The Modern History of Computing

First published Mon Dec 18, 2000; substantive revision Fri Jun 9, 2006

Historically, computers were human clerks who calculated in accordancewith effective methods. These human computers did the sorts ofcalculation nowadays carried out by electronic computers, and manythousands of them were employed in commerce, government, and researchestablishments. The termcomputing machine, used increasinglyfrom the 1920s, refers to any machine that does the work of a humancomputer, i.e., any machine that calculates in accordance witheffective methods. During the late 1940s and early 1950s, with theadvent of electronic computing machines, the phrase ‘computingmachine’ gradually gave way simply to ‘computer’,initially usually with the prefix ‘electronic’ or‘digital’. This entry surveys the history of thesemachines.

Babbage

Charles Babbage was Lucasian Professor of Mathematics at CambridgeUniversity from 1828 to 1839 (a post formerly held by Isaac Newton).Babbage's proposed Difference Engine was a special-purpose digitalcomputing machine for the automatic production of mathematical tables(such as logarithm tables, tide tables, and astronomical tables). TheDifference Engine consisted entirely of mechanical components —brass gear wheels, rods, ratchets, pinions, etc. Numbers wererepresented in the decimal system by the positions of 10-toothed metalwheels mounted in columns. Babbage exhibited a small working model in1822. He never completed the full-scale machine that he had designedbut did complete several fragments. The largest — one ninth ofthe complete calculator — is on display in the London ScienceMuseum. Babbage used it to perform serious computational work,calculating various mathematical tables. In 1990, Babbage's DifferenceEngine No. 2 was finally built from Babbage's designs and is also ondisplay at the London Science Museum.

The Swedes Georg and Edvard Scheutz (father and son) constructed amodified version of Babbage's Difference Engine. Three were made, aprototype and two commercial models, one of these being sold to anobservatory in Albany, New York, and the other to theRegistrar-General's office in London, where it calculated and printedactuarial tables.

Babbage's proposed Analytical Engine, considerably more ambitiousthan the Difference Engine, was to have been a general-purposemechanical digital computer. The Analytical Engine was to have had amemory store and a central processing unit (or ‘mill’) andwould have been able to select from among alternative actionsconsequent upon the outcome of its previous actions (a facilitynowadays known as conditional branching). The behaviour of theAnalytical Engine would have been controlled by a program ofinstructions contained on punched cards connected together with ribbons(an idea that Babbage had adopted from the Jacquard weaving loom).Babbage emphasised the generality of the Analytical Engine, saying‘the conditions which enable a finite machine to makecalculations of unlimited extent are fulfilled in the AnalyticalEngine’ (Babbage [1994], p. 97).

Babbage worked closely with Ada Lovelace, daughter of the poetByron, after whom the modern programming language ADA is named.Lovelace foresaw the possibility of using the Analytical Engine fornon-numeric computation, suggesting that the Engine might even becapable of composing elaborate pieces of music.

A large model of the Analytical Engine was under construction at thetime of Babbage's death in 1871 but a full-scale version was neverbuilt. Babbage's idea of a general-purpose calculating engine was neverforgotten, especially at Cambridge, and was on occasion a lively topicof mealtime discussion at the war-time headquarters of the GovernmentCode and Cypher School, Bletchley Park, Buckinghamshire, birthplace ofthe electronic digital computer.

Analog computers

The earliest computing machines in wide use were not digital butanalog. In analog representation, properties of the representationalmedium ape (or reflect or model) properties of the representedstate-of-affairs. (In obvious contrast, the strings of binary digitsemployed in digital representation donot represent by meansof possessing some physical property — such as length —whose magnitude varies in proportion to the magnitude of the propertythat is being represented.) Analog representations form a diverseclass. Some examples: the longer a line on a road map, the longer theroad that the line represents; the greater the number of clear plasticsquares in an architect's model, the greater the number of windows inthe building represented; the higher the pitch of an acoustic depthmeter, the shallower the water. In analog computers, numericalquantities are represented by, for example, the angle of rotation of ashaft or a difference in electrical potential. Thus the output voltageof the machine at a time might represent the momentary speed of theobject being modelled.

As the case of the architect's model makes plain, analogrepresentation may bediscrete in nature (there is no suchthing as a fractional number of windows). Among computer scientists,the term ‘analog’ is sometimes used narrowly, to indicaterepresentation of onecontinuously-valued quantity by another(e.g., speed by voltage). As Brian Cantwell Smith has remarked:

‘Analog’ should … be a predicate on arepresentation whose structure corresponds to that of which itrepresents … That continuous representations should historicallyhave come to be called analog presumably betrays the recognition that,at the levels at which it matters to us, the world is morefoundationally continuous than it is discrete. (Smith [1991], p.271)

James Thomson, brother of Lord Kelvin, invented the mechanicalwheel-and-disc integrator that became the foundation of analogcomputation (Thomson [1876]). The two brothers constructed a device forcomputing the integral of the product of two given functions, andKelvin described (although did not construct) general-purpose analogmachines for integrating linear differential equations of any order andfor solving simultaneous linear equations. Kelvin's most successfulanalog computer was his tide predicting machine, which remained in useat the port of Liverpool until the 1960s. Mechanical analog devicesbased on the wheel-and-disc integrator were in use during World War Ifor gunnery calculations. Following the war, the design of theintegrator was considerably improved by Hannibal Ford (Ford [1919]).

