PART FOUR: COMMUNICATION AND CALCULATION 692 Table 15.2 : The expansion of electronics and its effect on industry. INFORMATION 693 Table 15.2 (cont.) Courtesy of Mr G. W. A. Dummer and Pergamon Press PART FOUR: COMMUNICATION AND CALCULATION 694 It is interesting that none of these criteria relate to the theory underlying an invention; inventions often precede their theoretical explanation. For example, for several centuries (from Franklin to Marconi) electricity was the basis of many inventions, but the force behind it—the electron—was only discovered by J.J.Thompson at the beginning of this century. On the other hand, Maxwell’s elegant theory of electromagnetism preceded Hertz’s demonstration of wireless communications. Every invention of the first rank—in the context of this chapter, clocks, calculators, computers, photography, the phonograph, motion pictures, the telephone, radio and television, and satellites—has spread rapidly from its origin, and become used by the majority of the population instead of just being the plaything of a few devotees. Also, people do not actually ‘own’ most of these inventions. The only ones which are really self-contained are clocks, watches and calculators. Cameras, radios, television, the telephone and satellites are useless without elaborate support services. For example, the telephone instrument, which is all that we see and that most people are conscious of, is only the terminal apparatus of what is perhaps the most complex and capital-intensive industry ever known. Similarly, whereas we own motor cars, to use them we must have a highly developed and maintained infrastructure of highways, petrol stations, and repair shops. TIMEKEEPING Mankind first developed a sense of time from observations of nature. For the short term, he observed the movement of heavenly bodies—the sun and moon held particular importance, heralding the seasons and the months. For the long term, birth and death events—of themselves and their livestock—marked the passage of time. One of the earliest inventions was the astrolabe, which astronomers used to track stars and planets. The first such instrument may have been made in the second century by the greatest astronomer of ancient times, the Greek Hipparchus, and was brought to perfection by the Arabs. Early artificial means of timekeeping to provide an estimate of the hour were all analogue in nature, whether passive, like the sundial, or dynamic, like the sandglass or water-clock, which measured time by rate of flow. The sundial probably began with a stick thrust into the ground; the position of its shadow corresponded to the hour of the day. Very elaborate sundials have been constructed which compensate for the sun’s relative position during the year, and ingenious pocket versions have been popular from time to time. However, all such instruments are worthless if the day is cloudy, and after the sun goes down. Therefore, particularly for stargazers, independent means of time estimation were important. The simple sandglass, in which sand is made to run through a INFORMATION 695 small opening, was adequate only for short durations. However, running water can power a mechanism indefinitely. The greatest water-clock ever made was a building-sized astronomical device, constructed by Su Sung in China in 1094, to simulate the movements of sun, moon and the principal stars. Chinese philosophers thought that because water flows with perfect evenness, it is the best basis for timekeeping. However, in this belief they were wrong; the secret of accurate timekeeping is to generate and count regular beats, which is a digital rather than an analogue process. This may be surprising, and not just to the ancients in China, because except for intra-atomic events time can be regarded as a continuous phenomenon. The mechanical clock Three elements necessary to the invention of the mechanical clock were brought together early in the fourteenth century: 1. A weight-driven mechanism to provide constant force; unlike water-clocks, such a mechanism is relatively independent of temperature. 2. Energy (gravity pulling the weight) transmitted to a train of wheels, or better gears (Hero of Alexandria, one of the earliest known inventors, developed the principle of the gear wheel in the second century). 3. Back-and-forth motion, such as provided by a controlling device or regulator whose oscillations (beats) ‘keep track of time’, with an escapement to count the beats by blocking and releasing the wheel train (early mechanical clocks did not separate these functions). In summary, the essence of clockwork is the conversion of stored energy into regular movements in equal intervals; in modern terms, analogue-to-digital conversion takes place. The earliest mechanical clock on record was said to have been built in China in 725 by I Hsing and Liang Ling-tsan. In Europe, the earliest weight- driven clocks were built in the fourteenth century: they did not indicate time visually, but struck a bell every hour. One was built for Wells Cathedral in Somerset in 1335. The first public clock that struck the hours was built in Milan in the same year, and late in that century clocks appeared for domestic use. Each beat of these clocks was several seconds long, so their accuracy was poor. The first astronomical clock, built by Giovanni de’Dondi between 1348 and 1362, was well in advance of its time, showing the motion of the sun, moon and planets and providing