Bob Evans

Today, if you make a credit card purchase, make an airline reservation, sell some shares or claim on an insurance policy, the chances are that the computer behind the scenes will be an IBM mainframe, a descendant of the System/360 computer family introduced in 1964.

Bob Overton Evans, computer engineer and manager: born Grand Island, Nebraska 19 August 1927; engineer and manager, IBM 1951-84; partner, Hambrecht & Quist, and Technology Strategies and Alliances (later Rocket Ventures) 1984-2004; married 1950 Maria Bowman (three sons, one daughter); died Hillsborough, California 2 September 2004.

Today, if you make a credit card purchase, make an airline reservation, sell some shares or claim on an insurance policy, the chances are that the computer behind the scenes will be an IBM mainframe, a descendant of the System/360 computer family introduced in 1964.

Bob Evans "almost single-handedly persuaded management to abandon a less ambitious product plan for one that resulted in System/360", according to one IBM historian, Emerson Pugh. In 1985 Evans and his co-workers Fred Brooks and Eric Bloch received the National Medal of Technology at the White House for a development which was described as "revolutionising the industry".

Bob Overton Evans was born in Grand Island, Nebraska, in 1927. He graduated in Electrical Engineering from Iowa State College in 1949. While working for a utility company, he heard about IBM's "Defense Calculator", later to become the model 701, the company's first commercial computer system. He wrote to IBM asking for an interview and in September 1951 was hired as a junior engineer.

"Bo" Evans secured a reputation as a roving troubleshooter, fixing problems for early adopters of IBM's computers. He quickly advanced to more senior engineering positions, becoming manager of processing systems in 1959.

By 1960, although IBM had become the world's most successful computer manufacturer, its engineering strategy was in disarray. It was burdened with seven different models of computer, all of which required different software. Worse, there was a looming software crisis as the number of computers in existence began to outstrip the number of the programmers to write software for them.

In autumn 1961, IBM's president, Tom Watson Jnr, authorised a task force to determine its future computer strategy. A couple of months later, the task force recommended that IBM should scrap its entire line of computers, existing and under development, and replace them with a single unified range.

Amazingly, IBM's main board accepted the recommendation, which the Fortune journalist Tom Wise called "IBM's $5bn gamble", noting that "not even the Manhattan Project which produced the atomic bomb in World War II cost so much". Evans, now vice-president of development for IBM's Data Systems Division, was given company-wide responsibility for co-ordinating and troubleshooting the development. He imposed company-wide standards to ensure compatibility between all the thousands of subsystems, software programs, storage devices and peripheral equipment.

The new computer family - dubbed System/360, a name "betokening all points of the compass" - was launched on 7 April 1964. There were press conferences in 15 countries and New York journalists were shipped en masse to IBM's Poughkeepsie plant in a specially chartered train. System/360 was hugely successful, ensuring IBM's dominance of the industry for the next 25 years, during which it often enjoyed a 75 per cent market share.

In 1971, 10 years after the original System/360 decision, IBM set up another task group to review its computer products. The task group recommended that System/360 should itself be scrapped, and a five-year, $100m research programme was initiated for its replacement. In 1975, however, it became clear that so much software had been written for System/360 that it could not be replaced in the foreseeable future. Instead the technology was constantly improved, right up to the present day, while always maintaining software compatibility. In theory, and usually in practice, a program written for an IBM mainframe in the 1960s still works on IBM's current mainframe computers.

After the System/360 assignment, Evans became head of IBM's Federal Systems Division, where his assignments included support for the Apollo space programme. During his career he headed four of IBM's major divisions, including the Systems Communications Division in the 1970s where he oversaw IBM's Systems Network Architecture (SNA), a proprietary communications technology that was later overtaken by the Internet.

Evans left IBM in 1984 to become a general partner of the venture capital firm Hambrecht & Quist. In 1988 he became managing partner of a new H&Q subsidiary, Technology Strategies & Alliances (which later became Rocket Ventures), where he oversaw corporate turnarounds and was active in mergers and acquisitions. From 1981 until 1995 he was chief scientific adviser to Taiwan.

Martin Campbell-Kelly