The Soul of a New Machine Review
A human story about making a machine
Tracy Kidder’s The Soul of a New Machine is about the people who make computers and the computers they make.
It was first published in 1981, long before everyone had a smartphone in their pocket. It was a far-off age, where thirty engineers in a basement could, over the course of a year, create the Eagle, a new computer that would go on to save the company they worked for.
This is their story.
The Soul of a New Machine is more concerned, at all times, with the people who built the machine than the machine they built.
We’re introduced to an eclectic group of characters throughout the chapters, each playing a role in the saga of the Data General corporation and the birth of the new computer.
“No mode bit.” (p. 41)
Edson De Castro was the CEO of Data General, having started the company at the end of the 1960s. His sole instruction in building the Eagle was to make it backwards compatible with the company’s existing 16-bit computer, allowing the current users to seamlessly run their existing software on the new machine (something that the competitors’ machines did not).
“We’re building what I thought we could get away with.” (p.48)
Tom West is the originator and the leader of the Eagle project. He’s described by others throughout the book:
“a good man in a storm” (p. 7)
"He's fearless, he's a great politician, he's arbitrary, sometimes he's ruthless." Alsing (p. 103)
And perhaps the highest praise a leader can be given:
"West never bored us." Alsing (p. 275)
West’s leadership is the subject of much prose throughout. Without him, the Eagle never would’ve been made.
“It was kind of like recruiting for a suicide mission. You're gonna die, but you're gonna die in glory.” (p. 66)
Carl Alsing was tall, dressed carelessly, and spoke softly. He followed West onto the project, and was the one to give it the name Eagle. When hiring new engineers, he’d tempt them with the prospect of truly designing their own computer, before hinting strongly that they’d be working long hours. The leader of the Microkids, he’d tell people that they were only hiring the best before hiring them.
“Anytime you do anything on the sly, it’s always more interesting than if you do it up front.” (p. 83)
Steve Wallach was a slender man in cowboy boots, a computer architect who was tired of seeing his well-engineered designs not actually make it to production. He felt particularly slighted by Data General for his work on previous projects that had gone nowhere, and signed up to design the Eagle mostly out of spite. He would spend more time researching artful quotations to put on his technical documents than he did designing the computer’s memory.
"And I may not be the smartest designer in the world, a CPU giant, but I'm dumb enough to stick with it to the end." (p. 142)
Ed Rasala was big, with a firm handshake and fast speech. He was West’s trustiest lieutenant, in charge of the Hardy Boys and with them, Eagle’s physical manifestation.
The Secretary Extraordinaire
“I wanted to be part of that effort.” (p. 57)
Rosemarie Searle was the Eagle group’s main secretary, and perhaps more than anyone else embodied the idea of “signing up,” West’s term describing the attitude of “I want to do this job and I’ll give it my heart and soul.” (p. 63)
The Hardy Boys
“An oscilloscope is what cavemen used to debug fire.” - Jim Veres
Those who worked on the hardware of the new computer called themselves the Hardy Boys.
Ken Holberger - Called the Chief Sergeant Detective. (p. 189)
Jim Veres and Jim Guyer - The two Jims “make a marvelous debugging team, but only when they aren’t working together.” (p. 200)
Josh Rosen - Upon getting fed up debugging nanosecond-scale issues, left a note reading: “I'm going to a commune in Vermont and will deal with no unit of time shorter than a season.” (p. )
It takes people with devious minds. - Dave Keating (p. 156)
The Microkids, with their self-named microlounge and a van called the microbus, were responsible for the microcode - the incredibly low-level code that interfaces directly with the computer’s hardware. They delighted in pranking one another.
Dave Keating - Self-styled “devious mind.”
Chuck Holland - Hired because of a kinetic sculpture he was working on.
Dave Peck - A fast programmer with a raucous laugh, the “fastest in the East”. (p. 163)
Neal Firth - On liking abstract music in high school, said: “Maybe I'm right, maybe they are, but as far as I'm concerned they're entitled to their incorrect opinions.” (p. 168)
Jon Blau - When asked what he was getting out of working on the Eagle, he had this to say: “You have narrowed your field of vision to a small little world, trying to make it a vision of your own mind, and that's the kick - getting control of something.” (p. 160)
The story of The Birth of a New Machine is broadly about the Eagle project, a 32-bit computer that would commercially be called the Data General Eclipse MV/8000.
