The Trouble with Computers: Usefulness, Usability and Productivity.

Yoon, Yong J.

Man's hankering for intelligent devices dies hard. Since the time of our cave-dwelling ancestors, the adaptability of our physiology has not changed much; the size of our brain and the dexterity of our hands have not improved. But we have constantly added new tools for growing crops and building shelters, for governing and waging wars, and for building jet airlines. We have also learned how to design and apply computers, which have fascinated many visionaries including economists.

Economists have been constantly interested in the information technology that was made possible by computers. Herbert Simon and others economists believe that this is equivalent to the industrial revolution. But Landauer, a cognitive psychologist (not an economist), tells us that, considering the early enthusiasm and heavy investment in the computers, the performance record of computers in enhancing productivity has been poor. He believes the computers are in deep trouble.

Those nations, industries or people, who have invested heavily in the computers have not received proportional returns except to those who sell computers. Only the communications industry appears to have exploited the technology to significant advantage. This issue certainly concerns economists, especially when many economists suspect that information technology is partly responsible for the current income inequality and the impoverishing of low-skill workers in the United States and Europe.

Landauer reviews statistical data and asserts economic conjectures on the computers and information technology. Some of them are not very convincing or fully justified, but I found them interesting. Against the conventional view that the spillover effect of computers is higher than we see from statistical data alone, the author argues that the computer did not contribute much in productivity. The author also argues that the productivity decline in the '70s was caused by the poor investment in computer technology. Landauer thinks that evolution of the market, or market force, does not bring about the potential of computers. Economists would ask: then what would?

What is more interesting is his insight into the use of the computers in increasing labor productivity. The improvement in computer hardware has been impressive and reduces calculation costs dramatically. Their raw power for computation continues to double every few years. But, Landauer argues correctly that calculation is only a small fraction of input in the industry (or in the economy). Instead of asking who need computers, we should ask who need calculations. Computers would not contribute much in improving productivity because calculation is not an important input to most projects except perhaps bomb and airplane design. We cannot assume that their economic value is directly proportional to computational power. The premise of the whole book is that computers have not contributed nearly as much to labor productivity as we have believed they should.

The way computers are designed and have been developed and the computer's effects on the efficiency of their users seem to have intrigued Landauer, who is himself a software engineer and an AT&T consultant. He finds evidence that information technology does not help much in the first place; he analyzes why, and proposes remedies. The remedies involve user-centered design. His argument is convincing.

Basically, computers can do anything that can be reduced to numerical or logical operations. This makes automation feasible. But the gain from automation is limited, as most of the easily reached opportunities have been already exploited. Computers are helpful for new scientific knowledge, too. However, the next phase of computer application is what the author calls augmentation. This involves the wide range of things people do that cannot be easily taken over by a calculating machine. Most of the things people do - talk, understand language, write, persuade, decide, administer - fall into this category.

So far we have not been very successful in this phase. Landauer's explanation is that the software is too difficult to use. This is partly because the inclination of the developers in the past has been to just design a powerful feature, then let someone figure out how to teach its use. Computers are powerful and flexible, and there are many things that they can be made to do. Boolean algebra can map language for computers to work on, but cannot map the rich connotations we associate with language. It will take a while until we fully understand and benefit from the revolution in information technology made possible by computers since the 1960s.

The author emphasizes that the solution is to develop user centered software. He tries to help us understand this problem by analyzing some of the success stories in computer software development. But his analysis does not go beyond the surface of these case studies. His examples are helpful and awaken our superstition of believing more than what the computers can do. But the book lacks any insight into using computers at the phase two level. Perhaps Landauer is not to blame and just indicates where we are. But his repetitive illustration of statistics does not help much.

At the end of the book, Landauer offers an interesting vision and disclaimer. "This shared capacity was first manifest in language, later in writing, math, and science. By comparison with our forebears, each of us has become a genius. The human mind is not housed in an isolated block of tissues in a person's skull. It draws on the whole wealth of stored human knowledge and the whole power of our shared mental tools. The growth of this mental power has probably just begun. Computers offer so much' more than mere language and will make mental powers appear to our great-grandchildren the way those of chimpanzees appear to us today. But it won't work - not until it is based on thorough task analysis in a realistic setting, not until it's mocked-up and tried and revised, and again, then again, then again."

This vision is shared by those developers of internet and networking technologies.

Yong J. Yoon Public Service Commission, Washington, D.C.