Tracy Kidder manages to achieve in “The Soul Of A New Machine” what most writers of popular non-fiction would shy away from – narrate the story of a team of hardware-engineers racing against time to build a 32-bit mini-computer to maintain the relevance of their company in the crowded marketplace of the computer-industry of the late 1970s. It is a gripping story in computer history that would have been forgotten, were it not for this wonderful book.
It is the late 1970s and Data General is lagging behind its arch-rival DEC and its successful VAX line of mini-computers in bringing a next-generation 32-bit mini-computer to its current and prospective customers. Its aging Nova and Eclipse series of computers are not enough for the expanding needs of such customers. It decides to make a clean break with the secretive “Fountain Head Project” (FHP) computer, first sequestering the team of super-stars in a different office in the same city and then in another office in a different city altogether. Those who are not selected for this effort or who can not migrate to a different city are left to work on the legacy systems.
Fighting against these odds, one of the folks left behind manages to convince the leadership of the company to invest a little in expanding the legacy Eclipse series to a 32-bit architecture in a way that maintains backward-compatibility with the 16-bit software written by its customers. This charismatic yet eccentric leader, Tom West, then tries to assemble a team of engineers to build this system from scratch (code-named “Eagle”) in just a year. The engineers are understandably reluctant to work on what they consider to be a kludge of a system and is not a clear priority for the leadership. Tom manages to convince a few of them, but largely staffs the new team with fresh college-graduates who are willing to work a lot more for a lot less and who are happy to be working on something challenging right away that they perceive to be important to the company.
The unrealistic deadlines of the overly-ambitious project results in a death-march that takes a heavy toll on some of the team-members. They suffer burn-outs, ruined personal-relationships, health-issues, etc. However, most of them tell the author that they would gladly suffer it all over again for even less pay. They bring the new computer to life by pouring their own souls into it. The computer is finally released as the Eclipse MV/8000 after slipping its initial schedule by about six months. The competing FHP system on the other hand keeps slipping and is finally cancelled. The team does not, however, reap much rewards as a result of its heroic endeavor and in fact disbands shortly thereafter. Sad.
As a computer-programmer, I found the many tales of debugging in this book quite fascinating, especially the bugs that were hard to reproduce. The vicarious joy of seeing a system come to life after many long and arduous battles by the team makes this book a wonderful read. However, it is hard to overlook the clearly exploitative behavior of the leaders and unreasonable pressure on the team from having arbitrary deadlines imposed on them. In fact, Tom West proclaims to practise the “Mushroom Theory Of Management” (“keeping them in the dark, feeding them shit, and watching them grow”). Cringe.
That said, war-stories of people overcoming great odds to deliver something of quality makes for great reading. The author, a writer by profession, manages to convey hard technical ideas like instruction-pipelining, multi-level cache-coherence, memory-protection, etc. in a way that lay readers should be able to follow along to appreciate the challenges faced by the team. This is a remarkable achievement. The Pulitzer Prize the author received for this book is entirely justified, as far as I can tell. Bravo!