Setting the Stage
It was the end of the 1970s, and the computer business was booming. Data General, the company that had released the extremely successful 16-bit NOVA computer, was growing at an exponential rate, having joined the Fortune 500 at the end of fiscal year 1978. Profits were high, business was good, and there was only one problem - the future was refusing to come slowly.
Data General’s competitor, the Digital Equipment Corporation (DEC), had just come out with a 32-bit computer called the VAX 11/780. Without a 32-bit computer of their own to compete, Data General would quickly find themselves losing market share, as more and more customers moved to the better machine. Being behind technologically meant, in a very real sense, the death of the company; the market was not kind to any company slow to innovate.
Making a bet, Data General decided to invest in a research facility in North Carolina, a place with far lower taxes than the company’s birthplace and current location of Massachusetts. North Carolina became the place to be, the place where the future of the company would be built, and many engineers and staff moved there in a hurry to be a part of it.
This story, however, is about those left behind - who felt slighted and ignored.
This story is about people with something to prove.
Building a Computer
Tom West had already been burned once trying to compete with the folks in North Carolina, but decided to do so again, selling what would become the Eagle project as “insurance” to the company. After all, having a backup plan in case things went south was good business, was it not?
Of course, to the team he was putting together, he made it clear that nothing less than the future of the company was at stake.
Starting out, West had Alsing and Rosemarie and some senior engineers, but they wouldn’t be sufficient to meet the impossible deadlines West had set for the team in order to justify their existence to corporate. They needed more engineers.
The Eagle team itself was built differently from previous teams at Data General; instead of hiring experienced engineers from other companies, Alsing and West decided to hire inexperienced ‘kids’ fresh out of college. They’d come cheap, and work hard for the prospect of doing something they wouldn’t get to do anywhere else: build their own machine. As West said,
“There's some notion of control, it seems to me, that you can derive in a world full of confusion if you at least understand how things get put together.” (p. 175)
And that control was more valuable to potential new hires than the money they could make elsewhere.
Through an unorthodox hiring process West and Alsing soon gathered about 30 engineers, split into the Hardy Boys (who would design the hardware), led by Rasala, and the Microkids (who would design the microcode), led by Alsing.
Progress was made in fits and starts; Wallach designed the memory of the new computer, specifically how one user’s data would be protected from others. On previous projects Alsing had written large quantities of microcode in short bursts between long stretches of procrastination, but the microcode for Eagle was too much for him, and he focused on educating and leading his team. He’d play therapist and communicator more than engineer over the course of the project, learning to lead as his team learned to follow.
While the pressure was immense, the ‘mushroom management’ approach - keep them in the dark and feed them shit - was working. The new machine was starting to come together as the test models, named Coke and Gollum, were passing more and more of the diagnostics they’d need to pass in order to go to production. Along the way, flaky errors ranging from nanosecond-scale memory inconsistencies to loose wires had to be debugged, and more diagnostics written, tested, and inevitably diagnosed themselves when they failed.
And all of this happened under the ticking pressure of meeting an absurd deadline, the project dramatized by West until the engineers felt every moment that they had the cool metal of a gun to the backs of their heads.
In a reminder to himself and perfectionists everywhere, West wrote on his office whiteboard:
Not Everything Worth Doing Is Worth Doing Well. (p. 119)
The goal was to get Eagle out the door, to ship it. Not to build the perfect machine.
Deadlines slipped, because deadlines always slip, and each time they did West would ask Rasala for a schedule that would meet the new deadline, and each time Rasala would comply. He’d remark:
"The way to stay on schedule is to make another one." (p. 186)
The pressure, already high, became boiling when it became clear that the engineers in North Carolina, like those still in Massachusetts, weren’t going to meet their deadlines. All of a sudden, the project that West had sold to the company as ‘insurance’ became vital to Data General’s survival.
Originally due in April - West’s impossible promise to corporate in order to get the project approved - the date slipped, to May, to June, to July. The summer wore on, each bug in the Eagle found and removed more devilish than the last, the engineers flagging after a year of working long hours. Burnout was taking its toll. The last stretch was proving to be the hardest, the most draining, and the least exciting.
