SDI software to be debatedTo the Editor:
In recent months, scientists throughout the country have begun questioning the feasibility and desirability of President Reagan's proposed "Star Wars" defense against ballistic missiles.
Much of the debate has focused on the computer software required to support Star Wars: ten million lines of code, distributed among ground, air, and space components, which would control the detection, tracking, and interception of thousands of enemy missiles. Such software would have to accomplish its task within minutes, with little or no human intervention.
This Monday, Oct. 21 at 8 pm in Kresge Auditorium, four computer scientists will debate the feasibility of building software that can reliably accomplish this mission.
Two members of the Star Wars panel on Computing in Support of Battle Management -- Danny Cohen of the University of Southern California's Information Sciences Institute, and Charles Seitz of Caltech -- will face professors Joseph Weizenbaum of MIT and David Parnas of the University of Victoria, Canada. The formal topic will be: "Star Wars: Can the Computing Requirements be Met?"
Last June, Parnas resigned from the panel of which Cohen and Seitz are members, claiming that the required software could never be built.
In a series of memos attached to his resignation letter, Parnas exlained why he believed Star Wars software could never be fully specified or adequately tested before use, why it would inevitably contain serious bugs, and why research in artificial intelligence, program verification, software engineering, and automatic programming would not improve its feasibility.
In his resignation memos, Parnas also claimed that panel members have a vested interest in continued funding for Star Wars computer research, and have no experience in developing battle management software.
Next Monday's debate will be co-sponsored by MIT's Laboratory for Computer Science and the Boston chapter of Computer Professionals for Social Responsibility (CPSR). CPSR is a national organization of computer professionals concerned about the effect of computer technology on society, and especially on the arms race. The organization began in 1981 as an electronic mailing list at a Palo Alto research center, and gradually grew to a national group of 700 members in twelve chapters throughout the country. CPSR's Boston chapter meets monthly at MIT.
Ron Newman '79->