Active Regulation: General Systems Design Principles

By Gerald M. Weinberg

Active Regulation is Volume 3 in the General Systems Thinking series that begins with the world-wide best-selling, An Introduction to General Systems Thinking. (also available in ebook formats) That first series volume focused on the question,"Why do we see what we see?" The second and third books tackle the next question, namely "Why do things stay the same?"

Most people take for granted the stability they see in the world, but for systems thinkers, stability is one of the most fundamental properties, and puzzles, about the systems they struggle to understand and create. Indeed, it's such a fundamental property that it forms the topic of both this volume and its predecessor, Passive Regulation (Volume 2 in the series).
Originally part of the book, On the Design of Stable Systems, in its first, hardcover incarnation, Active Regulation does not just focus on computer systems, but systems of all kinds—human, natural, and technological.

In a highly readable, original presentation the Weinbergs here explore the subtle art and science of regulating systems, projects, and people in the most efficient and logical manner possible. The authors draw on their respective backgrounds in technology and social science to offer fresh insights and translate them into a language that anyone can understand.

In the course of this presentation, and with the help of dozens of figures, the Weinbergs introduce a host of laws and theorems derived from the best thinking of systems thinkers over the past century. For the most part, the regulation schemes examined in this volume are based on error-controlled regulation—in many ways the most effective of all regulation strategies.

Chapters include:
The Search for Regulation
The Homeostatic Heuristics
Other Regulatory Heuristics
Types of Regulatory Mechanisms
Regulation and Environment
When the Model Fails
Making Regulation Mysterious
Overly Simple Views of Regulation
Blindness and Reversed Vision

The book also contains an extensive bibliography and an appendix explaining some useful mathematical notation.

In addition to being a reference book for professional and lay people alike, Active Regulation is suitable as an undergraduate text in the humanities, social, natural, and engineering sciences. It is unique in its approach, highly readable, and offers practical ways of solving problems.

• Language: English
• Publisher: Gerald M. Weinberg
• Release date: Jun 6, 2011
• ISBN: 9781458155672

Gerald M. Weinberg

Gerald M. Weinberg (Jerry) writes "nerd novels," such as The Aremac Project, Aremac Power, First Stringers, Second Stringers, The Hands of God, Freshman Murders, and Mistress of Molecules—about how brilliant people produce quality work. His novels may be found as eBooks at or on Kindle. Before taking up his science fiction career, he published books on human behavior, including Weinberg on Writing: The Fieldstone Method, The Psychology of Computer Programming, Perfect Software and Other Fallacies, and an Introduction to General Systems Thinking. He also wrote books on leadership including Becoming a Technical Leader, The Secrets of Consulting (Foreword by Virginia Satir), More Secrets of Consulting, and the four-volume Quality Software Management series. He incorporates his knowledge of science, engineering, and human behavior into all of writing and consulting work (with writers, hi-tech researchers, and software engineers). Early in his career, he was the architect for the Mercury Project's space tracking network and designer of the world's first multiprogrammed operating system. Winner of the Warnier Prize and the Stevens Award for his writing on software quality, he is also a charter member of the Computing Hall of Fame in San Diego and the University of Nebraska Hall of Fame. The book, The Gift of Time (Fiona Charles, ed.) honors his work for his 75th birthday. His website and blogs may be found at http://www.geraldmweinberg.com.

Book preview

Active Regulation: General Systems Design Principles

by

Gerald M. Weinberg and Daniela Weinberg

SMASHWORDS EDITION

PUBLISHED BY:

Gerald M. Weinberg on Smashwords

Active Regulation: General Systems Design Principles

Copyright © 2011 by Gerald M. Weinberg

All rights reserved. Without limiting the rights under copyright reserved above, no part of this publication may be reproduced, stored in or introduced into a retrieval system, or transmitted, in any form, or by any means (electronic, mechanical, photocopying, recording, or otherwise) without the prior written permission of both the copyright owner and the above publisher of this book.

Smashwords Edition License Notes

This ebook is licensed for your personal enjoyment only. This ebook may not be re-sold or given away to other people. If you would like to share this book with another person, please purchase an additional copy for each person you share it with. If you're reading this book and did not purchase it, or it was not purchased for your use only, then you should return to Smashwords.com and purchase your own copy. Thank you for respecting the author's work.

All rights reserved. Without limiting the rights under copyright reserved above, no part of this publication may be reproduced, stored in or introduced into a retrieval system, or transmitted, in any form, or by any means (electronic, mechanical, photocopying, recording, or otherwise) without the prior written permission of both the copyright owner and the above publisher of this book.

