The Soul of a New Machine

The story in brief

In 1978 Data General, a minicomputer company in Westborough, Massachusetts, faced a threat. Its rival, Digital Equipment Corporation, had released the VAX, a powerful 32-bit machine, and Data General had nothing to match it. The company’s main effort to respond was a project run from a new facility in North Carolina. In Westborough, a group led by an engineer named Tom West, partly sidelined, set out to build a 32-bit computer of its own, codenamed Eagle, compatible with the company’s existing machines.

The writer Tracy Kidder spent about eighteen months with the team. The Soul of a New Machine, published in 1981, tells the story of how roughly thirty engineers, many of them just out of college, designed and debugged the machine, which shipped in 1980 as the Eclipse MV/8000. The book won the Pulitzer Prize for General Nonfiction and the American Book Award, as the National Book Award was then called, in 1982.

The decisions that made it

Tom West decided to pursue the project even though it was not his company’s official priority. He framed it as a modest extension of an existing product, which kept it below the level of scrutiny that might have stopped it. The book describes how he used the company’s internal politics to protect the project, presenting it in ways that would not provoke opposition.

West decided to staff the team largely with new graduates. Experienced engineers would have been expensive and might have doubted that the schedule was possible. Young engineers, eager to prove themselves, did not know enough to think it impossible. Kidder describes the hiring interviews as a kind of seduction, in which West’s lieutenants made the work sound so difficult and important that candidates wanted in.

West also decided to keep his distance from the engineers, delegating much of the detailed management to his deputies and appearing cold and remote. Some team members resented it. Kidder suggests that it was partly deliberate, a way of protecting them from pressure above and forcing them to solve problems themselves.

The team chose to build the machine for compatibility with Data General’s earlier computers, accepting design constraints in exchange for a product the company’s existing customers could adopt. And they committed to a schedule that was, by most standards, unrealistic, which led to months of very long hours.

What the book sees in them

Kidder’s central observation is about motivation. The engineers were not working for money; most received little extra for their efforts. They were working, in the team’s phrase, because they had “signed up”, committed themselves to the project and its success. Kidder compares their reward to the game of pinball: if you win, you get to play again. The prize for building one machine is the chance to build the next.

He also shows the costs. Engineers worked through nights and weekends, marriages and health suffered, and some burned out. When the machine shipped, the company’s marketing took over, and the people who had built it felt forgotten. Kidder does not resolve the tension between the exhilaration and the exploitation. He shows both.

The book is also about the nature of technical work. Kidder describes debugging, microcode and hardware design in detail, and makes clear how much creativity and craft they require. The engineers’ pride in their work was as much about elegance as about speed.

Since it was published

The book became a classic of narrative non-fiction and one of the first widely read books about the computer industry. Its picture of engineering teams, with their heroic hours, devotion to a mission and suspicion of management, anticipated much of later technology culture, including start-ups that celebrate long hours. Data General itself was overtaken as the minicomputer industry was displaced by personal computers and workstations, and was acquired by EMC in 1999. Later writers have returned to the book in debates about crunch in software and video game development, asking whether the devotion it describes is a model or a warning. The site’s essay Every Programming Language Is a Theory of the Programmer touches on the culture of the engineers who build computers.

What it asks of someone deciding

The book asks what really motivates the people who build things. Money mattered less to the Eagle team than challenge, autonomy and the chance to prove themselves. Leaders who understand that can create extraordinary commitment. The site’s idea page Wanting What We Want to Want raises the related question of whether people’s commitment reflects what they want on reflection.

It asks what such commitment costs, and who pays. The same devotion that produced the machine damaged some of the people who gave it. A leader who benefits from that devotion has a responsibility to notice its price.

And it asks how a team can succeed without official support. West’s skill at keeping the project below the radar allowed it to survive. Many important products have been built in this way, by people who found a way to do what their organisation had not yet decided to do.