let us know if you are here when I 1 read your name. 2 Bobbie Ann Brinegar? 3 MS. BRENEGAR: Here. 4 CHAIR HILLMAN: She is here, great. 5 Dan McCrea? 6 MR. McCREA: I'm here. 7 CHAIR HILLMAN: Rebecca Mercuri? 8 MS. MERCURI: Here. 9 CHAIR HILLMAN: Sanford Morganstein? 10 MR. MORGANSTEIN: Here. 11 CHAIR HILLMAN: Fernando Morales? 12 Mr. Morales, no? Okay. 13 If staff would let us know if he 14 comes in at any point during this 30 minutes, to 15 just let us know and we will make sure that he 16 gets added on. 17 Teresa Hommel? 18 MS. HOMMEL: Here. 19 CHAIR HILLMAN: Martha Mahoney? 20 MS. MAHONEY: Here. 21 CHAIR HILLMAN: And Sharon Shapiro. 22 MS. SHAPIRO: I'm here. 1 (inaudible) 2 CHAIR HILLMAN: Just one moment, 3 ma'am, and then we will call on you. I'm sorry, 4 I couldn't hear the request. 5 MR. CAMPBELL: Am I on the list? 6 CHAIR HILLMAN: I'm sorry. You are 7 on the list, Edmund Campbell. 8 We will start with -- 9 MS. SHAPIRO: I have to catch a bus. 10 So I don't have to go first, but I need to go 11 earlier. 12 CHAIR HILLMAN: She has a 13 transportation issue, so why don't we do this, 14 why don't we let you go first so that you don't 15 miss your transportation. 16 MS. SHAPIRO: Thank you. 17 Okay. What I am going to do is I am 18 going to have my colleague read my testimony so 19 that time is used wisely. 20 I do have some comments I would like 21 to make after it is read. 22 CHAIR HILLMAN: Okay. Who is 1 reading Ms. Shapiro's -- 2 MS. KAL HERSHAY (ph): I am. I am 3 Marjorie Kal Hershay and I am part of the 4 coalition of New York City/New York State HAVA 5 Coalition. 6 It is a group of organizations that 7 have been meeting over the last 2 1/2, 3 years. 8 And, thank you. 9 CHAIR HILLMAN: I'm sorry to 10 interrupt. But I did forget to state the rule 11 that each person has up to three minutes for the 12 presentation. 13 And so you will have to indulge me, 14 because of the number of people testifying, that 15 I will to keep strictly to the time frame. 16 MS. KAL HERSHAY: Sure. And you 17 have written copies of the testimony. 18 So let me just flick down through 19 the major points. 20 MS. SHAPIRO: Okay. I think I'll do 21 it since - okay. 22 My name is Sharon Shapiro and I am 1 disabled. 2 If I could go to the table, that 3 would be easier for me. 4 CHAIR HILLMAN: Sure. 5 Hello. 6 MS. SHAPIRO: I am the Director of 7 Advocacy of the Center for Independence of the 8 Disabled. 9 We have been involved in this issue 10 ever since and before the inception of HAVA. 11 We have met with New York State 12 officials. We have testified. We have written 13 papers about what people with disabilities need. 14 This morning, when the vendors were 15 giving their or saying dual comments, which I 16 found very interesting, they were saying that the 17 guidelines came or are too much too late. And 18 then later they said, oh, but we are prepared. 19 And I found that kind of 20 incongruent, because how can the guidelines be 21 too much too late and yet they are prepared? So 22 that I found somewhat incongurent. 1 I want to continue, though, by 2 saying that we are disappointed in New York State 3 in the joint bill that was presented to the 4 Governor. 5 The bill does not ensure large 6 print; it keeps the Full Faith Ballot Law, which 7 will restrict the machines that we can choose 8 from. 9 People with reach disabilities and 10 people who are using vision and hearing will have 11 trouble using the full faith ballot. 12 I also want to comment that in my 13 testimony, as you will see, I originally wrote 14 that I was disappointed that the Commission 15 didn't make access guidelines for voter 16 verification. 17 And the reason why I did that was 18 the access guidelines were in a different section 19 of the guidelines. So when I was looking up 20 voter verification access, I was looking under 21 the disability access regs and I completely 22 missed the voter verification. 1 So on one hand I'm very happy the 2 Commission has included it. On the other hand, I 3 recommend that you reference it in both places. 4 I just want to ask that New York 5 State and the Commission press upon New York 6 State to make these guidelines madatory. I fear 7 that since New York State will be having their 8 localities pick and choose machines, that they 9 won't go by the guidelines. 10 These guidelines need to be 11 mandatory now. 12 And the only other issue is the 13 voter paper trail. Right now there aren't any 14 extensive voter verified paper trails that have 15 been certified as being accessible. 16 What are we going to do? 17 And the state has, we have to get 18 machines in and yet we need access to paper 19 trails. We need to come up with a solution. 