It won’t be ready for this year’s U.S. presidential election, but the hope is that it will be ready before the next one, four years hence. We’re talking about an IEEE standard that specifies the performance requirements of electronic voting machines.

The standard being developed is the IEEE 1583 “Standard for the Evaluation of Voting Equipment.” It is a performance standard, which means it won’t specify how a voting machine should be designed or manufactured. Instead, the standard outlines what a voting machine must do—for example, maintain the confidentiality of the voter’s choice, withstand hacking, be easy for voters to use, and take into account voters’ physical disabilities or language difficulties. The standard will also define the levels of electromagnetic interference the voting machine must withstand. Such interference could be caused, for example, by mobile electronic devices brought into the polling place.

A voter inserts her voter-identification card into a management terminal that will clear her to use one of the touch-screen voting machines in the background.

Once IEEE 1583 is accepted by the U.S. Election Assistance Commission (EAC), it will be the law of the land, and all voting machines used in federal elections will have to meet the standard.

“We’re not saying what kind of technology a voting machine should use,” explains IEEE Member Stephen Berger, a member of the IEEE 1583 working group. “The standard stops well short of anything that would lock in one kind of approach to voting over another.” The standard presents requirements that voting machines must meet, Berger continues, yet the requirements are general enough to cover any computerized voting machine, including touch-screen devices and optical scanning machines. Punch-card, lever, or other mechanical voting machines will not be covered, however.

The EAC is a new government agency charged by the U.S. Congress with developing technical guidelines for voting machinery. IEEE 1583 is the first in a number of election standards. The agency must also develop standards for ancillary equipment such as central tabulating equipment and software for registering voters. However, guidelines for how voting equipment should be tested will be incorporated in IEEE 1583. Another task is to develop guidelines for the various activities involved in running an election, which include registering voters and counting ballots.

DOING BETTER  The idea for the IEEE standard originated with members of the IEEE New York Section. Chagrined by the problems with punch-card ballots in the 2000 U.S. Presidential election, they believed that a standard for electronic voting machines that could replace punch cards would be the best way to apply the IEEE’s technical know-how and influence. A group of engineers headed by Berger came together in mid-2001 to brainstorm how to improve voting equipment. Given the importance of the issue, the IEEE Standards Association (IEEE-SA) decided that it would sponsor a standard for such equipment—the first instance where an IEEE society did not take responsibility for a standard’s scope and content. After taking several months to enlist experts and get organized, Berger’s group began working with the National Association of State Election Directors, whose members are responsible for the voting machines and voting procedures in each state, and the U.S. Federal Election Commission (FEC), which oversees how U.S. elections are funded.

The IEEE 1583 e-voting standard will specify performance guidelines for electronic machines that will replace mechanical machines such as this lever-activated device still in use in New York City.

The IEEE was not alone in its concern for the security and performance of voting machines. In October 2002, the U.S. Congress passed the Help America Vote Act (HAVA) which included money–US$2.3 billion this year–for the states to spend on replacing old punch-card and lever-actuated voting systems.

The act, which called for setting up the EAC, also required that a representative from the IEEE sit on the commission’s Technical Guidelines and Development Committee (TGDC). The committee will write standards for the ancillary election equipment.

“We were very flattered that Congress saw that we had a contribution to make and gave us a permanent place to make it,” says Berger, who represents the IEEE on the committee. Though the act creating the TGDC was passed in 2002, the committee had its first organizational meeting only last June because of the time it took to choose its members.

The IEEE 1583 working group’s first task was to look at existing election-equipment standards, including those issued in a two-volume set by the FEC in 1990. These standards hadn’t kept pace with newer technology such as touch-screen machines or machines that print out a paper record for the voter to double-check before leaving the voting booth, notes Herb Deutsch, chair of the IEEE 1583 working group.

The FEC standards deal with how the machines must be built and says nothing about the software that voting-equipment manufacturers later developed. The FEC standard is a performance standard, as is IEEE 1583, but it covers the hardware and software involved in an election with a very broad brush.

Shortly after the IEEE’s working group was organized, it provided more than 30 pages of comments on the FEC’s 1990 electronic voting machine standard, which was being updated. The IEEE considered ways that electronics could keep each vote confidential and make the machine easier for the voter to use than punch-card or lever machines. Berger reports the FEC adopted most of the IEEE’s recommendations. The working group then used them as a jumping-off point for its own standard.

REACTING TO COMMENTS  The IEEE 1583 working group reached its first milestone in August 2003 when it completed its “Draft Standard for the Evaluation of Voting Equipment.” The draft was sent for voting to all working group members—more than 200 people from eight countries—and the IEEE-SA.

The IEEE standard will encompass touch-screen voting machines such as this one being used in Tallahassee, Fla., USA.

The working group, which meets about every three months, is composed of IEEE members and representatives from voting-machine manufacturers, standards groups, disability advocates, and other professional organizations. The latter include the Information Systems Security Association, whose members are experts on hacking a computer system and guarding against it, and the U.S. National Association of Radio and Telecommunications Engineers, whose members contribute their expertise on electromagnetic interference.

The 257-page standard was not approved, which is not unusual for a first draft. But along with the balloting, the working group received more than 1000 comments on the draft standard, including suggestions for changes. The group’s next job was to address the comments. Berger says the second draft of the standard may be ready for another vote—and more comments—by the end of the year.

Not everyone agrees, however, with how the standard is being developed. Some critics feel the standard may do more harm than good. David Dill, an IEEE Fellow, professor of computer science at Stanford University in California, USA, and a member of the IEEE 1583 working group, is uncomfortable that representatives from voting-machine vendors dominate the group. The group could endorse a standard whose guidelines are good for their companies, and not good for voters, he says.

“I have serious reservations about the appropriateness of the IEEE standards process for something that is as important as the foundation of democracy,” says Dill, who was also a co-editor of a special issue of IEEE Security and Privacy magazine devoted to e-voting [see “Magazine Tackles Security of E-Voting ” ].

Berger reports, however, that the working group is taking heed of such criticism. It delayed release of its second draft standard by a year to provide enough time for the comments to be considered and, when desirable, to be worked into the draft. One result was that requirements for equipment that provides a voter-verifiable paper receipt were added. 

According to Berger, the standard could also be of value outside the United States. “Other countries may want to adopt the standard, and if there are differences in their election laws or practices we might develop amendments to accommodate them,” he says. He notes that in some European countries, votes are not anonymous—a voter’s name is attached to each ballot—which differs from IEEE 1583’s confidentiality requirement. Representatives from Europe and Asia are participating in the IEEE 1583 working group and could develop their own ideas for voting equipment in their countries.

To learn more about IEEE 1583, visit IEEE-SA at http://standards.ieee.org.