From laboratory to office
At the start of the 1950s, computers were mostly experimental machines run by universities, government laboratories and the military. By the end, hundreds of them were installed in businesses, insurance companies, banks and government offices. The technology changed in important ways, but the larger shift was in who bought computers and what they used them for.
The first sales
Several groups tried to turn the wartime designs into products. In Britain, a machine built with Manchester University was developed commercially in the early 1950s, and a catering company known for its tea shops built its own computer for office work later in the decade. In the United States, the UNIVAC I was delivered to the Census Bureau in 1951 and became one of the best-known early commercial computers. IBM entered with the IBM 701 for scientific and defense customers, announced in 1952, and followed with machines aimed at business. The IBM 650, introduced in the mid-1950s, was used at many universities and companies.
Prices were very high, and many organizations rented rather than bought. A computer needed a dedicated room, trained operators, and often air conditioning. For most firms, the case rested on replacing large amounts of clerical work such as payroll, billing and inventory records.
Cards, tape and drums
A 1950s computer center looked like a mix of old and new equipment. Data arrived on punched cards, a format that businesses had used with tabulating machines for decades. Results were printed on paper or punched onto new cards. Magnetic tape was introduced during the decade as a faster way to store and move large files, and it became a standard tool of big installations.
Memory improved too. Magnetic core memory, built from tiny rings of magnetic material threaded on wires, became widely used in the middle of the decade and was more dependable than delay lines or special tubes. In 1956 IBM introduced a disk storage system called RAMAC, an early example of a hard disk drive that allowed direct access to data rather than waiting for a tape to wind to the right spot.
Programming becomes its own problem
Early programmers wrote in machine code, long strings of numbers, or in assembly language, which gives short names to machine instructions. This was slow and error-prone, and programs were tied to one model of machine. As computers became more common, the cost of programmers’ time started to exceed the appeal of hand-tuned code.
The response was higher-level languages and compilers, programs that translate readable instructions into machine code. Grace Hopper and her colleagues worked on early compiler systems in the early 1950s and promoted the idea that programs could be written in something closer to English. At IBM, a team led by John Backus released FORTRAN in 1957, aimed at engineers and scientists. It was designed to produce machine code efficient enough that people would actually use it, and its success helped make scientific programming a normal activity rather than a specialty.
John McCarthy and colleagues began work on Lisp in the late 1950s, which became important in artificial intelligence research. In 1959 a committee including representatives from industry and government began defining COBOL, a language meant for business data processing, which would remain in use for a very long time. ALGOL, an international effort, also appeared at the end of the decade and influenced many later languages.
Transistors and the first integrated circuits
Vacuum tubes were fast but hot, large and unreliable. The transistor, announced in 1947, began to replace them in computers in the second half of the 1950s. Transistorized machines were smaller, used less power and failed less often. By the end of the decade, manufacturers were selling “second generation” computers built this way.
In 1958 and 1959, researchers at Texas Instruments and Fairchild independently developed ways to put multiple components on one piece of semiconductor material. The result, integrated circuits, would reshape the industry in the next decade, but they were barely used in computers by 1959.
Government, defense and early research
Much of the money for computing still came from government. The SAGE air defense system, developed in the 1950s, linked radar stations to large computers and pushed progress in real-time computing, displays and communications. The launch of Sputnik in 1957 increased public interest in science and led the United States to create new research agencies the following year, including ARPA, which would later fund networking research.
The decade also saw the first serious discussions of machines that might think. A 1956 summer workshop at Dartmouth is usually credited with giving the field of artificial intelligence its name, although researchers had been exploring related ideas for some time. Early programs that played checkers or proved simple theorems drew attention, though their abilities were narrow.
Limits of the decade
It is easy to overstate how widespread computers were. Even in wealthy countries, most companies still did their accounting with people and mechanical office equipment. Computers were unreliable by modern standards, jobs were run in batches by operators, and users often waited hours or days to see the results of a program.
Still, a pattern had been set. Computers were something an organization could buy, and software was starting to become a field in its own right. The following decade pushed both ideas much further.