MASSACHUSETTS INSTITUTE OF TECHNOLOGY BOOK REVIEW - THE ART OF SYSTEMS ARCHITECTING BY MARK W. MAIER AND EBERHARDT RECHTIN November 2000 Sebastian Fixson, [email protected] Technology, Management and Policy Program, MIT Course: ESD.83 Research Seminar in Engineering Systems Prof. David A. MindellBook Review: The Art of Systems Architecting - 1 - For the last 50 years engineering has attempted to separate itself from research approaches and methodological tools seen as non-quantitative, sometimes even non-scientific. Emphasizing rationality, engineering has focused on phenomena that can be measured, calculated and optimized. In contrast, art has been categorized as an endeavor beyond scientific scrutiny. Art involves always non-rational criteria like emotion, senses or pure instinct. Two events at the end of the twentieth century seem to have begun changing this perception of separation between engineering work and the social world with which the systems interact and in which they operate. First, the end of the cold war has dramatically altered the financing options for almost all large-scale systems, in particular military and space projects, enforcing a stronger focus on heterogeneous customer demands. Second, the rise of the Internet has reformulated questions of who actually plans, builds, and controls many of these of systems. Both phenomena demonstrate that in today’s world the success of any large system depends not only on its technical performance, but also on the level of acceptance that it achieves from different constituencies, i.e. customers, government, the general public, etc. Recognizing these developments and the fact that the most far reaching decisions are made in the very early stages of system conceptualizing, Mark W. Maier and Eberhardt Rechtin propose to reconcile ‘artistic’ approaches with ‘classical’ engineering. Their book, The Art of System Architecting, calls for allowing ‘fuzzy’ approaches to support the early system development steps, which they call architecting. Distinguishing between engineering as a deductive process and architecting as an inductive process, the former being more science than art, the latter being more art than science, their book builds a bridge to reconcile these two worlds. Even though, their own strong background in systems engineering in the aerospace arena shines through most of the examples and their view on public decision making processes shows a strong grounding in the engineering thought world.Book Review: The Art of Systems Architecting - 2 - The authors direct their book towards experienced engineers, others “concerned with creating, building, or using complex systems,” and various other constituents of graduate studies. Consequently, the authors have structured the book in a textbook style. The book’s content is grouped into four parts. The first lays the groundwork and positions the book, the second presents a number of different application domains for the architecting approach, the third introduces various tools helpful for architecting systems, and the fourth explains how democratic decision making processes work and looks where systems architecting stands as a profession. The two chapters of part one form the foundation for the remainder of the book. In chapter 1, the authors locate one of the major reasons why systems architecting becomes increasingly important in the growing complexity of the systems being engineered. They define complex as being “composed of interconnected or interwoven parts” and a system as “a set of different elements so connected or related as to perform a unique function not performable by the elements alone.” To expand the classical engineering view, Maier and Rechtin argue for a purpose orientation that goes beyond mere systems performance, but rather includes explicit value judgments: “a useful purpose, at affordable cost, and in acceptable period of time.” They see the architect as an “agent of the client,” that works “jointly with the client on problem and solution definition. System requirements are an output of architecting, not really an input.” To tackle these often ill-defined problems, the authors propose in the second chapter a list of heuristics as tools for the systems architect. Abstracting experience, humans can use heuristics as a means to transfer knowledge that cannot be expressed in mathematical equations or hard logic. Chapter two presents an extract of the authors’ twenty-year effort of collecting and sorting heuristics helpful for systems architects. (Appendix A offers an extended list of heuristics.) Maier and Rechting find four types widely applicable. All four remind the architect of theBook Review: The Art of Systems Architecting - 3 - limited capacity of his own understanding of the situation. In essence, they recommend allowing for the possibility that the original problem statement may not have been the best, or even the right one, and in turn, argue for system clarity, simplicity, and flexibility. In its five chapters, the second part of the book presents five application domains for systems architecting. Chapter 3 is devoted to the authors’ home field; what they call builder-architected systems. The term implies, that builder and architect are identical. This situation usually applies for large-scale high-technology projects like new weapon systems or space programs. Having used systems engineering approaches since WWII, this industry had to cope with shrinking government funding after the end of the cold war. Forced to compete in more open markets, the term customer focus has become more important, and with it a new definition of project success: customer value. A Higher, Faster, Farther approach had to be exchanged with a Better, Faster, Cheaper way of doing business.i This leads the authors also to remind the reader that a good architecture, in addition to good system performance, can have strategic value in market competition. Chapter 4 shifts the focus from developing a product-like system to operating a manufacturing system. Maier and Rechtin point out that four major innovations over the last twenty years have changed the architectural rules for manufacturing systems. Ultraquality requirements, dynamically run manufacturing systems, lean production, and flexible manufacturing systems have added tasks for the manufacturing systems architect such as