20 And I have more items, but I don't 21 want to -- I wanted to focus on the issues that 22 you are addressing today. 1 CHAIR HILLMAN: Thank you so much. 2 We appreciate your comments. 3 MS. SHAPIRO: I have one more 4 comment. 5 My disability is cerebral palsy. 6 The dexterity issue that Mr. Dickson was 7 referring to would apply to someone like myself. 8 We need to be able to independently verify our 9 ballots. 10 CHAIR HILLMAN: Thank you very much. 11 Appreciate it. 12 Ms. Brinegar. I don't think I 13 pronounced your last name correctly. 14 MS. BRINEGAR: Brinegar. 15 CHAIR HILLMAN: And could we, just 16 to save on time, ask people to use the stand up 17 microphones. Stay where you are, but we will 18 just ask people -- 19 And, in fact, if you know you are 20 next, you can go up to the microphone to wait 21 your turn. It will help us move along. 22 MS. BRINEGAR: I am Bobbie Brinegar 1 speaking for the Verified Voting Foundation. 2 And I am pleased to have the 3 opportunity to provide our initial feedback on 4 the Voluntary Voting System Guidelines released 5 for comment this past Friday. 6 Our staff has been reviewing the 7 hundreds of pages of guidelines and we will 8 provide thorough written comments prior to the 90 9 day deadline. 10 Verified Voting Foundation is a 11 non-partisan, non-profit organization championing 12 reliable and publicly verifiable election 13 systems, including voter verified paper ballots 14 which enable meaningful recounts and effective 15 random manual audits. 16 We are concerned about public 17 visibility into the acquisition, regulation and 18 operation of voting systems. For that reason, we 19 petitioned the EAC to upgrade the guidelines with 20 measures that will make the voting system 21 certification process transparent and encourage 22 public participation. 1 The guidelines should specify that a 2 broad spectrum of representatives from non-profit 3 election protection organizations, universities 4 and the public be able to review and provide 5 comment on test plans for voting systems. 6 The test plans and results should be 7 made public before certification of voting 8 systems. And comments from test observers should 9 be published along with the official test 10 reports. 11 As a best practice, we also 12 recommend that any jurisdiction considering 13 acquisition of a voting system establish a 14 citizens' advisory board for that purpose, as 15 already occurs in some jurisdictions across the 16 nation. 17 Voting technology vendors should not 18 be able to elect a test lab and/or pay that lab 19 for testing, as it creates a clear conflict of 20 interest. 21 The test lab is not then independent 22 and has little incentive to find problems when 1 paid by the very voting technology vendors 2 seeking its approval for certification. 3 And test labs should conduct testing 4 using human beings in addition to automated 5 testing. 6 Additionally, Verified Voting 7 proposes a ban on wireless communication 8 capability in voting machines. 9 Our organization strongly suggests a 10 more complete treatment of ballot marking devices 11 within the guidelines, because many jurisdictions 12 are adopting this voting technology as the means 13 for providing the important disabled 14 accessibility for a verifiable precinct count 15 optical scan system based on paper ballots. 16 Finally, the Verified Voting 17 Foundation recommends the use of an accessible 18 voter verified paper ballot, which enables all 19 voters - including all disabled voters - to check 20 that their votes are recorded as intended and 21 makes it possible for election official to 22 perform meaningful recounts and effective random 1 manual audits. 2 Although clearly not the only 3 election reform needed, at this time electronic 4 voting systems cannot be secure, reliable and 5 verfiable without paper ballots. 6 Thank you for your dedication to 7 establishing superior voting systems guidelines. 8 CHAIR HILLMAN: Thank you. 9 Mr. McCrea. 10 MR. McCREA: Good afternoon. My 11 name is Dan McCrea. I am a Government Relations 12 Chair for the Miami-Dade Election Reform 13 Coalition. We are a non-partisan all volunteer 14 group advocating for all voters. 15 In my other life I am a general 16 contractor and a former elected City 17 Commissioner. 18 Much of what you are dealing with 19 today and the testimony that you have heard today 20 is hi tech. Let me return us to some low tech 21 best practice and suggest that it is essential 22 that the guidelines include it. 1 Our coalition undertook a study of 2 poll closing procedures in Miami-Dade County in 3 the November 2004 general election. 4 I have provided copies of the report 5 that we produced to your staff and we will 6 subsequently provide further written testimony to 7 you on it. 8 In that report we were -- In those 9 observations, we were particularly interested in 10 ballot accounting and electronic security. 