West himself grew thin and ragged as his subordinates toiled, fighting his own battles with the wider company to leave them the freedom to create the computer (almost) as they wished. He intervened in small but necessary ways, foreseeing issues with mass production or product integration long before everyone else and resolving them without anyone being the wiser.
It wound up being October when the Eagle team finally got the game Adventure running on their machine (incidentally it also passed all of the official diagnostics). Things moved quickly from there, aside from a scare with a particular kind of computer chip shortage, and before anyone knew it, it was done.
They’d gotten Eagle out the door.
The Microkids handed out microawards to each other. Rasala and the Hardy Boys debated whose machine it was they made - was it theirs? West’s? De Castro’s?
Who was truly responsible for Eagle’s creation and success?
Even more ambiguously, what would Eagle, now the Eclipse MV/8000, be used for? It was, after all, a general purpose computing device.
The engineers who made the computer argued over the first point, but seemed uninterested in the second.
Referring to building Eagle, West said:
It's for the kind of guy that likes to climb up mountains. (p. 181)
They had climbed the mountain, and all that was left to do was go home, get some sleep, and get started on finding the next mountain.
Kidder’s narrative lens informs the entire book. Every chapter begins with anecdotes from the characters’ lives or stories from Kidder’s research and investigation. He infuses scenes with descriptive life and people with quirky individuality.
His explanations of how computers work are well-woven into the story, sprinkled in a little at a time so as not to overwhelm. The only real drawback is that, as the book was published in 1981, some of the metaphors and references are dated, and occasionally the explanations go on for too long.
The book could be broadly divided into three sections:
The first section is an introduction to Tom West, Data General as a whole, and the problem the company was facing at the time.
The second section, taking up most of the book, involves building the new computer. It draws from a variety of different perspectives and is broken up by introductions to some of the various engineers at work, tracing the machine from an idea to an architecture to a kludgy, bug-riddled mess and finally out the door as a finished project.
The last section is shorter, and while it’s framed around the events that happened before, during, and shortly after Eagle is released to the world as the Eclipse MV/8000, it takes a meditative tone. Kidder ponders briefly about the role of computers in society and where all this progress is headed, including an amusing (in hindsight) mention of artificial intelligence.
Some things that stood out to me while reading the book:
Almost the entire engineering team on Eagle had been doing engineering of some sort since childhood, taking things apart and putting them back together, in a way reminiscent of how many geniuses spent their childhoods.
The architecture and specifications for the computer were determined almost organically, with Hardy Boys and Microkids going back and forth about what should be done in hardware and what in software; the design of the computer was the result of many negotiations over time.
All of the people who worked on Eagle worked ridiculous hours, often 60-80 per week or more, and they didn’t get paid overtime. Their motivations, while varied, were not monetary, and they worked absurdly hard. They felt responsible for the task, ownership of the product, and (for some) a sense that this was a worthy challenge.
When it comes to summarizing the book or otherwise answering the question of what it’s about, no simple answer or single word comes to mind.
The Soul of a New Machine is about computers, yes, but it focuses primarily on the people who make them, their struggles and frustrations and triumphs. It’s about humanity, in the broad sense, and about politics and management styles and the power and worth of leadership specifically. It’s about a group of individuals coming together as a team to accomplish something incredible - a herculean feat of engineering, in the time and with the resources they were given.
It reminds me personally of times I’ve embarked on a difficult endeavor with a team, and the sense of camaraderie and accomplishment that accompanies it. It captures the all-consuming nature of a project that is truly one’s own, and the sense of letdown (the book calls it postpartum depression) of finishing a great work and finding the vacuum it leaves in your life.
If you want to understand how computers work, this isn’t the worst book to read, although there exist many other resources better optimized for the task. On the other hand, if you want to read a story of a group of talented people racing against tight deadlines, struggling to give life to something more complicated than any could create on their own, that just happens to involve building a computer - I’d recommend checking it out.
While the engineers who worked on the Eagle felt empty after they’d finished it, eventually they’d move on to the next project, as do we all.
In the end, building the Eagle, as West says,
“…was a summer romance. But that's all right. Summer romances are some of the best things that ever happen.” (p. 287)
Affably Evil is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.