Managing the System Life Cycle is a treatise on managing systems development projects that use structured analysis, design, and programming techniques. Written by one of the luminaries of the structured school, the book gives me the feeling that the author is sharing his thoughts with the reader--manager to manager. Indeed, Yourdon shares ideas from his broad range of experience, makes direct suggestions to the reader, and demonstrates the well-developed sense of humor required of successful system developers.
The book assumes that the reader is managing a structured life cycle, that is, using structured tools such as data flow diagrams, entity relationship diagrams, and a data dictionary. Since the book focuses on project management, there is no discussion of actual structured techniques, although ample citations to the systems analysis and design literature are provided. The book concludes with a two-page bibliography.
This is the second edition. Yourdon wrote the first edition in 1980, before structured analysis techniques were well understood. In the preface, he points out that structured techniques now exist for all phases of the life cycle. The first edition emphasized the creation of four distinct models: the “current physical system,” the “current logical system,” the “new logical system,” and the “new physical system.” Yourdon states that, based on the real-world use of these methods in hundreds of projects, there is a danger in spending too much time and effort modeling the existing system--which is to be thrown away or replaced anyway. This edition instructs the analyst to concentrate on building a model of what the new system must do and recommends abandoning the strict sequencing of the phases of the life cycle in favor of overlapping the phases when possible. The author also takes aim at prototyping and voices concern that managers are condoning sloppy analysis and design by labeling their efforts “prototypes.”
The first three chapters provide a framework for the rest of the book. In chapter one, Yourdon states that projects using conventional techniques fail because of poor systems analysis, the absence of control over design and code, and bottom-up development and integration. In contrast, structured projects have meaningful paper models of the new system; visibility of the analysis, design, and programming efforts; objective criteria for judging “goodness” or quality; and an iterative approach to the activities in the system development process. Although I agree that structured approaches are generally superior to more conventional development methods, Yourdon has an almost religious reverence for the “structured specification.”
The second chapter is an overview of the structured techniques, including analysis, design, and programming, and of the structured documentation tools, including data flow diagrams, data dictionaries, entity-relationship diagrams, state transition diagrams, and process specifications. Together these make up the structured specification, which is partitioned, graphic, top-down in orientation, and implementation-dependent. Yourdon also provides an overview of several structured design components, namely, documentation techniques, design evaluation criteria (coupling and cohesion), design heuristics, and design strategies.
Chapter 3 discusses the structured life cycle, which consists of survey, analysis, design, implementation, acceptance test generation, quality assurance, procedure description, database conversion, and installation. To accommodate the overlapping of project phases, Yourdon suggests a range of management approaches, from the “ultra-conservative” one, where each phase is completed and approved before the next begins, to an “ultra-radical” approach, suitable for research and development projects.
Each of the following chapters focuses on a specific phase of the systems life cycle. For example, chapter 4 examines “Activity 1: Survey.” Each chapter starts with a dataflow diagram and an associated narrative. Definitions are given in data dictionary format, which is an interesting and efficient approach. A final section in each chapter focuses on problems and provides Yourdon’s ideas and suggestions to the reader.
Chapters 5 and 6 treat “Analysis” and “Design,” respectively. Yourdon lumps a good portion of the development activities into chapter 7, “Implementation.” At this point there is a noticeable diminution in the power of the book. Although the author would probably encourage prudent managers to read beyond this volume, in this and subsequent chapters he leaves the reader little choice. Chapter 8 discusses “Acceptance Test Generation” and chapter 9 looks at “Final Activities.” Chapter 10 contains Yourdon’s “Final Observations” on the problems of managing the stuctured life cycle.
This is a very concise and focused book. Thus, much is necessarily left out. In the preface, and again in chapter 1, Yourdon identifies automated tools and workstations as one of the major advances since the first edition of the book. Unfortunately, the rest of the book contains a total of two paragraphs and two footnotes on this important subject. Analysts and designers are increasingly turning to workbenches to assist them in using structured techniques, particularly to maintain the data dictionary, to create and maintain dataflow and entity-relationship diagrams, and to perform analysis on the various entities that comprise the structured specification. After citing estimates that 50 percent of professional analysts will be using such tools by 1995, he should have given the reader more information and encouragement on CASE tools and workbenches.
Yourdon is certainly a true believer in the structured methodology. His arguments are often tautological. The product of structured analysis is the structured specification used to guide the structured design and structured programming. If, however, the structured programs do not work right, then the structured specification must be wrong because the analyst did not do the structured analysis correctly (or forgot to say structured prayers).
In summary, this is a short, well-written, enjoyable book aimed at managers of systems development projects. Indeed, anyone engaged in managing any phase of a systems development project should find it worthwhile reading. It raises an important question: If Yourdon is a guru, and he thinks structured is the only game in town, then why are we not using structured methods in our organizations? At a time when many organizations are rethinking their approach to computing and are trying to manage information as an organizational resource, this book should be required reading.
And may the structured force be with you]
