The Soul of a New Machine

A must read for software engineers.
I gained so many gems from this book and made me reflect on my own career in software.

It's made me reflect on questions such as
- why did my manager act that way?
- why are we lacking X resource?
- why am I motivated even though it's not in my best interest to be?


Computers are so much more than the sum of their parts. Although I've not been fortunate to work on a project like Eagle, I resonate heavily with the feeling expressed.

The book is an immensely well written, humorous and thoughtful account of the Data general company. I'll be re-reading it in years to come.

About 6 years ago, a sort of scandal rocked the gaming industry related to a blog post by a woman known as “EASpouse”. The blog post criticized EA's labor practices at the time, which required employees to work massive amounts of unpaid overtime, as they were salaried employees. By massive, I mean about 12-16 hour days, 6 days a week, regularly. This was a big deal among gamers, because very few of us had ever had the opportunity to peek behind the curtain like this. It was likely that most of us viewed game development with a variation of the way that Roald Dahl as a child imagined the inside of the Cadbury Chocolate Factory near the boarding school he attended (which later led to Charlie & the Chocolate Factory).

The Soul of a New Machine, by Tracy Kidder shows that such working conditions are nothing new. The book follows the development process of Data General's micro-computer (sort of like a rack mounted server, except it's the size of the whole unit, but essentially only being one of the server nodes), that would be a successor to their Eclipse line of microcomputers, code named the Eagle, and later released as the MV/8000. The book goes into both the personal and technical aspects of the development process, profiling the various men (and a few women) involved in the project, and giving a description of the technical aspects of the process for the layman.

While the technical bits (pardon the pun), are enjoyable, the book's strength, and where it spends most of its time, is in profiles of the people. The book paints a bleak picture of the inner workings of Data General. The working conditions at Data General, particularly on this project, are brutal. Much as with EA Spouse, employees are salaried, with no overtime pay, and work 12-16 hour days, 6 days a week. As the project goes on, project leads and younger employees are worn down. Often, employees at Data General observe that the company brings in a lot of new fresh recruits, and few stay at the company after they turn 30. Many of these new recruits drop out for various reasons, and often employees discuss the company's sweat-shop like working conditions. As the project moves into the heat of summer, the air conditioning breaks, turning their windowless basement office into a sweltering oven, which they can't even leave the door open for, for security reasons. Only after the employees strike do they fix the air conditioning.

By the end of the book, several of the project leads, themselves burned out, leave the company, and while some of the employees on the Eagle team stay on, many more have left.

Tracy Kidder got an impressive amount of access at Data General when he wrote this book, and while he's honest and truthful about what happened there, Data General, at least to my 21st century mind, comes out of this book smelling like shit. I base this solely on what Data General does, and I know this because Kidder doesn't whitewash - he thankfully calls it right down the middle.

While the book is never accusatory, it makes clear that Data General is a predatory employer. It preys on young, semi-idealistic college Engineering graduates, who don't have a lot of job experience and are looking more for interesting problems to solve, interesting work to do, than a big paycheck. They promise them interesting problems, and briefly, very briefly, warn them that there will be long hours and possibly a limited social life, that this job will become their life. To meet the deadlines required of them they will have to give up friends, family, and the outside world, living only the job, for months or years at a time. Plus, because they're salaried, despite all the hours they get that would be overtime, they're only making their standard pay grade.

It chews up 22-24 year old kids, and spits them out at 30, burnouts who had great potential, but were consumed by their jobs. They don't say if many of these former employees stay in the industry, and some certainly do - Ray Ozzie, creator of Lotus Notes and current Chief Software Architect at Microsoft is a Data General veteran. However, those who leave the industry with a sour taste in their mouth will probably leave worse off then they would be if they worked somewhere else. Had they been actually paid overtime, they could have possibly built a nest egg that could have allowed them to retire early, or to at least take their time looking for work elsewhere.

