Book Review: Computer Lib/Dream Machines

This essay is reproduced here as it appeared in the print edition of the original Science for the People magazine. These web-formatted archives are preserved complete with typographical errors and available for reference and educational and activist use. Scanned PDFs of the back issues can be browsed by headline at the website for the 2014 SftP conference held at UMass-Amherst. For more information or to support the project, email

Book Review: Computer Lib/Dream Machines

by the Boston SftP Computer Group

‘Science for the People’ Vol. 10, No. 2, March/April 1978, p. 29–31

Computer Lib/Dream Machines By Thedor H. Nelson (Chicago, Hugo’s Book Service, 1974) $7.

Rapid technological change has in the past been accepted as the quickest and best method for developing a better society. Recently, widespread use of the electronic computer has rekindled the hopes of many people interested in using technology as a force for positive social change. Theodor Nelson’s dual book, Computer Lib/ Dream Machines attempts to deal with many of the potential consequences of today’s computerized society. 

Nelson argues for the demystification of computers for the non-computer oriented community. He aims to eliminate the dull and mediocre aspects of everyday life—in education, business, writing, and so on—by incorporating computers and computer programming into every phase of our lives. Nelson’s “simple soup-to-nuts overview of what computers are really about, without technical or mathematical mumbo-jumbo, complicated exercises, or talking down” is a far cry from the technical elitism found in most of the computer field.1

Demystifying computers is certainly a difficult but admirable task for an author to deal with; Computer Lib/ Dream Machines does provide useful insights for the highly motivated person who has had some exposure to computers beforehand. Unfortunately, it has been our experience that people with no previous contact with these machines have great difficulty reading or even developing an interest in Nelson’s book. He has arranged Computer Lib/ Dream Machines along the Whole Earth Catalogue format with short, highly specific articles on programming techniques, languages, and computer design intended to give people a basic technical understanding of what computers are all about. Such an understanding would help people, Nelson believes, to take advantage of the potential usefulness of computers. While a certain amount of technical knowledge can help us gain control over the way science and technology is used in our lives, basic programming and sketches of state-of-the-art projects are not the kind of information non-computer people seem to find most useful.

One of the great strengths of Computer Lib/ Dream Machines is the definition and unmasking of cybercrud—“putting things over on people using computers.” 

At every corner of our society, people are issuing pronouncements and making other people do things and saying it’s because of the computer . . . Cybercrud is, of course, just one branch of the great game of technological pretense that has the whole world in its grasp.2

In fact, a computer simply follows the instructions written by a computer programmer, doing nothing in and of itself. Computer analysis of a safety hazard, for example, is limited by the analysis program developed for the problem, by the data collected for the study, and by the assumptions underlying both the problem specification and data collection. Just as numbers can be juggled to prove contradictory points-of-view, so a computer can be made to state anything is safe or harmful just by changing the evaluation methods or the safety limits or by using biased data. The computer’s presence will in no way assure an accurate answer; rather, its power can reinforce the apparent legitimacy of errors. 

Similarly, innovative applications of computer technology are not necessarily the best, most efficient and most just methods of accomplishing a task. Nelson repeatedly neglects this basic idea, as in his strategy for a “humane” educational system. He identifies the American educational system’s major structural problems as fixed curricula and sequence which rob the student of motivation and enthusiasm, an emphasis on testing and scoring that reinforces rote learning rather than an interest in knowledge, and the division of material into subjects of specialization. The source of these problems according to Nelson, is the “inhuman teacher” and “it is to rescue the student … and allow him [sic] to relate directly and personally to the intrinsically interesting subject matter, that we need computers.”3

Nelson sees the computer as an educational tool of great flexibility, permitting each student to work at his/her own pace, thereby “individualizing” the instruction. The computer would also offer the advantage of “responding resources”—or “hypermedia” as he calls it—in representing written and pictorial material as needed on a video computer terminal. 

Putting aside the present technical limitations on Nelson’s educational proposals, the greatest problem in the American educational system may not be so much how things are taught, but what the system is trying to accomplish. As E.P. Cubberly, an educational historian and theorist at the turn of the century, stated: 

Our schools are, in a sense, factories in which the raw materials (children) are to be shaped and fashioned into products to meet the various demands of life. The specifications for manufacturing come from the demands of twentieth-century civilization, and it is the business of the school to build its pupils according to the specifications laid down.4

The most probable result of Computer Aided Instruction would not be the “individualized” instruction Nelson foresees, but rather a greater degree of individualized control and standardization of the product. This would be “necessitated” by the tremendous and costly programming effort involved in setting up the system. Imagine every lesson being carefully formulated by state or federal agencies and universally distributed, with minimum interference by human teachers (who now may at least sometimes present students with a dissenting view). 

