To the Editor:

Your Sept. 29 front-page article "Professor Shows Flaws in Touch-Screen Voting" propagates a number of misconceptions about voting machines that need to be clarified. You present one computer security expert, Edward Felten, showing how the machines can be hacked and expressing his opinion that the machines need to be equipped with voter-verifiable paper trails. The rest of the article then goes on to discredit his worries, by quoting voting officials and other experts. The impression conveyed is that there are two camps, pro-paper and anti-paper, of equal popularity, making equally strong arguments.

But this is a false impression. The vast majority of computer security experts are convinced the best currently available method to improve the security of elections is to supply the machines with a voter-verified paper trail, as recommended by Professor Felten. The Verified Voting Foundation has collected the endorsement of more than 10,000 people, including thousands who are professionals in the computing field. In a survey taken of the membership of the Association for Computing Machinery, more than 95 percent of those responding also take this position. The number of computer professionals who take the opposite position is minuscule, with good reason. Paperless systems are susceptible to undetectable manipulation by attackers, undetectable manipulation by the companies that make the machines, software errors, hardware errors and misconfiguration errors by poll workers. The certification processes in place are totally inadequate to prevent these problems -- indeed, there is no known certification process that could.

The article also claims that "only minor glitches" were encountered in the May primary election in Allegheny County using the new paperless voting machines. This statement is flatly contradicted by "Allegheny County Elections Board Ratifies Vote Totals" (Post-Gazette, June 13), which described very serious complaints that were raised against the voting process used in that election. In view of such obvious problems, why should we believe these machines are accurately carrying out the unauditable parts of their job -- like counting the votes? Why should voting require such a leap of faith when we could have voter-verified paper ballots? Twenty-seven states already do.

DANIEL SLEATOR

Professor of Computer Science

Carnegie Mellon University

Oakland

The letter also was signed by Roger Dannenberg, associate research professor, Department of Computer Science and School of Art; David Eckhardt associate teaching professor, Computer Science Department; Adrian Perrig, assistant professor of computer science, electrical and computer engineering, engineering and public policy; and Michael Reiter, professor of electrical and computer engineering and computer science, all of Carnegie Mellon.