11 Ballot accounting, when using 12 electric ballots, is particularly important 13 because there is not a fixed number of ballots 14 being issued. 15 The first step in establishing a 16 valid vote count in such elections is to know how 17 many ballots you are looking for, you are trying 18 to count. 19 Our study looked at poll workers 20 conduct of ballot accounting that is accurately 21 counting the number of voters who signed in to 22 vote and comparing that number, the total number, 1 of ballots counted in the precinct. We exposed 2 serious weaknesses. 3 77% of precincts did not produce a 4 match. We found that some of the 23% that did 5 produce a match forced that match. It is late at 6 night, poll workers are tired, and they would 7 enter a number that matched the ballot count, 8 even if it was not the signature count. 9 And 14% did not even count the 10 signatures, to make that important ballot 11 accounting procedure. 12 Even small discrepancies matter and 13 can hide larger problems. 14 In one egregious case, in Precinct 15 816, the number of voters and the numbers of 16 ballots cast produced a discrepancy of only two. 17 Had it been properly investigated, 18 they would have found that votes from a 19 malfunctioning machine were uploaded into the 20 certified canvass three times, causing 171 errant 21 votes in a single precinct to enter the canvass. 22 I think Professor Mahoney from the 1 University of Miamai is going to talk more about 2 that in a minute. 3 This amounted to electronic ballot 4 stuffing. 5 And while this case appeared to be 6 unintentional, what can happen unintentionally 7 can happen intentionally. 8 The ballot accounting problems that 9 were revealed are not unique to Miami or Florida. 10 We urge you to review our report for 11 its application nationally and to ensure the 12 guidelines outline sound ballot accounting 13 procedures and point out that they are 14 particularly important where electronic ballots 15 are used. Thank you. 16 CHAIR HILLMAN: Thank you very much. 17 Ms. Mercuri. 18 MS. MERCURI: I'd like to thank the 19 members of the United States Election Assistance 20 Commission for the opportunity to address you 21 today on the proposed voluntary voting system 22 guidelines, VVSG, recently posted on your web 1 site. 2 For those of you who do not know 3 me - I know many of you do already - I am a 4 computer scientist and engineer who has been 5 researching, writing, testifying and commenting 6 on voting systems for the last 16 years. 7 Much of this material is available 8 on my web sit at www.notablesoftware.com. 9 Currently, I am a Fellow of the 10 Radcliffe Institute for Advanced Study at Harvard 11 University, where my research has focused on the 12 concepts of transparency and trust and 13 computational systems. 14 I have been an active member of the 15 Institute of Electrical and Electronics Engineers 16 (IEEE) P1583 working group on voting system 17 standards, a portion of whose work product is 18 embodied in this VVSG that is being now 19 considered for adoption by the EAC. 20 I can tell you first-hand that it is 21 fairly certain that the VVSG document you have 22 release would not be accepted by the IEEE 1 Standards Association, nor its P1583 working 2 group, should it be sent through that channel. 3 And the reason that I can make such 4 a strong statement in this regard is that I can 5 point to many specific areas of grave concern in 6 your document that have been the focus of 7 hundreds of hours of debate among the engineers, 8 computer scientists, vendors, academics, 9 usability experts and election officials who 10 comprise the P1583 working group and its larger 11 overseeing standards association balloting group 12 known as SCC38. 13 These areas directly affect the 14 fundamental aspects of accuracy, integrity, 15 security and reliability of voting systems. 16 Some of this material represents 17 legacy text that was propagated, despite 18 strenuous objections, from the Federal Election 19 Commission's earlier 1990 and 2002 voting systems 20 standards. 