Building Careers and Shaping Public Policy

12 July 2001

The Honorable Sherwood L. Boehlert
Chair, Science Committee
U.S. House of Representatives
2246 Rayburn House Office Building
Washington, DC 20515

Dear Congressman Boehlert,

This is a letter to you and your Committee, expressing our support in principle for H.R. 2275, and its objective to reform the federal voting system. We are especially impressed that the definitions of the matters covered by H.R. 2275 include, "…products and systems used in every stage of the voting process from voter registration through recounts," and that the provisions of H.R. 2275 call for mobilizing technical support for the voting process.

Attached is the Entity Statement, "A Systems Engineering Approach to Reform of the Federal Voting System" created by the IEEE-USA's Committee on Communication and Information Policy (CCP). The CCP Entity Statement proposes a Systems Engineering Approach (SEA) for comprehensive reform of Federal voting systems. We clearly agree on the need for comprehensive reform and on the need to act quickly. We also agree on a requirement in the long term to create a new institutional structure along the lines of the Commission as proposed in H.R. 2275. When we compare the proposed legislation with the imperatives of the CCP Entity Statement, however, we identify important areas where we respectfully would make additional recommendations.

There are two natural timeframes implicit in the current situation: the Federal election of 2002 and the Federal election of 2004. It seems unrealistic to hope to create a new institution such as the Commission and require it to simultaneously displace existing institutions while integrating and reforming their ongoing activities, and do all this within a nine-month period, i.e., in time for the federal election of 2002. If significant changes in voting processes are going to come about before the Federal elections of 2002, they would have to be implemented through existing institutions.

Many institutions are currently attempting to provide reform within the limits of their authorities, mandates, and budgets. These entities embody institutional knowledge, memory, and expertise: resources the nation can ill afford to cast aside.

The several current actors include the IEEE Standards Association (IEEE-SA), the Federal Election Commission (FEC), and the National Association of State Election Directors (NASED); each is already at work on reforms. As you are no doubt aware, the IEEE-SA has established Standards Project 1583, a standards "Project for Evaluation of Voting-Equipment" already underway. The FEC has released for public comment, a draft of "Volume one, Version two," of its revised standard for voting systems. Under the aegis of NASED, the Election Center in Houston, Texas selects and supervises independent testing laboratories that evaluate voting equipment. The latter two institutions are operating under existing, though limited, authority with very limited budgets. To be effective, all are likely to require an agreed upon, formal coordination mechanism plus additional resources: sufficient authority to take required action during an initial transition, sufficient funding to act within that authority, and access to additional technical expertise. NIST has traditionally been a source of such expertise.

The institution(s) required to orchestrate comprehensive, fundamental reform in time for the Federal election in 2004 are likely to require characteristics very different from those needed for overseeing an initial transition. Nonetheless, they must be created and implemented at the earliest possible date. When initially created, these institutions can track, study and learn from the transition processes of 2002 while assessing the effectiveness of the existing actors.

In both the short and longer terms, it is important to take advantage of expertise wherever it exists. We endorse the provisions of H.R. 2275 that the Commission, after being created, is encouraged to operate through open, consensus standards-making processes of existing entities. These entities have demonstrated an ability to mobilize "citizen-action" in support of their activities: finding the right expertise at the right point to solve technical and other problems--often through voluntary contributions of governmental bodies (including NIST), not-for-profit agencies, private organizations and from individuals. The IEEE-SA, one such organization, is an ANSI-accredited, standards-writing body. (Recall that ANSI itself does not write standards.)

Our concerns also include matters that may be outside of the jurisdiction of the Science Committee. For example, there is a need to close-the-loop to ensure achieving performance goals, especially the mega-goal of ensuring that voting systems comply with agreed standards both quickly and uniformly in every state.

To achieve this goal, a complete SEA for Federal election reform must modify the incentives to the parties. Many local voting-jurisdictions, especially those with many constituents on the wrong side of the "Digital Divide," may not have sufficient resources to implement any of the voting systems that become certified for use under election reform. To avoid cries of "… unfunded mandate…," we urge that reform legislation include provisions for Federal funds to be made available through the states to permit such jurisdictions to implement certified voting systems. We urge further, that such funding not be made available unless the state in question can represent that with such funding, all of its jurisdictions will be able to implement certified systems prior to the next Federal election. As a further incentive and as a quid-pro-quo for Federal funding, we also recommend that legislation include the requirement (with safeguards of "reasonableness", and exceptions) that if any jurisdiction in a given state has not implemented a certified system, that whole state must use only hand-counted, paper ballots throughout the state in any Federal election.

We urge the Congress to ensure that the open, consensus standards developed under H.R. 2275 be implemented nationally, perhaps through the mechanism of OMB Circular 119 dealing with Federal implementation of standards arrived at through the open, consensus, standards-making process.

