The Electric Psychologist

THE DREAM MACHINE

J. C. R. Licklider and the Revolution

That Made Computing Personal.

By M. Mitchell Waldrop.

Illustrated. 502 pp. New York: Viking. $29.95.

At the University of Wisconsin many moons ago, I remember bringing stacks of punch cards of the ''do not fold, staple or mutilate'' sort to a largish air-conditioned basement room with cubbyholed walls. I placed my rubber-banded punch cards, which constituted a painstakingly assembled program for a large mainframe computer to execute, into a cubbyhole and returned the next day to pick up a long printout. It usually indicated some trivial error in my program that meant I would have to repeat the whole exercise.

Now, of course, we use the personal computer to send e-mail, buy books, make reservations, do research, perform simulations, play games, listen to music and write book reviews, all in the relative privacy of our own homes and at our own convenience.

In ''The Dream Machine,'' M. Mitchell Waldrop has written a sprawling history of the ideas, individuals and groups of people that got us from punch cards to personal computers. He writes about the theoretical notions of abstract computing developed in the 1930's and 1940's, the impetus World War II and the cold war provided, the development of programming and business applications in the 50's and, most elaborately, the dawning realization among a few in the early 60's that ''batch processing'' of punch cards would never allow the computer to achieve its potential in complementing human skills. Refinements of these ideas in the 70's and 80's, especially at the Palo Alto Research Center of the Xerox Corporation, are also sketched out. Waldrop, the author of ''Complexity'' and ''Man-Made Minds,'' holds his many-faceted story together by relating it to one J. C. R. Licklider, a visionary and polymathic psychologist from Missouri who during his career at the Massachusetts Institute of Technology, in industry and at the Pentagon was mentor to many of those who created computing as we know it today.

The book is full of sometimes arcane battles over project funding and the proper focus of research, contests in which Licklider was invariably on the side of those in favor of decentralization, connectivity, interactivity and graphic interfaces and opposed to bureaucrats from industry and the military who were content with the top-down nature of mainframe batch processing. He was a most persuasive evangelist. After an academic career in cognitive psychology, he was appointed in 1962 to head a group within ARPA (the Advanced Research Projects Agency), formed by the Pentagon after Sputnik to conduct research on ''command and control,'' the making and carrying out of military decisions in real time. Convincing people that his crusade for interactivity meshed well with ARPA's needs, he used his $10-million-a-year budget to foster his dream of interactive computing.

Avoiding classified research, he began by underwriting project MAC at M.I.T., a mammoth undertaking in large-scale time sharing. Long before Apple Computer, MAC stood for both Multi-Access Computer and Machine-Aided Cognition and included Marvin Minsky's artificial intelligence laboratory. He also supported the work of John McCarthy, the inventor of time sharing; programs at the RAND Corporation, and at Carnegie Tech (now Carnegie-Mellon) with Allen Newell and Herbert Simon of ''bounded rationality'' fame; and the kindred spirit Douglas Engelbart of SRI International, who later developed the mouse, hypertext and windows.

Licklider was a catalyst, an enzyme that sparked seminal research. Later, to further the fruitful collaborations he had engendered among his widely dispersed investigators, he proposed a computer network to link them. Presciently, he wrote of the virtues of many subsequent developments, including disks and CD's, online libraries and software floating free of individual machines, much like the Java applets now so ubiquitous on the World Wide Web. The network he suggested became Arpanet, which eventually grew into the Internet and the Web itself.

An appealing feature of ''The Dream Machine'' is its comprehensiveness. Most of the big names are here, along with brief synopses of their ideas. Using the common trick in popular science writing of humanizing the players, Waldrop gives us not only Alan Turing's account of an abstract computer but his tragic death at 41. John von Neumann's work on computer architecture is sketched, as are tales about his lightning mental calculations. Norbert Wiener's cybernetics (the name didn't stick, but the prefix sure did) is presented as is his proverbial absent-mindedness. The same treatment is accorded Vannevar Bush's anticipations of hypertext, Claude Shannon's information theory and the psychologist George Miller's and the linguist Noam Chomsky's rebuttal of behaviorism.

Many other lesser-known names appear, among them Lawrence G. Roberts, Wesley Clark, Edward Feigenbaum, Ivan Sutherland, Kenneth Olsen, Robert Fano, Robert W. Taylor, Alan Kay. One such is Paul Baran, one of the many ''fathers of the Internet,'' who invented packet switching in an effort to make communications less vulnerable to attack or disruption. The term refers to the way e-mail messages, for example, are divided into small packets before being sent. The packets are transmitted individually and follow the most efficient routes to their destination. After they've all arrived, they're assembled into the original message. Without any big switches and fixed routes, the Internet, as Licklider hoped, is decentralized and can work around any local disturbances.

While impressive, the prodigious research that went into the book will likely generate a few disputes about primacy; such books usually do. More of a problem for most readers is that the book occasionally becomes tiresome and threatens to sink into a sea of details and acronyms. But Waldrop's periodic return to his leitmotif of computer accessibility and to Licklider's personal trajectory, and his references to the ambient political climate, usually keep us oriented.

Eventually the answer to the question why Licklider isn't better known becomes clearer. He wasn't mediagenic, didn't start a company, didn't invent a gadget or write any software. He was, however, a kind of ideal bureaucrat, whose detailed knowledge and selfless enthusiasm were instrumental in getting America and the world online and out of that cold basement with the punch cards. Waldrop's account of his and many others' world-transforming contributions is compelling.