Writing on the head of a pin was a prophetic bit of showmanship, and pinhead-size computers will not necessarily have pinhead mentalities. This progress toward a seemingly hopeless goal takes on an inexorable quality when the writings of von Neumann are compared with the state of the art today. Starting out much faster but much larger than similar elements of the brain, computer components have been made even faster while simultaneously shrinking dramatically toward the dimensions necessary to produce quantitative equivalence. It happens that these goals work out well together, the one helping the other. Circuitry is now at the point where speed is ultimately dependent on that limiter of all physical activity, the speed of light, or of electrons through a conductor. Only by putting elements closer together can speed be increased; thus one quality is not achieved at the sacrifice of the other.
International Business Machines Corp.
This experimental “memory plane” consists of 135 cryotron devices built up in a 19-layer “sandwich.” Produced automatically, it is an example of continued shrinking of computer elements.
As an example of the progress being made toward speeding up computers, speakers at the recent Winter General Meeting of the American Institute of Electrical Engineers described a coming generation of “gigacycle” computers now on the drawing boards. Present electronic machines operate at speeds in the megacycle range, with 50 million cycles per second representing the most advanced state of the art. Giga means billion; thus the new round of computers will be some thousand times as fast as those now operating.
Among the firms who plan such ultraspeed computers are RCA, IBM, and Sperry Rand Corporation. To achieve such a great increase in speed requires faster electronic switches. Transistors have been improved, and more exotic devices such as tunnel diodes, thin-film cryotrons, magnetic thin-films, parametrons, and traveling-wave tubes are now coming into use. Much of the development work is being supported by the U.S. Bureau of Ships. Operational gigacycle computers are expected within two years!
Not just the brickmaker, but the architect too has been busy in the job of optimizing the computer. The science of bionics and the study of symbolic logic lead to better ways of doing things. The computer itself comes up with improvements for its next generation, making one part do the work of five, and eliminating the need for whole sections of circuitry. Most computers have a fixed “clock”; that is, they operate at a certain cyclic rate. Now appearing on the scene are “asynchronous” computers which don’t stand around waiting when one job is done, as their predecessors did.
One advanced notion is the “growing” of complex electronic circuitry, in which a completed amplifier, or array of amplifiers, is pulled from the crystal furnace much the way material for transistors is now grown. Pooh-poohed at first as ridiculous, the notion has been tried experimentally. Since a computer is basically a multiplicity of simple units, the idea is not far off at that. It is conceivable that crystal structure can be exploited to produce millions of molecules of the proper material properly aligned for the desired electronic action.
With this shrinking come the benefits of small size, low power consumption, low cost, and perhaps lower maintenance. The computer will be cheap enough for applications not now economically feasible. As this happens, what will the computer do for us tomorrow?
A figure of 7 per cent is estimated for the amount of paperwork the computer has taken over in the business world. Computer men are eyeing a market some five times that amount. It does not take a vivid imagination to decide that such a percentage is perhaps conservative in the extreme. Computer sales themselves promise to show a fourfold increase in the five-year period from 1960 to 1965, and in the past predictions have been exceeded many times.
As population grows and business expands in physical size and complexity, it is obvious that the computer and its data-processing ability will be called upon more and more. There is another factor, that of the internationalizing of business. Despite temporary setbacks of war, protective tariffs, insular tendencies, and the like, in the long run we will live in one integrated world shrunk by data links that can get information from here to there and back again so fast it will be like conversing with someone across the room. Already planners are talking worldwide computerized systems.