Some tech visionaries, such as Bill Gates or Larry Ellison, build empires out of the utilitarian. Less common are the Steve Jobs-types, who create stuff people wait overnight in lines to buy. Then there is the very small club of people who made those titans possible. A member in good standing is Frederick P. Brooks Jr., pioneer of modern software.
Brooks, a 79-year-old computer scientist and former IBM (IBM) executive, wrote an oddly named book 35 years ago called The Mythical Man-Month that laid out the organizing principles of how software gets made. "If a computer is a part of your life today," says Bill Buxton, principal scientist at Microsoft (MSFT) Research, "then you've been highly influenced by Fred Brooks and his work."
Now Brooks has a new book, The Design of Design, in which he takes his message beyond tech. It's a collection of essays that draws on his own adventures as well as conversations with designers of every species.
In the 1950s and '60s, Brooks ran product-development teams at IBM. The company would set budgets for projects based on "man-months." If 25 people required 10 months to write some code, that worked out to 250 man-months. If the project ran late, you just added people. The man-month was a convenient device that was, Brooks wrote, nonsense. Expanding a late-running design team slowed the process. It forced veterans to bring newbies up to speed right when they needed to hunker down. "It's a fundamental insight," says Andy van Dam, a professor of computer science at Brown University. "Fred updated that old saying that nine women can't make a baby in one month."
In 1961, IBM had him manage development of a new family of mainframes. About 1,000 people worked on System/360; Brooks insulated his core design team from the usual corporate bureaucracy. Unveiled in 1964, System/360 helped cement Big Blue's position as the world's biggest computer company. Man-Month, which documented the project, has sold 500,000 copies. Programmers still refer to Brooks' Law: "Adding manpower to a late project only makes it later."
With the new book, Brooks, now a professor at the University of North Carolina at Chapel Hill, shares more design tales from his IBM days and after. Brooks breaks down projects to assess what went right and wrong each time, and has suggestions for anyone who manages creatives. Sample tip: "Constraints are friends" that "shrink the designer's search space," a notion familiar to anyone who's ever felt paralyzed by open-ended essay questions. Another: Design is as much art as science. "Improving your process won't move you from good to great design," he says. "It'll move you from bad to average." And that's not how legends are made.
Father of the modern software business
Computer scientist, University of North Carolina
1999 Turing Award, computing's Nobel equivalent