21 That this wording and its 22 ramifications have continued to persist through 1 to this stage is, in large part, due to the fact 2 that the dissenting positions of esteemed 3 scientists numerous enough to have caused the 4 IEEE's proposed standard to fail to gain the 5 majority approval necessary for its adoption - 6 it's not just a few of us; there were quite a few 7 scientists and others as well - have not be 8 allowed, this position has not been allowed to be 9 represented in the overwhelming one-sided 10 appointments to the EAC's TGDC. 11 I strenuously object particularly to 12 the sections in the proposed VVSG that include 13 the allowance for the use of wireless and public 14 network technology that provides open 15 vulnerability of U.S. elections to terrorists 16 around the planet, the blanket exemption from 17 examination for commercial off-the-shelf hot 18 (COTS) products, many of which are known to have 19 exploitable security flaws, the inadequate Mean 20 Time Between Failures Rating that translate to a 21 9% Election Day equipment malfunction rate, the 22 lack of mandated security controls that are 1 commonly applied to computer systems used in 2 aviation, banking, health care and even lottery 3 and casino gaming, and the lack of any 4 requirement that voting systems be transparent to 5 and independently auditable by the election 6 officials who must certify the results that are 7 produced. 8 Objections to many of these items 9 were provided in the comment round held by NIST 10 during their work period on this document. But 11 these serious and very real concerns appear to 12 have been grossly ignored. 13 As a footnote, I would like to just 14 add that the HAVA money will all be spent before 15 any of the equipment pertinent to these 16 guidelines will be implemented or available for 17 sale using the EAC money. 18 This document, then, is the legacy 19 of the EAC because it will be used to influence 20 future equipment that will be purchased in later 21 years. 22 Because of the limited time 1 constraints for today's presentation, I will 2 defer my comments on these highly flawed sections 3 of the VVSG to a very detailed memorandum that I 4 plan to deliver within the public comment period. 5 I hope that you will review my 6 submission closely and I urge to you communicate 7 with me in order to gain a more holistic 8 perspective of the nature of the comments that I 9 will be providing, along with the related 10 concerns. Thank you. 11 CHAIR HILLMAN: Thank you. 12 Mr. Morganstein. 13 MR. MORGANSTEIN: Thank you. Good 14 afternoon. I'm Sandy Morganstein, President and 15 founder of Populex Corporation, a vendor that has 16 a 2002 certified election system that produces a 17 voter verified paper ballot, which we think is 18 accessible. 19 It's certified 2002 and there may be 20 some objection to that. But I think when it 21 comes to the discussion of accessibility, there 22 is certain areas where we don't need to split 1 hairs. 2 For example, if a voter votes in 3 private but someone helps put the ballot in the 4 printer but privacy is maintained, we can have a 5 discussion whether or not that is still 6 accessible. 7 But I am here today to talk about 8 the challenge you have as you finish your work on 9 these guidelines. 10 I think it is pretty difficult 11 listening, as you pointed out, Commissioner 12 Martinez, and as others have pointed out, that 13 you have people on all sides of the spectrum, 14 telling you what they think you ought to do, 15 however you are the ones who have to do it. 16 The problem is that I think the 17 standards need some tweaking. 18 And I am encouraging you - as I am 19 sure you are having these hearings for this 20 purpose - to undertake that tweaking during the 21 comment period that you will have in these next 22 90 days. 1 For example, we heard a lot today 2 about the word 'verification'. If you look at 3 the standards, it is not defined. 4 Some people will consider a paper 5 ballot verifiable if it simply reports whether or 6 not you have over votes or under votes. In the 7 case of an optical scan, that's usually what 8 verification means. 9 In other cases verification means, 10 will a ballot be read the way I think I marked 11 it. 12 And there is a tweak, if you look in 13 the glossary again, verification is not defined. 14 A larger area where I think you need 15 a tweak has been mentioned. Professor Rubin has 16 talked about the difference between a voter 17 verified paper audit trail and a voter verified 18 paper ballot. 19 If you look at the glossary again 20 and the definitions, there is no definition - 21 believe it or not, despite all this discussion - 22 of a VVPAT. It is not defined. 1 Voter verified paper audit record is 2 defined. That goes on to talk about an auditable 3 record. 4 If you take the standards as they 5 are drafted today, there is a possibility that 6 systems that are being rapidly deployed today - 7 from my mouth to God's ears, maybe ours - or our 8 competitors, for example, the AutoMark - are 9 either in one case not going to be covered by the 10 standards at all, if it is considered a VVPAT 11 after all, it is a voter verified paper audit 12 trail of something that happens on a computer 13 system, either it's not covered - or worse yet, 14 it's against the rules and cannot be deployed. 