(Formerly titled On the Design of Stable Systems)

Dear Reader: Even with many layers of editing, mistakes can slip through, alas. But, together, we can eradicate the nasty nuisances. If you encounter typos or errors in this book, please send them to me at: <[e-mail]> Thank you! —Jerry Weinberg

Contents

Preface to the E-book Edition

Original Preface

Chapter 1 • The Search for Regulation

Chapter 2 • The Homeostatic Heuristics

Chapter 3 • Other Regulatory Heuristics

Chapter 4 • Types of Regulatory Mechanisms

Chapter 5 • Regulation and Environment

Chapter 6 • When the Model Fails

Chapter 7 • Making Regulation Mysterious

Chapter 8 • Overly Simple Views of Regulation

Chapter 9 • Blindness and Reversed Vision

Chapter 10 •Epilogue

Bibliography

Appendix: Some Useful Mathematical Notation Explained

Further Reading

About the Cover

TO BEAUTIFUL JOE

Preface to the E-book Edition

With books, as with children, you never know what you've produced until they've grown up. When Jerry's book An Introduction to General Systems Thinking was first published, we imagined it would be used by practical systems designers in a variety of disciplines, not as a university text. To be sure, Dani used it in a graduate seminar in Anthropology, but she was biased. Then, after a few years, Jerry was surprised yet delighted to discover how many other courses were using it.

About the time of that discovery, we first published parent of this volume, under the title On the Design of Stable Systems. Looking at the original preface, it's easy to see that―perhaps influenced by the earlier companion volume―we thought we were producing an academic work. That impression undoubtedly influenced the title, leading us to the rather abstract and distant sounding On the Design of Stable Systems. (We actually received purchases from some horse owners who wanted to design stables for their mounts.) It also influenced the original publisher to promote the book solely as a text, where it was only moderately successful.

On the other hand, some very practical people managed to discover the book despite the publisher's promotion. We received many kind letters testifying to its practical usefulness—and complaining about the title. So, when the opportunity came to reprint the book, we sought to make two changes. First, we found a publisher who was more attuned to the people who design and build real systems in the real world. Second, we changed the title.

The new title, we hope, gives just the right impression of the book's contents and usefulness. It is a practical book about systems design―not just computer systems, but information systems in the larger sense: human organizations of all kinds, and systems in nature, like organisms, species, or forests.

With such a general scope, it obviously cannot be a nuts and bolts or "recipe book, yet it has many extremely practical applications to the daily work of people who design, for example, information processing systems, training programs, business organizations, parks, or cities.

The field of software development has recently become enamored of methodologies―integrated step-by-step approaches to developing and maintaining systems. Ours is not another methodology book―though the general principles we describe have an ancestral relationship to some of the popular methodologies. We believe they will make learning any methodology easier, because they are general principles―the sorts of things you need to know regardless of what methodology you use, or what kind of system you're designing or deciphering.

We considered using the title The Secrets of Systems Design, after Jerry's book The Secrets of Consulting, but we thought that would be confusing. It is about secrets, though―the deep thinking processes we use in our consulting and practice. These general ways of thinking about systems enable us to get quickly to the heart of existing organizations and information systems, to visualize new systems, and to design training and other interventions to catalyze the transformation from one to the other. Although we don't promise that such deep insights are easy, we've seen how much they can empower you. In the end, that's the only good reason for publishing a book, whatever the title.

Now, as we enter the e-book era, we've made some further changes to make the book more accessible. Based on reader feedback, we've come to realize we had written two books in one. Both books are about stability and design, but the second half of the original book was about stability through active regulation, which is more advanced than aggregate, or passive, regulation. We have separated out those final nine chapters for this volume: Active Regulation: General Systems Design Principles. We've done this to make each book more focused and less expensive. We hope they now better fits your needs as a reader.

Dani & Jerry Weinberg

Original Preface

This book is the result of an 18-year collaboration between two people, in two different disciplines, who share a fascination and love for the human animal. Whether from the vantage point of computers or anthropology, we are excited by the capacities of the human mind and alarmed by some of its products.

Our disciplines seem to begin at opposite poles―machine systems and social systems―but they converge as soon as people enter the picture. Thus the social scientist is concerned with the cultural meaning of math anxiety, while the computer scientist teaches programmers to overcome their people anxiety.

Both our disciplines daily come to grips with the subtle interplay between system and environment. Cultures and computers both exhibit the effects of adaptation to a constantly changing environment. And anthropologists and computer scientists equally balk at the difficulties of studying conservation and persistence.

This collaboration has been nourished, too, on another level. Our professional activities oscillate between the abstract and the concrete, the theoretical and the practical. The computer scientist designs program logic and enables people to work productively in teams. The anthropologist wrestles with concepts in the classroom and drinks wine with natives in the field. We are accustomed to switching modes, shifting communication styles, all the time. This instability in our professional lives has taught us to value uncertainty and to make a virtue of indeterminacy. General systems thinking has been, for us, a way of understanding the complexity of our own lives. We sincerely hope it will be the same for you.

