The Soul of a New Machine by Tracy Kidder

It is entirely possible that any reader of this review has in their pocket a computing device more powerful than the one whose design and initial construction are the story of The Soul of a New Machine by Tracy Kidder. That machine, code named “Eagle” in the book, would eventually be sold by Data General as the MV/8000. It was a big as a good-sized kitchen appliance, and needed a separate television screen (definitely a cathode ray tube screen at the time) to manage input and some forms of output. Why read a book that is nearly 40 years old about a device that’s hilariously outdated in a field that is notoriously fast-moving?

First, Kidder brings the process to life. Someone far from the business might think that the design of a new computer is a bloodless affair. Kidder shows the decisions, the drama, the fights, the passion and the humor that go into the construction of a machine that, at the time, aimed to set the standards of what engineering and science could do within the constraints involved in creating a minicomputer. Sometimes the engineers had disinterested discussions about the best way to address a question; at other times, the choices they made were the result of knock-down drag-out arguments about fundamental views on how people could and should work. There was even the occasional bit of skulduggery.

Second, Kidder shows people giving their all for a project that they believe matters. They worked crazy hours, they worked in shifts, in their off hours they thought about the problems it raised, they offered their creativity, they pushed each other, they stretched their capabilities. “The Eclipse Group and the many others who had worked on the machine — including, especially, Software and Diagnostics — had created 4096 lines of microcode, which fit into a volume about eight inches thick; diagnostic programs amounting to thousands of lines of code; over 200,000 lines of system software; several hundred pages of flow charts; about 240 pages of schematics; hundreds and hundreds of engineering changes from the debugging; twenty hours of videotape to describe the new machine; and now a couple of functioning computers in blue-and-white cases, plus orders for many more on the way. Already, you could see that the engineers who had participated fully would be looking back on this experience a long time hence. It would be something unforgettable in their working lives.” (p. 276)

Kidder captures the mix of old and young engineers working together, the excitement of making something that has never existed before. Time and again, the younger engineers marveled about the amount of responsibility that they were being given, how they were thrilled by the challenge and the sense that what they were doing mattered. They contrasted Data General to stodgier companies like IBM or to government departments; in either, they believed, it would have been many years before they would have had opportunities like the ones that Eagle offered them. The older ones had experience enough to know where common errors would arise, to appreciate the interpersonal aspects of engineering tasks. Some of the older ones were also cynical enough to think that the younger ones would work longer for less.

Third, The Soul of a New Machine is worth reading forty years later because many of the characteristics of Data General are still present in today’s computer and software companies. The men who started Data General left DEC, another computer manufacturer. “Did they quit because, after long and heartfelt labor on a new design, they found that DEC’s management would not build their new machine? … [D]id they design this new machine after they seceded, or had they done that job in secret, using DEC’s facilities, while still on DEC’s payroll?” (p. 16) They took care of the business side, making their lawyer invest in the company so he wouldn’t be eager to walk away (the lawyer, in turn, insisted that the founders sell some of their stock in the second public offering, so they wouldn’t have everything tied up in the company and be afraid of losing it all), and pioneered the use of stock options as compensation for employees. An ad about wanting to make a lot of money “said what many others presumably were thinking, but what none of them felt they should say publicly. For some years thereafter, most of Data General’s advertisements contained something brazen.” (p. 20) The company’s headquarters was ostentatiously plain; they took pride in being underdogs; and they pushed limits to see what they could get away with. Such a culture can produce great things — and the Eagle was, for its time, great — but it can also easily shade into a toxic workplace, or indeed a toxic industry. Tales of abusive behavior at the likes of Facebook have parallels at Data General; tech bros have been around for a long time.

Some things change even less: “One Microkid started talking about a model of computer that was ten years old. He said, ‘They’re really ancient.'” (p. 155) Other things seem almost archeological: “From a cabinet in a corner, Rosen gets an object almost exactly the size and shape of a 45-rpm record and he inserts this ‘floppy disk’ into the tall disk-drive machine that stands nearby.” (p. 212) Good luck finding either a 45-rpm record or a floppy disk these day, at least outside of a collectors’ market, You’re pretty old if you even know what they are. More archeology at a trade fair once Eagle has landed as a prototype and is about to take off as a product: “We wandered past array processors performing tricks in FORTRAN, past tape drives, Winchester disks, printers and consoles in sleek-looking packages, and the latest ‘modems’ for hooking up computers to telephones.” (p. 239) I once learned a trick or two in FORTRAN; as I write this review on a machine that is slightly archaic for having a disk drive (as opposed to solid-state storage), I am using a machine more powerful than the Eagle as a paperweight to hold the book open at the pages with the quotations that I want.

The Soul of a New Machine also shows the spirit of engineering, the specter of business, and the vitality of creation. Eventually, the machine no longer belonged to the engineers who made it, but to the company that planned to sell it and make lots of money, and then to the customers who put it to their various uses. Even later, it became as ancient as the devices its builders cut their teeth on before taking on the challenge of making something new. And then it passed out of service. Swashbuckling Data General eventually went out of business too; it was acquired in 1999 by a company that later became a part of Dell. Stodgy IBM is still around. The computing world is not what it was, and yet the young wizards of Kidder’s book are still recognizable today, hoping to make something great; it’s a tale that will never be obsolete.