15 We have provided to NIST and have 16 had discussions with NIST and to your web site 17 six or seven pages where we think these tweaks 18 need to be taking place. 19 We don't think there is any need for 20 any wholesale rewriting or any throwing out of 21 the great work that has been done. 22 We have talked to NIST about those 1 points already and they agree with us. And, in 2 fact, I am authorized to say - although I don't 3 speak for NIST - I am authorized to say that 4 following those discussions, they have asked that 5 you specifically call upon them again to provide 6 comments to you based on what's coming out of the 7 people who have now had a chance to review the 8 draft standards, particularly the people who have 9 been involved from NIST in writing the draft 10 standards. 11 They agree that these holes do exist 12 and perhaps they would be very easy to fill and 13 then the country will then have voting standards 14 that can last for some period of time. 15 Finally, you are up against 16 timelines, vendors are up against timelines. You 17 have talked about, Madam Chair, we are not doing 18 a Version I and we are not doing a Version II, 19 and others have said maybe we will come out with 20 clarification. 21 My recommendation to you is, if in 22 all your balancing, the time to get this done 1 versus the time to get it as best as it possibly 2 can, to err on the side to getting it the best 3 you possibly can. 4 There is confusion, as you have 5 heard today, what should we buy, what can we buy, 6 is it going to be grandfathered in, are we going 7 to be found out of compliance by some new set of 8 standards. 9 I encourage you, in that difficult 10 balancing act you have got - I know 2006 is right 11 upon us - to get it so the next set of changes - 12 obviously, there has to be some - will be as few 13 as possible. Thank you. 14 CHAR HILLMAN: Thank you. 15 Ms. Hommel. 16 And I do need to ask - and I 17 apologize for doing this - but the remaining 18 presenters to please keep to the three minute 19 time line. I really do hate to have to cut 20 somebody off, but we are just about at 5 o'clock 21 now. 22 MS. HOMMEL: Thank you for the 1 opportunity to speak before you today. 2 My name is Teresa Hommel and I am 3 Chairwoman of the Task Force on Election 4 Integrity of Community Church of New York. 5 I am also a computer professional 6 and I have worked in the field for 38 years. 7 My remarks concern the lack of 8 standards in the proposed guidelines that would 9 guarantee that no matter what kind of computer 10 technology is used in elections, that ordinary 11 non-technical citizens can appropriately witness 12 election procedures. 13 I believe that voters must be able 14 to witness that their own votes are correctly 15 recorded and cast and election observers must be 16 able to witness the storage, handling and 17 counting of votes. 18 The problem with computerized 19 voting, obviously, is that no one can witness the 20 electronic ballots or votes. 21 And the further problem is that if 22 election procedures are concealed, then an 1 election lacks legitimacy, the government lacks 2 legitimacy and there is no reason for the public 3 to have confidence in the integrity of the 4 announced election outcomes. 5 For these reasons, the guidelines 6 should have standards to require computerized 7 voting and vote tabulating systems to be designed 8 to facilitate appropriate observation by 9 non-technical citizens. 10 The proper use of voter verified 11 ballot printout would solve this - and I didn't 12 put this in my written testimony, but I want to 13 make clear what I say, what I mean by the proper 14 use. 15 My understanding of the voter 16 verified paper printout is that the voter can see 17 their ballot and it is the ballot of record 18 legally. And then it is counted and people can 19 watch the counting and know that the count is 20 correct. 21 If the legal ballot is the 22 electronic ballot, then nobody can watch it being 1 recorded or cast or handled or stored or counted. 2 And I believe that that is totally contrary to 3 any democratic principles of any legitimate 4 democracy. 5 So I believe that these guidelines 6 need to have strong suggestions, since they are 7 long term guidelines, that the voter verified 8 paper printout, whatever you call it, should be 9 easy to verify, easy to handle, easy to count and 10 actually should be counted. 11 Elsewhere, as appropriate, I would 12 urge you to deal with the problem that even where 13 state laws require voter verified paper ballots 14 to be created, no law considers these ballots to 15 be a valid record and no law requires all of them 16 to be counted. 