This book is a companion volume to An Introduction to General Systems Thinking and Passive Regulation, which means they may be used separately, although they go well together. If you have already read An Introduction to General Systems Thinking and Passive Regulation, you definitely have all the prerequisites you need to read this volume. If you haven't read the companions, you may well find you will benefit most from reading this volume first, since you already have it.

If you run into any difficulty with the mathematical notation in this volume, try using the Appendix for guidance. If that doesn't solve the problem, and you still feel that the notation is seriously eroding your enjoyment of this volume, then it may be best to set this one aside and read the companions first.

But keep in mind that a major goal of these books is to reduce your anxiety about mathematics, and to open doors any mathematical deficiency may have closed. If the companion volumes are not easily at hand, try reading through this one, ignoring anything you consider too mathematical.

The Questions for Further Research at the end of each chapter should be read as part of the text. They give you an impression of the scope of problems to which the chapter materials might apply. Should some problem strike you as particularly intriguing, make note of it and then use the references to take it up later. Since the book is intended to introduce you to new ways of thinking, many quotations and references have been given―not to lend a patina of scholarship, but to give you numerous pointers toward other paths to learning.

Following the suggestions of several readers of the earlier volume, we've made a few changes to improve the readability of the book. Chapters have been made shorter and more numerous with more topical divisions to provide places for rest and contemplation. References have been moved to a consolidated bibliography at the end of the book to be less distracting to those reading for the pure pleasure of reading. Recommended and Suggested readings have been dropped, as have notational exercises, because most people surveyed either ignored them or found them perhaps patronizing. Those readers interested in suggestions for further reading will find numerous clues in the body of the text.

For classroom use, there is a great variety of options. We have used this material in a course of its own, and also in a sequel to a course based on An Introduction to General Systems Thinking. Most of the people who use the companion volumes as texts have pretty good ideas of their own about how to use this one. Many have followed suggestions given in the earlier volume, some of which bear repeating here.

This text and the previous ones have been used, to our knowledge, in courses in management science, computer science, anthropology, sociology, urban studies, metallurgy, medicine, architecture, psychology, theology, and philosophy. The texts are suitable for any level, from sophomores on up, if adjustments are made by assigning differing amounts of supplementary reading and questions for further research. The research questions are usually suitable for either a short essay or a term paper. In higher-level courses we have asked students to prepare one or more of these questions for class presentation.

The very flexibility of these books and the generality of their material make them difficult to set in a university curriculum, yet many professors have overcome these problems and achieved great success. Some of them have told us that our suggestions in An Introduction to General Systems Thinking were useful in getting started. If you arc contemplating using this book as a text, look there first.But if you need further advice, the authors would be more than pleased to hear from you and to contribute what they have learned from their own experiences and from contributions of others. In any case, if reading the book itself doesn't give you many ideas, then perhaps we're all in trouble.

Bibliographic Note

The material in this book, accumulated over the better part of two decades, comes from an incredible variety of sources. It has survived more than twenty moves, plus another dozen or so seasonal crossings of the Atlantic. It has been manipulated by more than ten editorial assistants, using at least five bibliographic systems. To top off all this possibility for error, a substantial chunk of the original bibliographic material was misplaced in the most recent move, five years ago.

As a result of these perturbations, the bibliographic material has not been entirely stable. When preparing this manuscript, we discovered the large loss and a number of smaller problems. In order to reconstruct the bibliography, we and our assistants undertook several hundred hours of detective work, most of which was successful. In the end, however, there remain bibliographic problems which we cannot seem to correct with any reasonable amount of effort.

For example, some of the references are from editions of books not available anywhere in Nebraska, and it has proved impossible to check (or even provide) all page numbers. Several references to articles have been almost completely lost, with only the author's name and perhaps a date remaining. Sometimes even that much is missing.

Rather than omit material for which we are unable to provide exact, checked references, we have decided to give what reference material we have and to beg our readers to help us, wherever possible, to improve this information for subsequent printings. We intend no discourtesy to the original authors, and certainly do not offer their original material as ours. We have included direct quotations, even in those few cases where we have lost track of the author. We felt that authors would prefer that their thoughts be spread widely rather than being omitted because of an unfortunate bibliographic accident. We especially hope that we will be able to fill in all the blanks before several printings have passed.

In all, the problems have affected perhaps 2 percent of the bibliographic items, and most of those are merely missing page numbers. We don't think the problem will in any way affect the general usefulness of the book, but we apologize for any inconvenience it may cause any reader. [Note added for the eBook edition: Many of these problems are easily solved today with the internet and its search engines.]