a perpetual calendar. Spring-driven clocks appeared in the fifteenth century, but were even less accurate because a spring could not provide a constant driving force. The fusee wheel (see p. 392), a grooved, conical spindle around which a cord was PART FOUR: COMMUNICATION AND CALCULATION 696 wrapped to compensate for this irregularity, was first used in 1430 for Philip the Good’s Burgundy clock. By the middle of the fifteenth century, reasonably accurate spring-driven clocks became available, which could be made much smaller. About 1500, Peter Henlein, a German locksmith, built what is said to be the first portable clock. In 1584, Joost Burgi invented the cross-beat escapement, which increased accuracy almost a hundred times. However, this was soon made obsolete by the crucial invention of the pendulum at the beginning of the seventeenth century. Galileo had observed that a free-swinging weight gave a reliable and consistent period, dependent only on its length (as long as the swing was small). A pendulum 990mm (39in) long gave a period of one second, and would lose only one second a day if this length was kept within an accuracy of 0.025mm; thus, a perfect pendulum would have an accuracy of 1 part in 100,000. Galileo did not actually build such a pendulum-driven clock; Christiaan Huygens in the Netherlands seems to have been the first, in 1656. Huygens’s clock kept time to an accuracy of 10–15 seconds per day, a vast improvement on the verge- and-balance system which varied more than 15 minutes a day; in 1670, the minute hand was introduced. In 1675, Huygens invented the balance wheel which allowed further miniaturization, and led to the development of the pocket watch. During the seventeenth and eighteenth centuries, great attention was devoted by maritime nations to improving the accuracy of timekeeping. Although ships at sea could determine their latitude by ‘shooting’ the sun or stars, to determine longitude they also had to know the time difference between a fixed meridian and their position at sea. Therefore, navigation in the east-west direction was very uncertain (perhaps this was why Columbus thought he had discovered India). A prize was offered by the British Admiralty for a clock which, carried aboard a rolling and pitching ship, could still keep time to an accuracy sufficient for the determination of longitude. The prize was won by John Harrison in 1762 with his No. 4 chronometer which made two beats every second. Modern clocks and watches About 1790, Jacquet-Droz and Leschot invented a wrist-watch in Geneva, but it was not until Adrien Philippe of Paris brought his 1842 invention of the stemwinding mechanism to Patek in Geneva in 1844 that a practical wrist- watch became possible. Today, because of its convenience and as an attractive adornment, the wrist-watch has all but supplanted the pocket watch. A clock based on a tuning fork was invented by Louis Breguet in France in 1866, but although this further miniaturized the regulation function, a watch using this principle was not made until 1954 by a Swiss, Hetzel. Alexander Bain invented the first battery-driven electric clock in 1840, although a battery was not incorporated into a clock until 1906. In 1918, the domestic electric INFORMATION 697 clock using a synchronous motor was introduced; this depended on the mains frequency for its accuracy, and stopped if the current was interrupted. In 1929, Warren Morrison in the United States used the natural period of vibration of a quartz crystal to give new accuracy to timekeeping, better than 0.0001 second per day. A ring of quartz was piezoelectrically driven to oscillate at 100,000 vibrations per second, and the resulting electrical signal frequency- divided a number of times to drive a human-readable display. The atomic clock was developed by William F.Libby in the United States, the first model being constructed at the National Bureau of Standards in 1948. The atomic clock was the first artificial timekeeper which was better than the stability of the earth’s rotation. The functioning of present-day atomic clocks depends upon the natural resonance of caesium atoms. Since 1972, the international time standard has been the atomic second, defined as the duration of precisely 9,192,631,770 periods of the radiation corresponding to the transition between two hyperfine levels of the ground state of the caesium-133 atom. This is an accuracy of 1 part in 2×10-11 per month, more than ten times that of the best quartz clock. A master clock which controls up to fifty slave clocks was marketed in 1963 by Patek Philippe; it is synchronized with an atomic clock by radio signals/Since 1960, the international standard for length has also been based on atomic resonance: one metre is defined as 1,650,763.73 wavelengths in vacuum of radiation corresponding to the orange-red line of the krypton-86 atom. In 1961, an electronic clock was patented by P.Vogel et Cie of Switzerland, followed by an electronic watch patent a year later. However, it was not until 1971 that the first electronic digital watch was marketed by Time Computer Corporation in the United States. This watch, called Pulsar, cost $2000, and used light-emitting diodes (LED) to show the time; because they use so much current, the wearer had to push a button to read the display. Most people are used to interpreting the two hands of an analogue timepiece; the digital watch provides a direct, numerical readout. Present