Tracy Kidder spent 1978 and 1979 with a team at Data General in Westborough, Massachusetts, watching them build a 32-bit minicomputer in a basement. The machine was code-named Eagle, and it existed because Digital had shipped the VAX and Data General had nothing to answer it with. The book came out in 1981 and won the Pulitzer.

That’s the setup, but it doesn’t really explain why the book is as good as it is.

A non-technical writer in a hardware lab

The thing that makes it work is that Kidder isn’t a technical person and doesn’t pretend to be. Reading it, it’s fairly obvious that he doesn’t understand most of the terms being thrown around him, or what is actually happening on the boards in front of him. Normally that would sink a book like this one.

What he does instead is go after the only thing he can properly observe, which is the people, and it turns out he is exceptionally good at that. He captures what makes these engineers tick without ever really explaining what they are building. It’s obviously well researched, but the research isn’t the point. The point is how it felt to be in that room. You come away feeling like you were there, and that isn’t something I get to say about many books.

Part of you is in that machine

I only marked two passages in the whole book. This is one of them, an engineer on the team talking about recognition:

“Look, I don’t have to get official recognition for anything I do. Ninety-eight percent of the thrill comes from knowing that the thing you designed works, and works almost the way you expected it would. If that happens, part of you is in that machine.”

I got bitten by this when I was ten years old. I wrote a piece of software and published it on a local BBS, and then people started using it, and then they started sending me feature requests. That was more or less that. Whatever that feeling is, I’m still riding it with every new product I work on, and it hasn’t worn off in the slightest.

I suspect that’s the actual reason this book has stayed in print for forty-five years. Almost everyone who builds things recognises that sentence.

The conditions under which it was built

The other passage I marked is Kidder on Ruskin:

In The Nature of Gothic, John Ruskin decries the tendency of the industrial age to fragment work into tasks so trivial that they are fit to be performed only by the equivalent of slave labor. […] What is usually meant by the term craftsmanship is the production of things of high quality; Ruskin makes the crucial point that a thing may also be judged according to the conditions under which it was built.

Kidder writes Tom West, who ran the project, in decidedly romantic terms, and it’s very hard not to fall for it. He is something close to the ideal of a senior engineering manager. He shields the team, he absorbs the risk, he actually understands the machine, and he keeps everyone pointed at a target most of the company doesn’t believe in.

At the same time, and more so as the book went on, his approach struck me as badly outdated. The Eagle team worked punishing hours. People signed up for it in a way that consumed them, and a number of them left not long after the machine shipped.

I run engineering teams now, and I’d like to think the way I do it is more humane, more empathetic and more sustainable than what is in this book. No crunch time. No weekend work. Everything treated as a marathon, with nobody burning out at the end of it. Launches and deployments that are deliberately boring and have no drama attached to them. I’m proud of that, and I think it produces better software as well as better lives. I’ll still admit that reading Kidder I caught myself feeling the pull of the other thing, which is worth being honest about.

Which is where Ruskin comes back in. The Eagle was a genuine achievement. The conditions under which it was built were not good. Ruskin’s point is that you’re allowed to hold both of those at once, and that the second one counts.

Monozukuri, again

A few years ago, after a trip to Japan, I wrote about kaizen and monozukuri and what they mean for building products. I described monozukuri as craftsmanship, and specifically as the idea that the craftsperson puts a piece of themselves into everything they create.

That is almost word for word what the Data General engineer says. Part of you is in that machine.

Ruskin’s cathedral builders, a Japanese craftsperson and a microcode engineer in a Massachusetts basement in 1980 are all saying the same thing: good work has some of the worker in it, and you can’t get that by extracting it from people. That’s something close to what I have been circling around on this site for years, and I didn’t expect to find it set out this clearly in a book about a minicomputer.

All quotes are from The Soul of a New Machine by Tracy Kidder.