The Mythical Man-Month

Book Card  -  Volume 13  -  Book Review Number 4  -  Fri, Mar 23, 1990 11:03 PM

TITLE: The Mythical Man-Month
AUTHOR: Frederick P. Brooks, Jr.
PUBLISHER: Addison-Wesley

In this industry computers last a few years, maybe, before they are obsoleted. Software is even more volatile; successful programs will see significant updates once or twice a year. So a book about software that has lasted more than a decade is both a curiosity and an anomaly.

Such a book is "The Mythical Man-month", subtitled "Essays on Software Engineering".

In addition to their fleeting existence, computer books typically share another common trait: they are incredibly boring. Not so this book. Consider, for example, a few chapter titles:

Chapter 1The Tar Pit
Chapter 7Why Did the Tower of Babel Fail?
Chapter 9Ten Pounds in a Five-Pound Sack
Chapter 14Hatching a Catastrophe

I suppose I should admit to one of the main reasons that I like this book: finally, I've found someone in the computer industry who knows how to quote (Answer a question, quick: Name ten, or five, or three programmers who even know how to write). I could take - we all could take - lessons from this fellow on how to use the written word. Making technical information readable, interesting, understandable: now THAT'S a challenge.

"The Mythical Man-Month" is about software engineering, about managing large computer programming projects. Brooks knows whereof he speaks: he managed the development of OS/360 back in the mid-60's. OS/360 is the operating system for IBM's System/360 computer line, and is still spoken of with reverence in some circles. OS/360 was completed behind schedule and over budget; it required about 5000 man-YEARS to develop. When it was all over, Brooks took time out to reflect on what went wrong, and why. What he learned filled this book. The following paragraphs are, perhaps, representative:

"The second fallacious thought mode is expressed in the very unit of effort used in estimating and scheduling: the man-month. Cost does indeed vary as the product of the number of men and the number of months. Progress does not. Hence the man-month as a unit for measuring the size of a job is a dangerous and deceptive myth [Sentence italicized in original -PRN]. It implies that men and months are interchangable.

"Men and months are interchangable commodities only when a task can be partitioned among many workers with no communication among them. This is true of reaping wheat or picking cotton; it is not even approximately true of systems programming.

Bottom line recommendation: If you develop computer software, read this book. If you manage people who develop computer software, read this book. And if you work with people who (manage people who) develop computer software...