Computer Lib/ Dream Machines contains an enlightening discussion of the practices and policies of the computer industry, particularly of its dominant company, IBM. This corporation has become the model of efficiency and growth for much of the business world. Over the last twenty years, IBM has defined and symbolized what a computer is or could be: a complex, expensive machine attended by an army of experts. “IBM seems to have molded computers in its own image,” Nelson writes, “and then persuaded the world that this is the way they have to be.”5 Like large corporations in other industries, IBM absorbed the bulk of its market by subtly forcing those who chose IBM equipment to stay with IBM as long as they used computers. Users of IBM equipment were sometimes captured through devious marketing practices here and abroad; by blatant deception; and by a disregard for human rights. IBM designs its equipment to be incompatible with accessories designed by other manufacturers; some of these accessories add up to half the cost of a computer. Further, in a marketing technique familiar to all new car buyers, IBM delivers machines with no detailed manuals describing their internal workings. The buyer must purchase service from IBM for even minor troubles. 

Nelson points out the absurdity of the idea—fostered by IBM—that computers must be large and costly. It is the smaller, less expensive (under $ 10,000), application-oriented machines currently appearing that Nelson hopes will pave the way towards a just society. While there can be no doubt that computer technology will soon be widespread, we can see no reason that this necessarily constitutes a remedy for the present injustices in the United States. Funneling hopes for social progress to the manufacturers of mini and microcomputers, who must exist in a highly competitive market, cannot be any more realistic than previous faith in the benevolence of IBM. Singling out the history of IBM provides a clue into the future of smaller companies which capture a substantial portion of the minicomputer market. We cannot accept Nelson’s implication that a small computer must come from a relatively small manufacturer, or that this supposedly small corporation will therefore hold public interest over profits. 

Another aspect of Computer Lib/ Dream Machines is most disquieting—a pervasive, perhaps unintentional sexism. Although a great many women are involved in all aspects of the computer field, we see little of them in this book, except for an occasional picture of a naive Dorothy (snipped from the Oz books) being amazed by a decidedly male Tin Man. All photographs and drawings of people doing something relatively complex (e.g. programming) are of men. On the other hand, Nelson seems to agree that the natural sex for a keypuncher is female—at least his drawings do nothing to dispute this stereotype. 

Probably the most serious injustice to women is the total lack of acknowledgement of their major contributions to the field. While we don’t subscribe to the view of history as a succession of giants making individual breakthroughs, it is nevertheless galling to see Nelson carefully give credit to John Backus for developing the programming language FORTRAN, Kemeny and Kurtz for the language BASIC, and never even mention the women who literally invented computer programming. In the early 1800’s an Englishman, Charles Babbage, tried and failed to build the first true digital computer. In the process, his friend and colleague Ada Augusta developed the basic principles of programming.6 Also, in the early 1950’s, debate raged over whether computer programming would best de done directly with machine languages or through software programs7 which would allow the programmer to work in languages suited to the problems she or he was trying to solve (e.g., FORTRAN or BASIC). Dr. Grace Hopper set up the first major software development group, against great opposition, and virtually all programming since then has been done in the way she championed.8

Nelson’s preoccupation with computers as a technological innovation has distorted his sense of purpose: to identify and attack the many basically oppressive aspects of computer technology. In writing this book he has attempted to suggest a mechanism for social change. Like most technocrats who optimistically apply technological solutions to social ills, Nelson addresses neither the existence of our most important social issues nor their political causes. In focussing on how computer technology could make a better world, he manages to sidestep the more probable oppressive aspects of the computerized society. 

In Nelson’s computer future, will the political and economic structures that cause poverty vanish? If so, then he must present some idea of how computer technology might accomplish this task. Computers, like much technology, can be used for beneficial social applications, and they can have a place in a socialist society, but progress in the form of computers or any other technology cannot by itself pave the way toward a just society.

>>  Back to Vol. 10, No. 2  <<


  1. Nelson, Computer Lib/ Dream Machines. p.2.
  2. Computer Lib, p. 8.
  3. Dream Machines. p. 18.
  4. E.P. Cubberly, Public School Administration, (Boston: Houghton-Mifflin Co., 1916), p. 338. This reference was found in Clarence Karier’s article, “Testing for Order and Control in the Corporate Liberal State.”
  5. Computer Lib, pp. 52-56.
  6. Pylyshyn, Zenon, ed., Perspectives on the Computer Revolution. Prentice Hall, 1970, pp. 16-28. See also Alice Hilton, Logic, Computing Machines. and Automation, World Publishing Co., 1963, p. 226.
  7. Machine languages/ software programming languages – the fundamental difference here is that a computer must eventually use a set of instructions which correspond to the actual electronic operations the machine can perform. This is the “machine language.” This language is, however, very tedious for people to work with and so “software” programs (as opposed to the electronic “hardware”) have been developed to translate program specifications which people write, in a more convenient language, into the machine language which is actually used when the program operates.
  8. Rosen, Saul, ed., Programming Systems and Languages, McGraw Hill, 1967. p. 4.