17 For example, you heard from Mr. 18 Kosinski earlier, our New York State proposed 19 law, which has not yet been signed by the 20 Governor, requires only 3% of the voter verified 21 paper ballots to be counted and includes no 22 standards whatsoever as to when a larger count 1 should be done as a result of discrepancy between 2 the paper and electronic count. 3 Computerized votes and vote 4 tabulating systems also conceal election 5 procedures from our own election officials and 6 prevent them from complying with their 7 responsibilities to oversee our election. 8 For example, in HAVA, Section 9 301(a)(5), there is a requirement that equipment 10 that counts ballots should meet the FEC 2002 11 accuracy standards, which allow a maximum error 12 rate of one in 500,000 ballot positions. But 13 there is nothing in the guidelines to ensure that 14 election officials can determine what the 15 accuracy of their counting equipment is and 16 whether it is in compliance with HAVA or not. 17 CHAIR HILLMAN: Could I ask you to 18 summarize. We have got your written testimony 19 and you have passed three minutes already. 20 So if you could just summarize to 21 your conclusion and we will read your written 22 testimony. 1 MS. HOMMEL: I also urge you to ban 2 any communication capability whatsoever in any 3 voting and vote tabulating systems. 4 I do not believe that it can ever be 5 handled securely. 6 Earlier this year my America Express 7 card was broken into. Last week we saw in The 8 New York Times that 40 million Master Card 9 accounts have been compromised. 10 And if the most secure computer 11 installations in our country, which are the 12 financial institutions, cannot control their 13 computers, what makes you think that your local 14 poll worker can. 15 In conclusion, I urge you, the 16 Commissioners, to set forth standards to 17 facilitate meaningful observations by 18 non-technical citizens and election professionals 19 of the entire recording, casting, storage 20 handling and counting of votes. 21 And I don't mean some bait and 22 switch piece of paper which is not the legal 1 ballot, but the ballots themselves. 2 And I urge you to declare that if 3 this cannot be done, that computerized voting is 4 inappropriate technology for use in this country. 5 Thank you. 6 CHAIR HILLMAN: Thank you. 7 Ms. Mahoney. 8 MS. MAHONEY: Thank you. I am 9 Martha Mahoney. Can you hear me? 10 CHAIR HILLMAN: Yes. 11 MS. MAHONEY: Thank you. You have 12 been given a copy of the Miami-Dade Election 13 Reform Coalition's report on the 2004 election. 14 I want to talk to you about three 15 issues that were revealed there and their 16 implications for the system guidelines. 17 I will submit written comments in 18 much greater detail later. 19 There were three things that are 20 really notable from Precinct 816. 21 One was, there was a vote counting 22 problem in which the votes from a single machine 1 added 85 ballots to the canvass three times 2 instead of once, so that 945 people signed in but 3 1116 ballots were certified in the final canvass. 4 Nobody noticed and the failure to 5 notice is one of the points that I am going to 6 draw to your attention. 7 There was an inaccurate total public 8 count that has not been addressed or explained 9 anywhere that was only possible because it wasn't 10 flagged. 11 It wasn't brought to the attention 12 of poll workers by the system and there were 13 discrepancies in machine memory that have only 14 been reviewed by the vendor without independent 15 experts or independent evidence of any sort. 16 First, in the IVOTRONIC (ph) that 17 was being used, there was a known security flaw 18 had been criticized in the compu-air (ph) study 19 that would allow votes to be added repeatedly to 20 the canvass from a single machine, without 21 notifying the poll workers this was happening. 22 We don't know if that flaw caused 1 this or if there is an additional way to add 2 votes to the IVOTRONIC without notifying poll 3 workers. 4 At the bottom line, your Human 5 Factors need to address poll worker usability 6 more than they do. 7 I see the focus on voters. But poll 8 worker usability can be crucial to the integrity 9 of the canvass, as a system matter. 10 And I am calling on you to think 11 about these unflagged problems. 12 The guidelines emphasize recording 13 votes accurately and counting them, but they 14 don't emphasize preventing them from being 15 counted repeatedly by poll workers. This is a 16 gap. 17 Now the second problem I want to 18 draw to your attention is that there were 945 19 voters who signed in. And when you added the 20 individual public counts on the machines, they 21 should equal 945. 22 There was an ambiguous record that 1 showed one more ballot, so maybe it would have 2 reached 946. 