received a unified and stylish industrial design language, but also to documentation, public exhibits, and architecture. IBM hired Charles and Ray Eames, probably best remembered today for their iconic chairs, to produce one of the earliest exhibits on the history of computing. Its landmark buildings, which hoisted the firm’s logo like a flag in the leading cities of the free world, were designed by star architects such as Mies van der Rohe.

IBM’s greatness also rested on its commitment to science. For its research headquarters, in Yorktown Heights, IBM turned to industrial architect Eero Saarinen, responsible for such futuristically iconic structures as the Gateway Arch in St. Louis and the TWA terminal at JFK airport. The lab curves in an oval shape, glowing in the dark like a flying saucer. Into the 1950s IBM retained a hands-on, product-centered engineering culture, long after Bell Labs and General Electric hired scientists and built up centralized research and development centers. By the 1960s, however, its international network of research facilities set the standard for corporate commitment to research. Its researchers, envisioned by Saarinen as “tweedy pipe-smoking men,” enjoyed the enviable conjunction of university-like research facilities with IBM’s generous pay and benefits and freedom from teaching duties. The firm’s Nobel prizes came from basic research into fields such as superconductivity and electron microscopy. IBM’s willingness to fund basic science reflected the many possibilities for payback across the huge range of products it developed and manufactured, from semiconductors and core memories to disk drives, keyboards, punched cards, printers, and dictating machines.

 

Threats to Greatness

The most successful computing firms
of the 1970s and early-1980s targeted
niche markets where IBM’s core ar-
chitecture did not compete well. DEC
established a huge new market for
minicomputers, SDC and Cray target-
ed supercomputers, while Wang built
a thriving business around word pro-
cessing and office automation. IBM
offered credible projects in most of
ful of products and services that domi-
nate their respective niches. Backward
compatibility is more important than
innovation, as Microsoft showed when
turning Windows and Office products
into lucrative monopolies. Most tech
firms develop a platform as startups
and grow it steadily over many years,
defending it against threats from up-
start competitors.

By the 1970s IBM held a similarly dominant position in the business-oriented mainframe market, but only because it launched a massive research and development effort during the early 1960s. In April 1964 when IBM introduced its System/360 range it rendered its entire installed product base obsolete. Two years earlier IBM had reportedly budgeted $5 billion to develop the new range, twice its entire annual revenue. According to Watson Jr. the actual costs were so high that in 1965 the firm unexpectedly found itself just “a few weeks” from needing “emergency loans to meet the payroll.” (See “Father, Son & Co” in the Further Reading section.)

Before this IBM offered several incompatible ranges of computers for different markets and sizes of company. Customers increasingly expected to receive bundled with their machines an extensive collection of systems software packages, such as programming languages and operating systems. The cost of developing these tools across so many different platforms was rising. Incompatibility also threatened customer loyalty. Computers became obsolete every few years, but successful application programs could evolve over decades of use. Any computer upgrade that required customers to recode their applications was a chance for IBM’s competitors to steal an account away, particularly as the newly introduced COBOL language was intended to let users move applications between computers from different vendors.

It was this project that introduced the phrase “computer architecture,” to describe design features at a more abstract level than their implementation in a specific computer model.

System/360 imposed a common ma-
chine instruction set, word length,
character set, and peripheral inter-
face over an industry-spanning range
of machines. For example, smaller
machines implemented in micro-
code instructions that were handled
directly in hardware by the larger
ones. From smallest to largest, the
first wave of System/360 machines
differed by a factor of 500 in memory
capacity and a factor of approximately
20 in processor performance.

The machines were all supposed to run a single operating system, with the bigger ones loading more modules and turning on more features. That part of the plan didn’t work out so well. Fred Brooks’ classic study of the problems of OS/360, The Mythical Man Month, was a founding text of software engineering and helped win him an ACM A.M. Turing Award.

Despite these problems, IBM’s gamble paid off magnificently. IBM’s core platform evolved through countless new models over the next 50 years, though the 370, 380, and 390 ranges and on to today’s System z machines. These never dropped backward compatibility with the 360 range and IBM never lost its dominance of the mainframe market. Its most direct competitors of the 1960s, including General Electric, which was still a much larger firm, were unwilling to match this investment. Most exited the computer business over the next decade. IBM was dominant at home and abroad.

 

Supporting Great Design

For Watson Jr., greatness included a commitment to elegance in design and the finest in modern architecture. An IBM System 360 featured prominently in the recent TV show Mad Men, where its stylish complexity symbolized the rise of analytic approaches to advertising. Today the clean, confident design aesthetic of the 1950s and early 1960s is more popular than ever. Period houses, furniture, and consumer products sell for a premium. No company did more to popularize that aesthetic and bring it into the American mainstream than IBM.

As documented by John Harwood in his book The Interface: IBM and the Transformation of Corporate Design, 1945–1976, IBM’s design chief, Eliot Noyes, assembled one of the most influential teams in history. Their skills were applied not only to the firm’s computers and office products, which