While some poor decisions related to processor architecture helped to kill Data General right before the dawn of the 21st century, it is my suspicion that the boom in Silicon Valley may have inspired a brain drain. Nicer weather, a less oppressive corporate culture. For people who wanted more money, there was the change to come in on the ground floor of companies which had the potential to be worth millions and get significant stock options. For those who preferred challenge, they could face whole new challenges when designing new systems and new architectures at the new companies in the Valley.

In summary, the book is a high resolution snapshot of the early days of the computer industry, before the internet started to permeate our lives in subtle ways - computerized tax processing, credit cards, ATM machines, and so on, leading up to the more overt ways it would later find its way in - Bulletin Board Services, E-Mail, and finally, proper web pages. People interested in the history of the computer industry will certainly find this fascinating. People who don't care about the history of computing can still find something in the profiles of the people in this project, and how the project's process slowly wears them all down.

Een journalist die meer dan een jaar in de loopgraven zit, samen met de mensen die een nieuwe computer aan het maken zijn. Die de computer volgt van ontstaan tot oplevering, in een periode dat er nauwelijks sprake was van personal computers: eind de jaren 1970, toen IBM oppermachtig was en Data General de gevaarlijke en hippe New Kid on the Block was.

Maar nu zit Data General mogelijk zelf in de problemen: DEC's VAX is sneller en beter dan het beste dat Data General heeft. En dus beslissen ze om op een nieuwe plaats een volledig nieuwe afdeling op te richten om een fantastische nieuwe machine te maken die vele kere beter zal zijn dan de VAX (maar ook veel complexer). De achtergebleven mensen, die niet ana het nieuwste van het nieuwste werken, moeten ondertussen maar incrementele verbetering aan de bestaande Exlipse-lijn doen.

En dat ziet Tom West, niet zitten. Zijn meesterplan: de bestaande 16-bit Eclipse hermaken in een 32-bit-versie, die backwards compatible zal blijven. Als een soort Plan B, voor in het geval het niet zou lukken met de nieuwe machine.

Hij stelt een klein team samen van oude rotten, aangevuld met een hele stapel mensen die net afgestudeerd zijn, “zo jong dat ze nog niet weten wat onmogelijk is”, en past er de paddenstoel-magament-theorie op toe: “keep them in the dark, feed them shit, and watch them grow”.

Het boek slaagt er nooit in om Tom West helemaal te verklaren: hij blijft een mysterieus iemand, in zijn jeugd rebels, een muzikant, heel de wereld afgereisd voor het Smithsonian, en dan bij Data General terechtgekomen waar hij eerst zelf in de loopgraven zat maar redelijk snel ‘management' werd. Voor zover we zien moet hij totaal schizofreen geleefd hebben: op het werk droog en een imago van bullebak en tiran, maar buiten het werk een compleet andere persoon.

West usually drove out of Westborough fast after work. “I can't talk about the machine,” he said one evening, bent forward over the steering wheel. “I've gotta keep life and computers separate, or else I'm gonna go mad.”

This book tries too hard to do too much. It ultimately is the story of a computer being built with typical exploitative labor practices and worst-practice management that's a case study for what not to do for its readers. It also is a character story, or rather, many character stories where the author dramatizes interactions and extrapolates on motivation to serve their narrative. Finally, the author is present in the story and feels compelled to wax philosophic and glamorize the character leads from their interactions in a way that is distracting. If you got rid of the things I disliked, the story would be more focused and probably boring as hell. I guess that's why you slap "Based on true events" on a movie and people come away thinking they had an accurate history lesson.

Wow I loved this. It's 1979 and a rag-tag team of engineers is on their way of hacking out a new 32-bit minicomputer (then the size of a couple of fridges) in the basement the Data General in Massachusetts. And a journalist gets to watch. The dust jacket reads, Kidder wrote the book “with a reporter's eye, a novelist' heart, and a technician's understanding” and it's all that and more. It's about wire-wrapped prototype boards and chasing bugs, management styles and team motivation, burning out and finding new drives, vivid portraits of people full of excitement for building something new and innovative.