Learning to tell the computer who's boss
Hanover, N.H. — There's a quiet intensity in John Kemeny's voice as he reviews some of the issues confronted during a quarter-century of monitoring, and contributing to, the rise of the computer as an educational tool.
Leaning forward, hands folded atop the long oaken table that dominates his conference room/office in Darmouth College's venerable Baker Library, Dr. Kemeny observes that for 20 years now he has believed that ''computer literacy is an essential part of a liberal education. And I do not believe that you can in any sense be computer literate if you've never programmed.''
That's a point of view some in academia take issue with - Yale professor Roger Schank, for instance. He is the author of ''The Cognitive Computer,'' a new book on artificial intelligence that's causing a stir within the computer fraternity. Dr. Schank has argued that ''computer literacy,'' including a knowledge of programming, is about as crucial to the average computer user as ''automobile literacy'' is to someone learning how to drive.
From his pastoral Ivy League campus, which now nurtures as many high-tech marvels as pine trees, Kemeny calmly pursues his point. ''It's not that I'm necessarily arguing that people are going to program for the rest of their lives. But you can't really understand what computers are about without having written some programs yourself.''
Back in 1964, Kemeny and his Dartmouth colleague Thomas Kurtz took a big step toward giving such beliefs practical expression. They co-invented the BASIC computer language and thus swung open the door to programming for thousands of relative novices. About the same time, the two scientists developed the first time-sharing system, which made Dartmouth's central computer available to a wider range of users and lent crucial momentum to a then embryonic high-tech revolution.
Throughout his academic career, including an eventful stint as Dartmouth's president (1970-81, during which period he chaired President Carter's commission on the Three Mile Island nuclear plant mishap), Kemeny has stood for the wider dispersal of computer know-how. In his view, people in this age have to know what those blinking screens and whirring disks are all about if they're to avoid the hazards of overestimating, or underestimating, what the machines can do for them.
In line with these views, Kemeny teaches, as often as three times a year, a computer literacy course for students who are not going into the sciences. The subject matter deals, in part, with the logic of programming, and students have to write or modify an average of 30 to 35 programs during the term. His current class has an enrollment of 110.
''Now, they're obviously not computer experts when they come in,'' Kemeny says, with an accent that hints at his upbringing in Hungary, ''but they have a totally different feeling about computers when they finish.''
He points out, for instance, that people who've been exposed to programming are not likely to be easily taken in when such terms as ''intelligent machines'' are bandied about. ''I don't like to be repetitious in class,'' says Kemeny with a half smile, ''but one thing I am very repetitious about - at least once every two weeks - is that computers are basically very stupid.''
If you ask a computer to solve a problem, he says, but forget to ask it to tell you what the answer is, it won't. Computers totally lack common sense, he asserts.
Students should also be given a chance to grasp just how much a program can reflect the attitudes, even the biases, of its conceivers, Kemeny says. ''That's one of the subjects I often speak of - about the hidden value judgments that are in computer programs. The programmer who writes a sophisticated piece of software in effect has built in a view of your world for you. Hopefully the one you want.
''But it could easily have assumptions or certain ways of looking at things which are not the way you wanted to look at it,'' he cautions. ''And unless you understand it, with enough sophistication to at least ask the right questions about it, you don't know what you have.''
He draws an example from the workings of a college bureaucracy. The software used by the campus accounting office is likely to view financial aid as an expense to the school - the standard budgetary approach. But an alternative view might be to look at such aid as uncollected or partial tuition on the revenue side of the ledger - an unorthodox approach which Kemeny in fact championed at Dartmouth, and which, he says, has revamped the school's financial aid policy.
But if the former perception is built into accounting software, you can't get at it, he argues. Then a school's governing board is, in effect, constrained to view things through the eyes of some tradition-bound programmer/accountant.
And he sees the implications of biased software ballooning as you move into governmental areas such as defense planning. ''Frankly, it scares the daylights out of me,'' says Kemeny, airing his concern that a ''star wars'' technology could leave the ''ultimate decision'' about war or peace with a computer. ''I just don't trust any computer that way. In particular, I don't trust the people who write the programs.''
Dr. Kemeny's concern with nuclear weapons has deep personal roots. He was among the scientists who took part in the Manhattan Project in the mid-'40s, which led to the first atomic bomb, and was at one time an assistant to Albert Einstein at Princeton's Center for Advanced Studies.
Returning the conversation to the classroom venue at Dartmouth, Kemeny offers one more insight on the value of exposing students to computer programming. It's a great teaching tool, he says. ''If they have to write the program themselves, they get a much deeper understanding of how it works than if somebody just explained it to them. It's a variant of the old adage that the best way to learn a subject is to teach it.
''What the students have,'' he says, casting about for the right words, ''what we care about, is that they're teaching a computer how to solve the problem. And they learn an enormous amount.''
Next Friday: A look at the booming use of computers among students and faculty at Dartmouth.