With sharply lowered machine cost, interpretive computing on terminals will become common for small problems, especially for students. Data base query systems do not demand expert programmers as users, and the data base itself could contain material of high interest to the general public. System software delivered thus far has been well received. The future LSI-boosted machine, embodying the fruits of data base research, assisted by the intelligent terminal, should drive down the cost and time per query, to within reach of the layman; and data bases then need no longer be reserved exclusively for executives of large firms. One should see very large, efficient query systems open to the general public, accepting queries in a restricted form of English. Such an effort may require initial sponsorship by the Federal Government. With data base techniques firmly in hand, the office terminal and the home terminal cannot be far behind. Eventually, a proliferation of display terminals may allow serious attempts at the elimination of printed hard copies, before the decade has run its course. **** 7. MAN-MACHINE INTERACTION Low cost machine intelligence will stimulate artificial intelligence research, but probably not enough to solve some of the nagging problems in the field in a decade. True breakthroughs may need to await the new structuring of entire ensembles of logical devices, to create either a semblance of, or an alternative to, biological intelligence. Such breakthroughs will surely be tried, but success cannot be presumed at this time. Thus problems in natural language comprehension,voice and handwriting recognition, language translation, theorem proving, and deductive reasoning probably will remain incompletely solved. However, the computer can handle simpler aspects of these problems; human help can be enlisted for the harder ones. Even in conventional data processing, whenever the machine is stuck because of insufficient information or too many alternatives, it can try to supply to the user reasonable guesses and their dire consequences, or just ask a pertinent question. The human then provides the needed direction. Man-machine interaction, combining the best of the twin worlds of computer precision and human perception, should become the fastest growth area of computer usage during the next decade. The lack of training of the user here is almost never a problem, the difficulty lies in equipping the machine to handle human-intuitive concepts. The most natural communication channels are of course sight and sound. Voice input may remain limited, but voice output can be quite general. Image processing and computer graphics will permit efficient two-way communication, and intelligent optical display terminals will be wide-spread, many of these in color. The possibility of man-machine dialogue without resorting to written messages should profoundly affect the use of computer systems, especially in education. Initially, these man-machine projects should rest upon applications, to lend proper weighting of priorities. General man-machine intelligent processing may then be distilled from a number of successful applications. **** 8. REALISM lN THE CLASSROOM The wide availability of low-cost computing power, even in conventional packages, can be of untold benefit to education. This is true not merely in the obvious areas such as research in computer sciences, fulfillment of explicit computing needs, or preparation of tutorial material. The computer should be exploited on a large scale as a new vehicle to challenge and sharpen younger minds, by exposing the latter to realistic problem solving situations. With the arrival of the pocket calculator, examinations in freshman physics no longer need to confine triangles to artificial side ratios such as 3, 4 and 5. A more subtle form of artificiality remains in education, however, in the nature of the problems being posed and solved in schools. Students today are given only well-formulated problems, the solutions of which requiring a small number of steps. The graduate soon finds, however, that realistic problems seldom come gift-wrapped, with attached answer sheets. Indeed, the identification of the problem is often the major aspect of real problem-solving. The culture shock of the new graduate will be greatly lessened, if realistically complex situations are treated regularly in the classroom. Professor Kemeny has pointed out that just the exposure of a student to a data base of statistical facts can stimulate him to draw and verify tentative conclusions (4). Such inductive reasoning is the main ingredient in problem identification. The possible consequence of a problem-solving step can be seen explicitly using a parameterized model. The ancient Greeks treated geometry as a mindsharpening device; in the computer we have a new, vastly more powerful tool. An excellent way to practice multilevel reasoning, for example, is by programming and running a computer. The student soon learns to expect punishment and rewards as the multilevel consequence of his decisions. He also learns to separate the forest from the trees, and to shift his forms of attention from one to the other, never losing sight of the final goal. This way he acquires the **** [Image] Videodisc players coupled with a microprocessor and keyboard will make tremendously powerful and flexible educational and recreational devices in the 198O's.