A Robot's Job

Nearly every science nowadays, even fleshy anthropology, is bogged down in figures which mean little until digested mathematically. Sometimes this chore is simple, requiring only an adding machine, or a pencil, paper and persistence. More often, as science takes off into thinner & thinner abstractions, each calculation is a double-jointed equation. There may be thousands to solve, each a mind-racking job. Most established research centers have bales of figures lying around which no one has time or courage to analyze.

To distill these bales into useful conclusions, scientists are turning to giant calculating machines which eat up equations as quickly as small boys gobble peanuts. Last week Harvard University dedicated its new Computation Laboratory, devoted solely to overgrown abaci, their design, construction, care & feeding. Two hundred scientists, engineers, mathematicians gathered to hear the latest plans, and to yearn for more calculators. Mark I, Harvard's first, was operating. The electronic entrails of Mark II, under construction for the Navy, were still in the semiassembled stage.

Fast & Narrow. Dr. Howard H. Aiken, director of the laboratory, does not like to hear his machines called "mechanical brains." "These humanitarian terms are unfortunate," he says severely. But he does admit that they work more or less like fast, narrow-minded brains. Like the brain, the machine accepts information, generally in the form of figures represented by small holes punched in a paper tape. It salts them away in a kind of "memory." (Dr. Aiken prefers "the relatively modest term: storage of numbers.") Then it combines them into conclusions, as human brains try and often fail to do. Unlike most human brains, it stops when it makes a mistake.

The machine's range of acceptance is strictly limited. It cannot examine a field and a pretty girl, and conclude from the data available which would be more worth cultivating. Such semi-tangibles are not for it. Figures alone it accepts, in floods and mazes. Quick as a midget's wink, it adds, subtracts, multiplies, divides, raises to powers, extracts roots (square or better). It blends the figures together, mixes them with constants such as the speed of light. "It's a robot," says Dr. Aiken, "and does just what it's told."

For Harvard & Navy. Mark I is working round the clock. All day & night a river of figures streams through its insides. Four hours it works for Harvard. The rest of the time it works for the Navy, on such problems as calculating the way of a rocket in the stratosphere, influenced by air resistance, wind pressure, gravity, and the earth's rotation, and other more subtle factors.

Military problems are not all it can do. Physicists, mathematicians, engineers drool with desire to punch their problems on its willing tape. Economist Wassily W. Leontief of Harvard wants the calculator to put its mechanical mind on the nation's economic planning.

It is almost impossible at present, he says, to calculate all the effects of a new tariff or tax. Too many factors have to be considered, combined, evaluated before we know how all the nation's industries will be damaged or benefited. The same is true of a public-works program, or a wage change. The calculating machine would race through the interlaced figures, come up with conclusions much better than the primitive guesses of economists. But Economist Leontief admits that the robot's forecasts may be fallible. Says he: "No machine will tell if the miners are going to slow down."

This file is automatically generated by a robot program, so reader's discretion is required.