Federal Panel Rebuffs Guidelines That Insist on a Paper Trail

By Cameron W. Barr
Washington Post Staff Writer
Tuesday, December 5, 2006

A federal advisory group rejected a measure yesterday that would have discouraged states from using electronic voting systems that lack an independent means of verifying their results, according to a spokeswoman for the National Institute of Standards and Technology.

Members of the Technical Guidelines Development Committee, a group created by Congress to advise the U.S. Election Assistance Commission, deadlocked 6 to 6 on the proposal at a meeting held at the NIST headquarters in Gaithersburg. Eight votes are needed to pass a measure on the 15-member committee.

NIST spokeswoman Jan Kosko said the proposal was introduced by Ronald Rivest, a computer science professor at MIT who heads a subcommittee on transparency and security.

Rivest told committee members that software errors in paperless machines could go undetected, leading to a situation in which "an election result is wrong and you have no evidence to show that it's wrong."

Committee member Brit Williams, a computer scientist who has conducted certification evaluations of Georgia's paperless electronic voting system, opposed the measure. "You are talking about basically a reinstallation of the entire voting system hardware," he said.

Five states -- Delaware, Georgia, Louisiana, Maryland and South Carolina -- use machines without a paper record exclusively. Eleven states and the District either use them in some jurisdictions or allow voters to chose whether to use them or some other voting system.

The proposal was based on draft recommendations developed by scientists at NIST, who said in a document released last week that voting machines that do not produce a paper record of each vote "cannot be made secure."

The committee is beginning to prepare technical guidelines that the Election Assistance Commission is expected to adopt next year but that would not be implemented until 2009 or 2010. The guidelines are not binding, but many states require election officials to use voting systems that meet national criteria.

Echoing the work of the NIST scientists, Rivest's proposal suggested guidelines that reject systems "in which the correctness of the election results is dependent on the correctness of the software."

The Associated Press contributed to this report.

© 2006 The Washington Post Company