- Lundstrom, David E.
A Few Good Men from Univac.
Cambridge, MA: MIT Press, 1987. ISBN 0-262-12120-4.
-
The author joined UNIVAC in 1955 and led the testing of the
UNIVAC II which, unlike the UNIVAC I, was manufactured in
the St. Paul area. (This book uses “Univac” as the name
of the company and its computers; in my experience and in all
the documents in my collection, the name, originally an
acronym for “UNIVersal Automatic Computer”, was always written
in all capitals: “UNIVAC”; that is the convention I shall use
here.) He then worked on the development of the Navy Tactical
Data System (NTDS) shipboard computer, which was later commercialised
as the UNIVAC 490 real-time computer. The
UNIVAC
1107 also used the NTDS circuit design and I/O architecture. In
1963, like many UNIVAC alumni, Lundstrom crossed the river to join
Control Data, where he worked until retiring in 1985. At Control Data
he was responsible for peripherals, terminals, and airline
reservation system development. It was predictable but sad to
observe how Control Data, founded by a group of talented innovators
to escape the stifling self-destructive incompetence of UNIVAC
management, rapidly built up its own political hierarchy which chased
away its own best people, including Seymour Cray. It's as if at a
board meeting somebody said, “Hey, we're successful now! Let's build
a big office tower and fill it up with idiots and politicians to keep
the technical geniuses from getting anything done.” Lundstrom
provides an authentic view from the inside of the mainframe computer
business over a large part of its history. His observations about
why technology transfer usually fails and the destruction wreaked on
morale by incessant reorganisations and management shifts in
direction are worth pondering. Lundstrom's background is in
hardware. In chapter 13, before describing software, he cautions
that “Professional programmers are going to disagree violently with
what I say.” Well, this professional programmer certainly did, but
it's because most of what he goes on to say is simply
wrong. But that's a small wart on an excellent, insightful,
and thoroughly enjoyable book. This book is out of print; used
copies are generally available but tend to be expensive—you might
want to keep checking over a period of months as occasionally a
bargain will come around.
December 2004 