digital-watch technology is quartz- based, and instead of gear-train and hands, uses an integrated circuit to calculate the time, showing numbers representing hours, minutes and seconds continuously on a low-current-drain, liquid-crystal display (LCD). Digital timepieces may also provide indications of the month, date and day of the week, alarms, corresponding time in other zones, and even a calculator function. Inexpensive digital watches were offered in kit form in the mid-1970s by the British inventor Sir Clive Sinclair, but few could be made to work. The first acceptable consumer products came from Timex Corporation and Texas Instruments in the US, and Seiko and other companies in Japan; they offered an unbeatable relation between price and performance, and put the mass- market, mechanical watch industry in jeopardy. However, even in the most expensive digital watches, the LCD display is inelegant; therefore, many watchmakers are now using digital electronics to control the hands of conventional-looking watches. PART FOUR: COMMUNICATION AND CALCULATION 698 COUNTING, CALCULATING AND COMPUTING The first mechanical calculator (indeed, the world’s first digital computer) was the abacus, whose exact origins are lost in prehistory; however, counting devices were in use in the Far East by 450 BC. An abacus is a small wooden frame into which wires are fastened; each wire represents a numerical place value (units, tens, hundreds, etc.), and is strung with a number of beads, each of which represents a digit with that place value. By manipulating the beads, addition and subtraction may be performed at a rate which in skilled hands is astonishing. In 1946, a Japanese clerk using a soroban (abacus) easily won a speed contest against an American soldier using an electric desk calculator. Many early civilizations developed analogue or digital systems for measurement or calculation. The megaliths of Stonehenge and other stone circles were erected about 2000 BC to predict important astronomical events. Polynesians in outrigger canoes fitted with sails navigated thousands of kilometres of the Pacific Ocean using direct observation of natural phenomena, sometimes supplemented with ‘maps’ made of knotted cords attached to a wooden frame. In the sixteenth century, the Incas of Peru developed the quipu, an intricate pattern of knotted cords, for calculation. Logarithms and the slide-rule Michael Stifel, a mathematician of Jena, coined the term exponent in 1553; and as early as 1544, he had discovered the properties of positive and negative exponents by noting the connection between two sequences, one of which is an arithmetical progression and the other geometrical. About 1600, the Scottish mathematician John Napier invented logarithms; and the Swiss watchmaker Joost Burgi independently invented antilogarithms. The beauty of this system of calculation was that problems of multiplication were reduced to those of addition, and division reduced to subtraction. There was a ready market for the tables designed to calculate the products of sines which Napier published after twenty years of labour, because every advance in science and technology demanded greater precision. In 1624, Napier’s friend and collaborator Henry Briggs published the first tables of logarithms to the base ten (decimal logarithms), which became the basis of all surveying, astronomy and navigation until recently. Napier also developed a calculating system using his ‘natural’ logarithms, which came to called Napier’s bones. In 1622, the English mathematician William Oughtred invented the slide-rule, an analogue calculator based on logarithmic scales. The slide-rule was much quicker than using tables, but the accuracy of practical-sized instruments was limited to three or four figures. The slide-rule in its case swinging from the belt was a trademark of students and INFORMATION 699 practitioners of engineering until the mid-1970s, when the inexpensive, pocketsized, scientific calculator made it obsolete. Mechanical calculating machines The invention of the first calculating (more accurately, adding) machine which did not require the memory skills of the abacus operator is attributed to the French mathematican and philosopher Blaise Pascal in 1642. His priority has been disputed: a letter written to Johannes Kepler by Wilhelm Schickhardt of Tübingen in 1624 describes a similar device, which he called a calculating clock. Pascal’s calculating machine used gear wheels, each of which had the numbers 0 to 9 engraved on its face (see Figure 15.1). There was a wheel for units, one for tens, one for hundreds, and so forth; gear wheels were used for carrying. In 1671, Gottfried Wilhelm Leibniz conceived a calculating machine which could multiply numbers and such a device was built in 1694, although it proved unreliable. The eighteenth century saw no radical innovations in calculating devices; however, a great deal was accomplished in the related field of horology, where greater timekeeping accuracy stimulated more precise machining of gears and small mechanisms. This refinement in mechanical technology would stimulate Charles Babbage to conceive the first true digital computer early in the nineteenth century. In 1805, Joseph-Marie Jacquard in France invented the perforated cards which, linked together, were used to control weaving patterns in the Jacquard loom (see p. 823). Jacquard’s system was the direct ancestor of punched