『The Mythical Man-Month』要約

Frederick P. Brooks Jr.’s The Mythical Man-Month draws on his experience managing IBM’s OS/360 operating system project in the 1960s, one of the largest software efforts of its era. Its central argument, often called Brooks’s Law, holds that adding manpower to a late software project makes it later, because new people must be trained and communication paths multiply roughly as the square of the team size. The title essay explains why managers persist in this error: they treat months as interchangeable units, ignoring that men and months are not equivalent. Brooks introduces several durable concepts. Conceptual integrity argues that a system’s design should reflect one coherent vision rather than a committee’s compromises, and he suggests separating architecture from implementation. The surgical team model proposes a small, tightly organized group around a chief programmer, rather than many equal contributors. He describes the second-system effect, in which designers overload a successor system with features, and the tendency to schedule optimistically, dismissing the “man-month” as a dangerous myth. Other chapters address communication overhead, documentation such as the project workbook, and the role of a “chief programmer.” In the anniversary editions, Brooks revisits his arguments, notably conceding in “No Silver Bullet” that no single technique or technology would yield order-of-magnitude productivity gains within a decade. He identifies essential difficulties of software—complexity, conformity, changeability, and invisibility—as distinct from accidental ones. The book remains a foundational text in software engineering and project management, valued for its candid, experience-based treatment of why large programming projects fail.