Bob Bemer

Computer pundit and 'father of ASCII'

Bob Bemer has a distinctive place in computer history. He did not make any lasting scientific contributions, nor was he a captain of industry. Instead, he was a computer pundit with a flair for communication and a talent for spotting technical trends and issues that put him at the centre of several developments that shaped the world of computing.

Robert William Bemer, mathematician and computer programmer: born Sault Sainte Marie, Michigan 8 February 1920; six times married (three sons, three daughters, one stepson, one stepdaughter); died Possum Kingdom Lake, Texas 22 June 2004.

Bob Bemer has a distinctive place in computer history. He did not make any lasting scientific contributions, nor was he a captain of industry. Instead, he was a computer pundit with a flair for communication and a talent for spotting technical trends and issues that put him at the centre of several developments that shaped the world of computing.

His most celebrated observation, made in the 1970s, concerned what later came to be called the "millennium bug" or the Y2K problem. Bemer realised, probably before anyone else, that problems lay in store for computer programs in the year 2000 because of the custom of using only two digits to record a calendar year.

Robert W. Bemer was born in Sault Sainte Marie, Michigan, in 1920. He graduated in mathematics from Albion College, a Michigan liberal arts school, in 1940. This was followed by a course in aero engineering at the Curtis Wright Technical Institute, Santa Monica. In the early 1950s the West Coast aerospace industry was a hotbed of numerical computation and soon became the computer industry's biggest customer. Bemer worked as a mathematician and programmer for aircraft firms that included Douglas, Lockheed and Marquardt, as well as for the Rand Corporation, the US Air Force's think-tank.

In 1956 he joined IBM's programming research group in New York. At that time "automatic programming" was the talk of the computer industry, and Bemer created one of IBM's first programming systems. This did not result in a significant product, but it led to Bemer's leadership of IBM's flagship programming system for data processing, Comtran (short for Commercial Translator).

Most other computer manufacturers were also devising such systems, and, in order to prevent a proliferation of programming dialects, the US government established a committee to design a common language. Bemer was at the heart of this committee, introducing several features he had designed for Comtran, and he was also responsible for the name of the new programming system, Cobol (Common Business-Oriented Language). Cobol was to be the lingua franca of commercial programming for the next 25 years.

In the early 1960s business computers began to use online terminals, for example at airport check-in desks. There was no standard code for these terminals, and as IBM's director of programming standards Bemer was in a position to spearhead a committee of the American National Standards Institute to establish a common standard. Known as ASCII (American Standard Code for Information Interchange, pronounced "as-key"), the code is still the most widely used code in computing, including on the World Wide Web. Bemer did not suffer from a retiring nature, and his motor car bore a vanity licence plate, "ASCII", to which, in case anyone should be in doubt, he added the legend "Yes, I am the father of ASCII".

From 1965 Bemer occupied a succession of senior software management positions with Univac, General Electric, and Honeywell (when it acquired General Electric's computer division). At that time the so-called software crisis was the most pressing issue in the world of computing, almost every big programming project having experienced time and cost overruns. In 1968, the academic and industrial software communities held a conference in Garmisch, Germany, to study the problem. Bemer proposed an ambitious solution that he termed the "software factory". While his software-factory project went unrealised even in the firms he worked for, the concept was a recurring theme in software debates for the next 20 years.

Bemer somewhat overreached himself by trying to get the US government to declare 1970 "the year of the computer" to highlight the year-2000 problem. His proposals made no headway, leading him to publish his first warning about the millennium bug in 1971. He restated the arguments again in 1979 in an article "Time and the Computer" in Interface Age, a widely circulated industry magazine. However, its alarmist tone seemingly went unheeded.

A larger-than-life character, who married six times and had many children, Bemer retired in 1982. He formed a small software company, and moved to a spectacular cliff-top house overlooking a reservoir, some 120 miles west of Dallas. In the late 1990s, the world finally woke up to the millennium bug, and Bemer was flattered with several television appearances, and was the subject of newspaper profiles. His software company also experienced an upturn in business, enabling him to sell out to a larger concern, which gave him the title of chief scientist.

Historians have not yet unravelled the true story of the millennium bug, which turned out to be something of a non-event. It may be that Bemer's warnings in the 1970s were not, in fact, unheeded.

Martin Campbell-Kelly