The Soul of a New Machine
Tracy Kidder
There are certain stories our world deems interesting to tell and to hear: stories of battle, stories about artists, stories about show business, stories about politics, stories about crime, and of course love stories. An author would have to be crazy to dedicate himself to the proposition that there is an engrossing story to be told about about building a house, or designing a digital computer.
Tracy Kidder is that author. His book House tells about the building of an ordinary family home in Amherst, Massachusetts. This book, The Soul of a New Machine, tells the story of the design of a digital computer at the minicomputer company Data General in the 1970s.
Before the 1970s a computer was a huge, honking machine, so expensive that an organization that needed to do computing, such as a university, would own exactly one mainframe computer, probably an IBM (AKA "the Evil Empire"), and use it for all campus computing. In the 1970s a new kind of computer came to be. These were called minicomputers, but they were only "mini" in comparison to an IBM mainframe. The most successful minicomputer company was Digital Equipment Corporation, known to its users as "DEC" (pronounced like "deck") . The company didn't like this and insisted that we call them "Digital", but "DEC" -- one syllable, "Digital", three -- it was a fight they were never gonna win.
Three engineers who left DEC started a new minicomputer company called Data General. They had some initial success with the Data General Nova, a minicomputer designed to undercut the DEC PDP-8. (PDP stood for "Programmable "Data Processor" -- that name was chosen to conceal the truth that the PDP machines were computers from university computer centers, who didn't want anyone but themselves doing computing on campus.) But by the mid-1970s DG minis were running a distant second to DEC's own improved product, the PDP-11, which dominated the minicomputer market. The PDP 11/20 I spent my undergrad years programming occupied two standard electronic racks, each a little bit bigger than an ordinary kitchen refrigerator, and stored data on removable magnetic disks with a capacity of 1 MB, each the size of a large serving platter. The computer had 48 kB of memory. As an undergrad I exercised much of my ingenuity in squeezing my programs and data into those 48 kB. The PDP-11/20 cost about $20,000 -- this at a time when $5,000 was a lot to pay for a new car.
The PDP-11 had a serious limitation -- it was a 16-bit machine. That meant that the largest chunk of memory you could deal with was 65536 bytes (= 2 multiplied by itself 16 times). The 16-bit limitation was built into the PDP-11 architecture in a pretty fundamental way. It was obvious to everyone that the Next Big Thing in minicomputers would be 32-bit minicomputers, which would be able to address 4 GB of memory (an amount that seemed almost infinite at the time). DEC set out to design a new 32-bit minicomputer that would be as much like the PDP-11 as possible -- this was eventually their VAX minicomputer line.
Data General knew that the VAX was coming onto the market. In order to survive, they would need to come up with a competitive 32-bit minicomputer. And they would have to do it fast, before DEC owned the market. The Soul of a New Machine is the story of that effort, led by Tom West. It is a genuinely exciting story, not just because it is well told by Kidder, but also because it was a high-stakes effort under the most severe sort of pressure. West hired a bunch of young still-wet-behind-the-ears engineers who had no experience with a project this complex. He pioneered "Move fast and break stuff." long before Mark Zuckerberg ever said that. For instance, he told his team "Not everything worth doing is worth doing well," or "If you can do a quick-and-dirty job and it works, do it."
In the short run, the project was a success. DG brought the Eclipse MV/8000 to market. In the long run, though, it was a failure. VAXen came to dominate the minicomputer market, and DG eventually was forced out of the minicomputer business. But, to be clear, The Soul of a New Machine is not a business story. It's an engineering story. If you don't believe that an engineering story can be exciting, maybe you should give it a try.
Comments
Post a Comment
Add a comment!