The 18th century was an age of quaint things as well as revolutions, novelties and technical toys. It brought forth Robert Burns, the United States of America and the musical box.
Have you ever examined a musical box closely? There is a cylinder that looks at first glance like a mechanical hedgehog; merging from it are many little steel pegs or pins like the hedgehog's spines. There is also a steel comb whose teeth pluck the pins and so make them give out tinkling sounds. Open the box, and the rattling old clockwork starts, the cylinder revolves, the pins are twanged by the teeth of the comb, and for sixty charming seconds you can hear "Annie Laurie." The notes have been tickling in music boxes with varying precision for two hundred years - a reliable program.
A decisive word has just been said: the word "program." Unless you are professionally associated with the electronic computer and its curious ways, you could hardly be expected to know that this word, "program," when used in electronic circles, has a momentous, fateful sound about it. In this context it does not evoke concerts, but the sequence of instructions in a particular working process. The program of an electronic computer does not consist of a planned series of silvery notes, but of a series - if all goes well, an equally planned series - of strictly practical counting operations.
Efforts are being made to equip electronic computers with fantasy, initiative, creative thought and all the other admirable qualities which have made it possible for men to fly into space, and recently even to manufacture non-dripping faucets. But, for the time being, you cannot ask a computer to work out all by itself a problem you wish to have solved. You cannot simply order it to reckon out the compound interest payable in 35 days on a savings bank account, taking due note of the local bank charges. Any novice bank clerk can do that, but an electronic computer costing close to a million cannot - or at least not until you have exactly explained the calculation to it beforehand. You most command it firmly, in this fashion: "Take the interest. Divide it by the number of days in a full year. Multiply the result by 35 days." And so on. If you have taught the computer how all this is to be done and set it going, then, but only then, will it give you the answer to this and similar problems more quickly than the new bank clerk can pick up a pencil.
Electronic computers, in fact, are idiots. Accomplished idiots, however, with a special talent: they can count swiftly and accurately. All they ask is to have every calculation process pre-digested for them, bit by bit. They will obey a sequence of commands from the beginning to the end of a lengthy problem's exposition: this is what the expert calls a "program." The mathematicians who concern themselves with working out the programs are called "programmers."
Let us assume that the programmers have written down a program for a payroll. The machine most work out the pay slips for all 2,000 employees in a business. In each case, it must take as a basis the number of hours each of them has worked, calculate the wage payable according to the current hourly rates, make tax and other deductions, and print the result on a ticket.
To solve the problem the machine must be industrious indeed. It must lack up in its records the hourly pay rate applicable to each worker, multiply the rate by the hours worked, store the figure thus obtained, extract the tax scale currently chargeable for the case in question, and so on. One number most be shifted to point A, another to point B, a third brought from the auxiliary storage and subtracted from the first. What is left over is stowed away in an intermediate storage, moved over to the arithmetic unit a dozen operations later and added them to another number altogether. Activity in a computer is like that at a post office at Christmas time.
But how are the signals operated? How are the sums switched from one set of lines to another? By hand as with a telephone switchboard? Efforts were made to do it that way in the Stone Age of the electronic computer. All the parts of the machine - the storages, the various departments of the arithmetic units, the input and output devices - were provided with plugs and sockets. The operator took a firm hold on the cords and plugged away: the input storage was connected up to certain addresses in the intermediate storage, a wire leading from the latter was plugged into the multiplier matrix which was connected to a further storage address, from which wires led to the adder matrix. A programmer thirty years ago was as busy as a telephonist at her switchboard exchange in the early 1900s.
The switchboard was soon replaced by the automatic selector. But mathematicians and technologists had complaints in plenty to address to the computers before that happened. Plugging-in was too much trouble. It was not only that plug-and-socket operations were far more involved than those which had to be carried out by a telephone girl. That was a disadvantage that had to be accepted. What annoyed them was the waste of time. If the girl at the exchange needed twenty seconds to put through a call lasting three to ten minutes, no one could justly complain that she was dawdling. But if a mathematician had to spend hours plugging in a simple program and the computer solved the problem (quite literally) with the speed of a flash of lightning zipping through the plugged connections, the disproportion in the amount of time consumed was simply preposterous.
The teleprinter network was already partly automated when the first electronic computers were still in the plug-and-socket era. It was only a matter of time before someone thought up automatic program controls for the computers too. After all, even the musical box - through its cylinder - is provided with one.