previous | contents | next

viii Preface


A final goal comes from our feelings as computer scientists that the variety of computer systems is a phenomena worthy of study in its own right. This book carries, therefore, an invitation to taxonomy-to asking how to classify the diversity of forms of computer systems that are coming into existence. Taxonomic endeavors usually take place in a field of natural systems, particularly biological systems. It may seem strange that a domain of artificial systems calls for taxonomic activity. But the demand for empirical classification exists whenever there is a population of significant size and rich structure. Rudimentary classification efforts have occurred for many populations of artifacts-for ships, for aircraft, for houses. This book should amply confirm that computer systems are complex and diverse enough- and undergoing enough continual proliferation and evolution-to command significant taxonomic endeavor.

Enough is said in the first two chapters about the new notations introduced in the book, so that nothing substantive need be added here. We apologize for inflicting new notation on the reader. We feel that good notations are really quite important for the aspects of computer structure described in this book. Much would be gained by the whole field of computers-by users, programmers, engineers, planners, buyers, sellers, manufacturers, students, and scientists-if relatively uniform notations came into common use. Although we have no illusions about the perfection of the notations we have introduced, we would be most happy if they cause a rise in concern for standard notations and nomenclature.

A large number of distinct systems are described in substantial detail. We have redescribed many of the systems in the common notation introduced in the book. The accuracy of all these descriptions is a major problem. Even where the papers are reproduced from the literature, this problem of accuracy remains-although then it is not ours alone. Even though we have taken pains to obtain accurate information on the systems and to portray them faithfully in our various descriptions and figures, there is no way we can be responsible for their ultimate accuracy. The PMS and ISP figures, in particular, cannot be guaranteed to be accurate representations of the systems they purport to describe. Ultimately, one would like to have simulation languages for such notations and to verify (up to the usual criteria of a debugged program) that a system given by, say, an ISP description, simulates the behavior of the target machine. But that day is still far off.

Our most fundamental acknowledgment is to the contributors to this volume, not only for the articles they have written, but for the computers they have designed and built, thereby creating a population of fascinating artifacts worthy of study. An additional reason for reprinting their articles rather than simply describing their computer systems is the importance of having available the views of the designers themselves about the nature of their systems.

The research on the basic ideas underlying the notations was supported by Advanced Research Projects Agency of the Office of the Secretary of Defense (F 44620-67-C-0058) and is monitored by the Air Force Office of Scientific Research.

We would like to extend an acknowledgment to the organizations that have produced all of these computers, oftentimes it would seem in defiance of the laws of economics. Perhaps, as the old saw has it, a computer manufacturer is simply a computer's way of breeding another computer. This might account for the tenacity

previous | contents | next