Grasping Reality on TypePad, by Brad DeLong

...began work as computers at the Royal Observatory in Greenwich, under the directorship of Astronomer Royal George Biddell Airy:

We were at our posts at 8 a.m. to the moment. I had not been many minutes seated on a high chair before a roomy desk placed on a table in the centre of the Octagon Room, when a huge book was placed before me, very different indeed to what I had anticipated. This large folio book of printed forms, was specially arranged for the calculation of the tabular right ascension and north polar distance of the planet Mercury from Lindenau’s Tables.... After very little instruction from Mr. Thomas, the principal computer in charge, I began to make my first entries with a slow and tremulous hand, doubting whether what I was doing was correct or not. But aft er a little quiet study of the examples given in the Tables, all this nervousness soon vanished, and before 8 pm came, when my day’s work was over, some of the older computers complimented me on the successful progress I had made...

Two boys sent out to support their widowed mother, the high chair and the huge ledger, the twelve hours of eye-straining, handcramping calculation (alleviated only by an hour’s dinner break), the standardized printed forms that divided computation into steps like the manufacture of pins—it could be a vignette from Dickens, and both Airy and his predecessor in the office, John Pond, have been cast by contemporaries and historians alike in the roles of Bounderby or Scrooge.

But the reality of massive calculation of the sort that went on in astronomical observatories since at least the medieval period in parts of Asia and in Europe since the sixteenth century (and since the nineteenth century in insurance offices and government statistical offices) was considerably more varied — as varied as the nature of work itself in different historical and cultural contexts. The only constant was that calculations on the large scale needed to reduce astronomical observations, compute life expectancies, and tally statistics on everything from crime to trade was indeed work: the first Astronomer Royal John Flamsteed, appointed in 1670, called it “labour harder than thrashing.” Before and even after the invention and diffusion of reliable calculating machines, the challenge to astronomers and other heavy-duty number crunchers was how to organize the work of deploying many algorithms, over and over again. These combined experiments in labor organization and algorithmic manipulation ultimately transformed both human labor and algorithms.

Let us return to young Edwin Dunkin perched on his high chair in the Octagon Room of the Royal Observatory at Greenwich. Edwin’s father William Dunkin had also been a “computer”—a word that until the mid-twentieth century referred primarily to human beings, not machines—and had worked for Airy’s predecessors, Astronomers Royal Nevil Maskelyne and John Pond, to calculate tables for the Nautical Almanac, a navigational tool for the globalized British navy that had been produced under the direction of the Astronomer Royal since 1767.

Unable to supply the labor necessary to compute the Almanac’s numerous tables from the Greenwich Observatory’s own resources, Maskelyne organized a network of paid computers throughout Britain to perform the thousands of calculations according to a set of “precepts” or algorithms, to be entered on pre-printed forms that divided up calculations (and indicated which values had to be looked up when from which one of fourteen different books of tables) into a step-by-step but by no means mechanical process. What is noteworthy about Maskelyne’s operation (which involved a computer, anti-computer, and comparer to check each month’s set of calculations) was its integration into an established system of piecework labor done in the home and often involving other family members. Each computer completed a whole month’s worth of lunar position or tide prediction calculations according to algorithms bundled like the patterns sent to cottage weavers to produce finished textile wares.

Just as the mid-eighteenth-century manufacturing system, in which many workers were gathered together under one roof and subjected to close managerial supervision, began to replace the family textile workshop long before the introduction of steam-driven looms, so the development of Big Calculation traced a parallel arc a good half-century before algorithms were calculated by machines. The careers of William and Edwin Dunkin, father and son computers in the service of the British Astronomers Royal, span the transition between piecework and manufacturing—but not yet mechanized—systems of labor organization...

#shouldread
#economichistory
#riseofgtehrobots