Acknowledgments

Many people have contributed in many ways to the publication of this book. We wish to express our special thanks to the following friends who have been closest to the project over the years: Sheila Abend, Lorri Campbell, Jim Greenwood, Mike Gunn, Linda Hollcroft, Joan Kaufmann, Shanna McGoff, Cheryl Plum, Mona Thompson, Melissa Weiksnar. Encouraging, questioning, searching out literature, criticizing, typing, preparing the manuscript―in all these ways, their help has been invaluable. We also owe an important debt to our many students, and to the readers of the previous volume who took the time to send us their comments.

Gerald M. Weinberg Daniela Weinberg, Eagle, Nebraska

Chapter 1 • The Search for Regulation

No one will quarrel with the assertion that social existence is controlled existence, for we all accept a certain basic assumption about human nature—namely. that without some constraint of individual leanings the coordination of action and regularity of conduct which turn a human aggregation into a society could not materialize. (Nadel 1968:401)

A pure aggregate strategy may be all right for auks, aubergines, or auriferous rock, but human societies from the australopithecines to the Austro-Hungarian Empire have some coordination of action and regularity of conduct. No person who has felt the constraint of individual leanings can lack curiosity about the deep reasons and strategies behind the regulation necessary for survival―regulation above the level of a mere summation of individual activities. In this chapter, we begin to move away from description of aggregate behavior toward understanding some of the structured strategies built upon them.

The Problem of Multidimensional Regulation

The study of equilibria will always be important in the treatment of systems of high complexity, for the equilibria, in their various forms, are those states, or sets of states, in which the system's behavior no longer depends to a major degree on the time. By effectively losing a variable, the functional relation becomes simpler; and the change may reduce the impossibly complex to the manageable. (Ashby 1964:95)

Concealed by our impression of a relatively stable, structured world is the unceasing regulatory activity of aggregates. Aggregate survival is the most elementary form of regulation, both in the sense of the simplicity with which it may be understood and modeled and in the sense that it is the foundation on which other regulatory mechanisms are built. As Ashby indicates, where there is such a great complexity in the environment, it is advantageous to be able to lose a variable, so that the problem of regulation can be made tractable. Relying on aggregates, then, is one of the simplest ways of losing variables.

But why should losing a variable be such a helpful action for a system to take? Engineers are well aware, or should be, of the simplifying effect of removing one of the constraints from a problem. Engineers are annoyed by the introduction of such non-engineering concepts as beauty, not so much because they are against beauty but because people asking for beauty usually think beauty is something that can be painted onto the otherwise strictly engineering solution design. They do not realize that for a truly beautiful bridge, beauty must have been an integral part of the design―and the cost estimates―from the beginning.

To see how the addition of dimensions affects the design problem, let us consider a very hypothetical design problem given to a lighting engineer. The present street lighting system in our neighborhood has proved to be unsatisfactory because the young girls living there have recently acquired BB guns and like to use the lights as targets. The village cannot replace bulbs fast enough to keep up with the girls, and because the United States Constitution guarantees the right of the populace to bear arms, it has proved impossible to legislate a prohibition.

The village elders come to the lighting engineer for help, and she begins by listing the various forces in the environment affecting the lifetime of a streetlight. For simplicity of our argument, suppose the list contains only two items:

1. The BB guns.

2. Heat generated by the bulbs themselves.

The engineer then finds out that the effects of each of these two disturbances depend on the thickness of the individual bulb, T. A thin bulb dissipates heat quickly and is thus better able to survive the destructive effect of its own internal heat. A thick bulb, on the other hand, has a better chance of surviving a hit from a BB. Moreover, there is a limit to the thinness of the bulb, below which it will shatter spontaneously. Similarly, there is a limit to the thickness, above which it will transmit no light.

The problem facing the engineer is to design a system of lighting supplying a certain minimum amount of light over a certain minimum time―long enough, at least, so replacements can be made on a reasonable schedule. As in all engineering problems, however, there is one more constraint—cost. If the village elders are willing to put up enough cash, the engineer can simply employ a pure aggregate strategy and design a very large light stanchion containing thousands of bulbs. Given that the girls cannot afford infinite numbers of BBs, nor infinite time to be shooting at streetlights, such a pure aggregate solution would guarantee a minimum of light for some minimum time.

With a cost factor, however, the lighting engineer is constrained to some maximum number of bulbs per stanchion. This we can call N0. Just to keep our problem within the bounds necessary for exposition, let us assume that there are no better bulbs available and that the engineer has no light-transmitting material available other than the bulbs themselves. Thus she can cluster the bulbs in any way she likes, but she cannot introduce other materials into her solution.

Given that she