3 But when you look at the results 4 tape at Tab 3a,you will see that the voting 5 system, not just that machine, reports a total 6 public count of 947. 7 When a machine can't add 12 numbers, 8 it should be like a student in the back of the 9 class. It should raise its hand and say, 10 teacher, I have a problem with my math. 11 We do not train human beings to 12 check the math of the machines. These machines 13 have to bring tabulation and other addition 14 errors to the attention of the poll workers or 15 the integrity of the system is at risk. 16 Finally, there were conflicting 17 records in the different memories on the machine. 18 They went to ES&S and were studied. 19 The report was not released to the 20 state and county for six months until after it 21 came out in the paper. 22 My concern here is that the vendor 1 did their evalation without independent expertise 2 or independent evidence. And we need to call for 3 procedures or to consider the roll of the 4 independent records in a way that will help solve 5 that gap. 6 Because when you have one record 7 that says 84 ballots and one that says 85 and 8 this is not - I am not right now judging the 9 ultimate conclusion of the report. I will put 10 that in writing - there has to be an incentive 11 for the vendor to reach one conclusion or 12 another. 13 The roll of independent evaluation 14 is extremely important. 15 If you get one conclusion you might 16 conclude the system worked better or some other 17 advantage happened. 18 We need to think further about this 19 problem. 20 I will submit written comments. 21 These were things I think that no 22 one envisioned: a tape that couldn't add 12 1 numbers; machines going in repeatedly. 2 Much of what I have said comes under 3 human factors. Please consider extending it 4 further. Thank you. 5 CHAIR HILLMAN: Thank you. 6 Mr. Campbell. 7 MR. CAMPBELL: Good evening. My 8 name is Edmund Campbell. I am a union activist 9 of the South Florida AFL-CIO Central Labor 10 Council and a member of the Miami-Dade Election 11 Reform Coalition. 12 I am here to bring to your attention 13 problems that I saw while I was doing some poll 14 closing at the 2004 November election in regards 15 to early voting throughout. 16 Election Day is a 12-hour day and at 17 the end of the day there was voting verification 18 and printing of results for the precinct. 19 Early voting, on other hand, is day 20 after day of an incomplete process, which ends 21 before Election Day with equipment being taken 22 away without any results being printed and posted 1 at the polling site. 2 DRE's are left powered up for the 3 duration of the election in public areas. Early 4 voting demands physical security over extended 5 time and incremental ballot counting every day. 6 We need appropriate procedures to 7 protect security and accuracy. 8 I even heard poll workers actually 9 say, close is good enough. So what if the count 10 is off. 11 The thing that is unique about early 12 voting is that it takes up to 18 days of 13 operating in a public place, while individual 14 election officials are making rules governing the 15 new trend in American voting reform, using 16 procedures that was not meant for more than a one 17 day election, when the poll opens early in the 18 morning and closes that night. 19 We have started the convenient era 20 with early voting. 21 If the EAC hopes to create 22 guidelines designed to protect the integrity of 1 our voting system, members need to address early 2 voting separately from Election Day and call for 3 early voting best practices and procedures that 4 will directly address this unique condition. 5 Thank you. 6 CHAIR HILLMAN: Thank you very much. 7 That concludes our public comment 8 presentation. 9 Colleagues, as you can see, our 10 guidelines affect many constituencies. 11 And we encourage all of your 12 comments to come in throughout the 90 day period 13 so that we can be as informed as possible. 14 Any closing comments? 15 VICE CHAIR DeGEORGIO: No. I just 16 wanted to compliment you, Madam Chair, on a job 17 very well done. 18 CHAIR HILLMAN: Thank you. I 19 appreciate it all. 20 If there is nothing further, then 21 our hearing stands adjourned. 22 (Thereupon, at approximately 5:10 1 o'clock, p.m., the above meeting 2 was adjourned.) 3 * * * * * 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 1 2 3 4 5 C E R T I F I C A T E 6 7 8 I, ELLEN REACH, a Stenotype Reporter 9 and Notary Public of the State of New York do 10 hereby certify that the within transcript is a 11 true and accurate record of the Public Meeting of 12 the U.S. Election Assistance Commission, held on 13 June 30, 2005. 14 15 16 17 18 19 _____________________ 20 ELLEN REACH 21 22 |