Review of "The Soul of a New Machine, by Tracy Kidder." Atlantic-Little and Brown, 1981

Abstract

The biggest problem is staying on top of the literature. I subscribe to over a dozen periodicals and rely on our librar y for many others. Those, coupled with the books I read , severely restrict my " just-for-fun " reading. What a delightful surprise, then, to discover a "must-read" professiona l book climbing the bestseller list. Tracy Kidder's new book is must reading for anyone involved in project management. Kidder spent parts of 197 9 observing the small group of computer engineers who designed and developed the Data General Eclipse MV/8000 , a 32-bit "supermini" computer. The "Eagle," as it wa s code-named, had to be developed in record time if Dat a General was to remain competitive with vendors who ha d already shipped 32-bit machines. Kidder presents an intriguing view of Data General, its top and project-leve l managers and the tightly knit group of computer engineer s who built Eagle. Kidder thoughtfully examines the managerial , technical and ethical dimensions of the project. This is not project management as prescribed in textbooks . Estimates and schedules are set by management as impossible or nearly impossible targets, formal project-control tools are rarely employed, vertical communication is a variant of the mushroom theory and is never congratula-tory. But the impossible schedule is nearly met, and th e Eagle, a market success, is delivered before the competitio n can seriously erode Data General's market position. The key to success was not project control but motivation , motivation fostered in several ways. First, many of the project team members, "the Hardy Boys " and "the micro kids," were college recruits ; they didn't know "what' s supposed to be impossible," but they were told that the y were bright and being given an unusual opportunity to " get a machine out the door with their name on it ." Second, th e team was isolated, physically and organizationally, from th e rest of the company, but the members were frequently reminded of strong competition for resources from anothe r group working on what Data General considered to be a higher priority project. Third, the team members' jobs wer e to a large extent self-defined, and the group operated fairl y autonomously. Finally, intragroup competition was regularly kindled. These strategies contributed to the develop