The First Computer Was a Job Title

Before a computer was something you restarted, it was somebody you employed. They required mathematical ability, patience and wages. If the result was wrong, you could ask them how they had reached it, although the answer might involve several pages and an understandably defensive expression.

The word described a person who calculated long before it described an electronic machine. The Centre for Computing History traces a recorded use to 1613. By the time machines inherited the name, people had spent generations doing the work.

This is more than an entertaining piece of vocabulary. It changes where computing history begins. Before the famous equipment came organisations that had already divided problems into operations, allocated them to workers and devised ways to check the results.

The machines entered an existing workplace.

Calculation had to be organised

Large-scale calculation is not simply a matter of finding someone good at arithmetic. A substantial problem must be broken into manageable parts. Inputs have to be recorded, intermediate results preserved and discrepancies investigated.

That work existed in astronomy, engineering and military research. Its tools included paper, tables and mechanical calculators. The human computer did not necessarily calculate unaided, any more than a modern programmer performs their job without equipment.

The Smithsonian’s account of ENIAC describes the human computers, largely women using mechanical calculators, who were performing the calculations the electronic machine was intended to accelerate. Demand had outgrown what the existing arrangement could deliver.

The problem was therefore partly one of production. How could useful answers be obtained quickly and reliably enough? The mathematical method mattered, but so did the system through which it became a result.

This is a familiar distinction to anyone who has watched an organisation discover that employing clever people does not automatically make their work fit together.

The stars did not classify themselves

At Harvard College Observatory, women employed as computers examined photographic plates and analysed stellar observations. The arrangement gave them access to important scientific work without necessarily granting equivalent authority over its direction.

Henrietta Swan Leavitt’s contribution makes the limitations of the job label especially clear. Her study of variable stars revealed a relationship between the time taken to cycle in brightness and their luminosity.

As Harvard’s account explains, the relationship, confirmed through further observations in 1912, became fundamental to measuring astronomical distances. A task involving repeated observation and comparison helped produce a way of establishing the scale of the universe.

It would be misleading to describe all computing work as discovery. Much was repetitive. But it would be equally misleading to assume that repetition excluded insight.

The person spending sustained time with the evidence might notice something the person assigning the task had not anticipated. A modest place in the organisation did not establish a modest contribution to knowledge.

The machine required another kind of understanding

When ENIAC arrived, the need for human expertise did not obligingly disappear. Somebody had to turn a mathematical problem into operations the machine could execute.

Six women became its original programmers. The ENIAC Programmers Project, which collected oral histories from members of the group, documents their work and the long struggle to secure recognition for it.

Programming the early machine was not a matter of typing a request and waiting for an answer. It required understanding the equipment, sequencing operations and configuring the system. The imposing hardware did not explain to itself what it was supposed to accomplish.

There is a temptation to treat that work as a supporting activity because it happened around the machine. But a calculator that cannot be made to perform the required calculation is an elaborate demonstration of unrealised potential.

The physical object was easier to photograph than the intellectual work of making it useful. That is not the same as being more important.

A transition somebody had to manage

Dorothy Vaughan’s career offers another account of what changed. She led the segregated West Area Computing Unit at the National Advisory Committee for Aeronautics, NASA’s predecessor, from 1949 to 1958.

NASA’s biography describes both her mathematical leadership and her subsequent expertise in FORTRAN programming. As electronic computing developed, the work changed, and Vaughan helped make that transition practical.

It is easy to turn this into a reassuring story about adaptability. Technology arrives, a capable person learns the new skills, and the future proceeds efficiently.

But that reading can make the institution’s responsibilities disappear. Vaughan and her colleagues worked within racial and gender restrictions. Their achievements do not demonstrate that those restrictions were harmless because determined people sometimes overcame them.

Nor does one successful transition establish that everyone affected by automation can move effortlessly into the next role. Training, opportunity and the authority to apply new skills have to exist.

A machine does not make those arrangements merely by arriving.

“Routine” can conceal a great deal

There is a revealing habit in how we describe work before and after automation. Beforehand, a task may require dependable judgement and considerable practice. Once a machine performs part of it, the earlier activity can be retrospectively reduced to something anybody could have done.

Sometimes the task really was repetitive and tightly specified. Automating it can be an uncomplicated improvement. Nobody should be required to spend a career repeating arithmetic so that future historians have something moving to write about.

But routine work can also contain decisions that become visible only when someone tries to automate it. Which observation is suspect? Does the result make sense? Has the method been applied to a case it was not designed for?

These are not arguments against machines. They are reasons to investigate the work honestly before declaring that all of it has been captured.

A job description is a summary written for an organisation’s purposes. It is not a complete inventory of what a person knows.

The archive has its own preferences

The history we inherit is shaped by what can be preserved. Machines leave cabinets, components and photographs. Organisations leave titles and formal reports. Understanding how work happened may require notebooks, correspondence and the testimony of people whose names were not prominent at the time.

That helps explain why recovering the human computers matters. Adding them to the story changes the account of invention from a sequence of machines into a history of people developing methods and institutions around them.

It also complicates the comforting distinction between the visionary and the person doing the detail. Sometimes the detail contains the discovery. Sometimes the person making the plan work has to solve problems absent from the plan.

Recognition need not become a competition in which the engineers lose whatever the programmers gain. A more accurate history has room for several kinds of contribution.

The title moved. The work continued.

The eventual shift in the meaning of “computer” was real. Electronic machines transformed the speed and scale of calculation, making work possible that human computing teams could not realistically undertake.

But the continuity deserves equal attention. People still had to formulate problems, establish methods and decide whether a result could be trusted. Some tasks disappeared, others changed, and new specialisms developed around the capabilities of the machine.

That is a more useful history than either technological triumph without workers or a story in which nothing important changed. The machines mattered enormously. So did the people who understood what to ask of them.

Today, calling someone a computer sounds as though you are complimenting their memory while expressing reservations about their social life. Once, it identified a contribution institutions depended upon.

The machine inherited the title. Remembering the people who held it first is a useful precaution whenever somebody announces that the work now does itself.