We are prepared to provide you and others in the Congress any assistance you might request.

Sincerely,

Ned R. Sauthoff, Ph.D.

cc: Members of the House Science Committee

A SYSTEMS ENGINEERING APPROACH TO REFORM
OF THE FEDERAL VOTING SYSTEM
(12 July 2001)

The Committee on Communication and Information Policy (CCP) of The Institute of Electrical and Electronics Engineers-United States of America (IEEE-USA) recommends developing a national voting system standard to the highest feasible quality level. To do so the nation must use core systems technologies similar to those implemented by the United States for other missions that are critical to the national interest. These systems technologies include systems engineering, quality assurance, human usability engineering and information security technologies.

To implement this recommendation the nation must:

Follow a formal process:

• Use performance standards based on best-practices of modern systems engineering under expert technical leadership; and incorporate the efforts of the consensus standards community of which the IEEE is a part.
• Recognize, in accord with the precepts of usability engineering, that user-error is a system failure; that failure of multiple users to "follow instructions" is caused by improper instructions or by improper system design; and further, that both types of failure must be foreclosed to the maximum extent possible through a program of system testing.
• Achieve simplicity, rather than complexity, in the features of the process to facilitate the reliability, integrity and security of the system.
• Adequately fund the effort throughout all stages of its development, implementation and operation.

Achieve at least the following performance goals:

• Maintain voting privacy. This requirement currently precludes the use of the Internet for balloting, but not necessarily for support of voting administration where it may be useful. This requirement also implies protection of all the electronic devices used in the system from inappropriate capture of electronic emanations.
• Approach absolute voting accuracy. Provide immediate ballot feedback to all voters, giving each voter confidence that the ballot reflects the choice intended. Such feedback must extend also to those with visual and other disabilities; e.g., for those with visual impairment, an oral feedback via a headset of the vote when entered may be appropriate; similar forms of feedback must be available to all voters.
• Allow for manual counting of ballots. Manual counting is necessary as both a check on the integrity of electronic or other types of systems, and as a backup against system failure. Research to-date suggests that this is best achieved by creation of a paper record of each ballot at an appropriate stage of the balloting process.
• Result in statewide compliance in all states.

Provide for proper oversight:

• Any voting system intended to be used by any jurisdiction must be assessed and certified, or not certified, against nationally-mandated performance standards. This will require oversight by an expert body. All problems or failures that occur in the use of certified systems must be reported to the oversight body in a timely fashion.
• Once created, these standards must remain current by incorporating changes in requirements appropriate to changing circumstances, including reported problems and failures of previously certified systems; if performance standards are modified, all elements of all voting systems must be re-assessed and re-certified.
• The National Institute of Standards and Technology (NIST) or an equivalent body must be given these permanent roles: 1) to provide expert technical leadership in developing a national voting system; 2) to act as the expert body ensuring that all voting systems intended to be used by any jurisdiction are assessed and certified, or not certified, against the nationally-mandated performance standards; and 3) to ensure that once created, nationally-mandated standards remain current.

Provide federal funding to states:

• Such funding is to be for jurisdictions on the wrong side of the "digital divide" within a given state, but only where such funding will result statewide in implementations of fully-compliant systems on a schedule timely for the next federal election.

Provide for freedom of choice (with responsibility):

• Each state must be free to authorize each of its jurisdictions to use a voting system of its choice from among those that are certified as above, but all systems implemented must be from among those that are so certified.
• Until all jurisdictions throughout a given state are using systems that are certified, all jurisdictions must use hand counted paper ballots.

Recognize the current context:

• The process can build on and extend the experience and expertise of groups such as the Federal Election Commission (FEC), a group that developed existing standards for voting systems and a process for qualifying systems as complying with those standards.

This statement was developed by the Committee on Communication and Information Policy of The Institute of Electrical and Electronics Engineers-United States of America (IEEE-USA), and released subject to review by the IEEE-USA Board of Directors. It represents the considered judgment of a group of U.S. IEEE members with expertise in the subject field. IEEE-USA is an organizational unit of The Institute of Electrical and Electronics Engineers, Inc., created in 1973 to promote the careers and public-policy interests of the more than 230,000 electrical, electronics, computer and software engineers who are U.S. members of IEEE.

| Top of Page | Policy Log | Public Policy Forum | IEEE-USA |

Last Update:  11 June 2001
Staff Contact: Deborah Rudolph, d.rudolph@ieee.org

Copyright © 2001, The Institute of Electrical and Electronics Engineers, Inc.
Permission to copy IEEE-USA policy communications is granted for non-commercial uses with appropriate attribution, unless otherwise indicated.