cards, the key element of the tabulating machines which were the principal product of IBM during the first half of the twentieth century, and the main form of external storage for digital computers until the 1960s. Most of the concepts which underly the modern digital computer were first enunciated by Charles Babbage, an English mathematician. Based on the division of human labour which Gaspard Riche, Baron de Prony, had used to calculate logarithms and trigonometric functions by the method of differences, he conceived of a machine to accomplish the task without human intervention. In 1822 he built a prototype of what he called the Difference Engine (see Figure 5, p. 32), which calculated the first thirty terms of a polynomial. Based on this promising demonstration, the government funded a three-year contract to build a working machine which could both calculate and print tables. The Difference Engine was never completed, although the government spent £17,000 and Babbage at least as much out of his own pocket. Apparently, this failure was not because the machinists of his day could not construct gears and levers to the required precision, but because of inadequate project management. PART FOUR: COMMUNICATION AND CALCULATION 700 In 1833, Babbage had a new and much more ambitious conception, the Analytical Engine, whose attempted realization occupied much of the rest of his life. Although a mechanical solution, the only one available in his day, could never have accomplished the task, the Analytical Engine was the ancestor of all current digital computers. The machine was to have four parts: a store, made of columns of wheels in which up to a thousand 50-digit numbers could be placed; the mill, where arithmetic operations would be performed by gear rotations; a transfer system composed of gears and levers; Figure 15.1: Model of Blaise Pascal’s 1642 calculating machine. INFORMATION 701 and a mechanism for number input and output. For the latter, he planned to adapt Jacquard’s punched cards. Babbage also anticipated many other of the most important features of today’s computers: automatic operation, external data memories, conditional operations and programing. Between 1836 and 1840, Babbage devised some fifty user programs for the Analytical Engine, using two strings of Jacquard cards, one of which was for the arithmetic operations and the other to specify variables; the strings of cards could be looped, and could move forwards or backwards independently. In 1840, Babbage gave seminars in Italy; as a result, Luigi Menabrea wrote a paper on the Analytical Engine which, translated into English by Ada Augusta, Countess of Lovelace, is the best description we have of Babbage’s ideas. (Her name has been given to Ada, the most powerful, general-purpose computer language so far developed.) In 1846, Babbage stopped working on the Analytical Engine, and began to design a new Difference Engine. Georg and Edvard Scheutz built one on the basis of a published description (which included no drawings), and showed it to Babbage in 1855; a copy of the Scheutz engine was used to calculate life expectancy tables, but proved unreliable. Babbage then tried to simplify the design of the Analytical Engine; models of the mill and printing mechanism were under construction at the time of his death in 1871. In 1847 the English mathematician George Boole published The Mathematical Analysis of Logic, which showed that logic could be expressed using binary symbols. As in the binary number system, Boolean algebra has only two symbols, 0 and 1; however, instead of using them as binary digits for arithmetic computation, they are interpreted as the outcomes of logical operations, 0 denoting ‘false’ and 1 ‘true’. (In 1938, Claude Shannon of the Bell Telephone Laboratories realized that the Boolean ‘true’ and ‘false’ could be represented by the ‘on’ and ‘off states of bistable electronic components, and this became the basis of computer logic design.) Herman Hollerith devised a suite of punch-card machines in 1880 to assist in processing the United States census; but the cards stored only data rather than programs. In 1896 he formed the Tabulating Machine Company, which became International Business Machines in 1957 and is now called just IBM. Digital computers Alan M.Turing, working at the University of Manchester, published in 1936 a paper, ‘Can a Machine Think?’, which has been a challenge to computer philosophers ever since. In it he proposed a theoretical ‘universal machine’, whose purpose was nothing less than to determine whether any mathematical statement could be proved true or not. Later, Turing became the first mathematician to be employed in the British cryptanalysis service where he developed relay-based machines to break the Enigma-enciphered transmissions of the German navy. The resulting device, Colossus, became operational in . bodies the sun and moon held particular importance, heralding the seasons and the months. For the long term, birth and death events of themselves and their livestock—marked the passage of time models of the mill and printing mechanism were under construction at the time of his death in 1871. In 1847 the English mathematician George Boole published The Mathematical Analysis of Logic,. so their accuracy was poor. The first astronomical clock, built by Giovanni de’Dondi between 1348 and 1362, was well in advance of its time, showing the motion of the sun, moon and planets and