Stanley Fifer reports that the first semi-automatic mechanicalanalog computer was built in England by the Manchester firm ofMetropolitan Vickers prior to 1930 (Fifer [1961], p. 29); however, Ihave so far been unable to verify this claim. In 1931, Vannevar Bush,working at MIT, built the differential analyser, the first large-scaleautomatic general-purpose mechanical analog computer. Bush's design wasbased on the wheel and disc integrator. Soon copies of his machine werein use around the world (including, at Cambridge and ManchesterUniversities in England, differential analysers built out of kit-setMeccano, the once popular engineering toy).

It required a skilled mechanic equipped with a lead hammer to set upBush's mechanical differential analyser for each new job. Subsequently,Bush and his colleagues replaced the wheel-and-disc integrators andother mechanical components by electromechanical, and finally byelectronic, devices.

A differential analyser may be conceptualised as a collection of‘black boxes’ connected together in such a way as to allowconsiderable feedback. Each box performs a fundamental process, forexample addition, multiplication of a variable by a constant, andintegration. In setting up the machine for a given task, boxes areconnected together so that the desired set of fundamental processes isexecuted. In the case of electrical machines, this was done typicallyby plugging wires into sockets on a patch panel (computing machineswhose function is determined in this way are referred to as‘program-controlled’).

Since all the boxes work in parallel, an electronic differentialanalyser solves sets of equations very quickly. Against this has to beset the cost of massaging the problem to be solved into the formdemanded by the analog machine, and of setting up the hardware toperform the desired computation. A major drawback of analog computationis the higher cost, relative to digital machines, of an increase inprecision. During the 1960s and 1970s, there was considerable interestin ‘hybrid’ machines, where an analog section is controlledby and programmed via a digital section. However, such machines are nowa rarity.

The Universal Turing Machine

In 1936, at Cambridge University, Turing invented the principle of themodern computer. He described an abstract digital computing machineconsisting of a limitless memory and a scanner that moves back andforth through the memory, symbol by symbol, reading what it finds andwriting further symbols (Turing [1936]). The actions of the scanner aredictated by a program of instructions that is stored in the memory inthe form of symbols. This is Turing's stored-program concept, andimplicit in it is the possibility of the machine operating on andmodifying its own program. (In London in 1947, in the course of whatwas, so far as is known, the earliest public lecture to mentioncomputer intelligence, Turing said, ‘What we want is a machinethat can learn from experience’, adding that the‘possibility of letting the machine alter its own instructionsprovides the mechanism for this’ (Turing [1947] p. 393). Turing'scomputing machine of 1936 is now known simply as the universal Turingmachine. Cambridge mathematician Max Newman remarked that right fromthe start Turing was interested in the possibility of actually buildinga computing machine of the sort that he had described (Newman ininterview with Christopher Evans in Evans [197?].

From the start of the Second World War Turing was a leadingcryptanalyst at the Government Code and Cypher School, Bletchley Park.Here he became familiar with Thomas Flowers' work involving large-scalehigh-speed electronic switching (described below). However, Turingcould not turn to the project of building an electronic stored-programcomputing machine until the cessation of hostilities in Europe in1945.

During the wartime years Turing did give considerable thought to thequestion of machine intelligence. Colleagues at Bletchley Park recallnumerous off-duty discussions with him on the topic, and at one pointTuring circulated a typewritten report (now lost) setting out some ofhis ideas. One of these colleagues, Donald Michie (who later foundedthe Department of Machine Intelligence and Perception at the Universityof Edinburgh), remembers Turing talking often about the possibility ofcomputing machines (1) learning from experience and (2) solvingproblems by means of searching through the space of possible solutions,guided by rule-of-thumb principles (Michie in interview with Copeland,1995). The modern term for the latter idea is ‘heuristicsearch’, a heuristic being any rule-of-thumb principle that cutsdown the amount of searching required in order to find a solution to aproblem. At Bletchley Park Turing illustrated his ideas on machineintelligence by reference to chess. Michie recalls Turing experimentingwith heuristics that later became common in chess programming (inparticular minimax and best-first).

Further information about Turing and the computer, including hiswartime work on codebreaking and his thinking about artificialintelligence and artificial life, can be found in Copeland 2004.

Electromechanical versus Electronic Computation

With some exceptions — including Babbage's purely mechanicalengines, and the finger-powered National Accounting Machine - earlydigital computing machines were electromechanical. That is to say,their basic components were small, electrically-driven, mechanicalswitches called ‘relays’. These operate relatively slowly,whereas the basic components of an electronic computer —originally vacuum tubes (valves) — have no moving parts saveelectrons and so operate extremely fast. Electromechanical digitalcomputing machines were built before and during the second world war by(among others) Howard Aiken at Harvard University, George Stibitz atBell Telephone Laboratories, Turing at Princeton University andBletchley Park, and Konrad Zuse in Berlin. To Zuse belongs the honourof having built the first working general-purpose program-controlleddigital computer. This machine, later called the Z3, was functioning in1941. (A program-controlled computer, as opposed to a stored-programcomputer, is set up for a new task by re-routing wires, by means ofplugs etc.)

Relays were too slow and unreliable a medium for large-scalegeneral-purpose digital computation (although Aiken made a valianteffort). It was the development of high-speed digital techniques usingvacuum tubes that made the modern computer possible.

The earliest extensive use of vacuum tubes for digitaldata-processing appears to have been by the engineer Thomas Flowers,working in London at the British Post Office Research Station at DollisHill. Electronic equipment designed by Flowers in 1934, for controllingthe connections between telephone exchanges, went into operation in1939, and involved between three and four thousand vacuum tubes runningcontinuously. In 1938–1939 Flowers worked on an experimentalelectronic digital data-processing system, involving a high-speed datastore. Flowers' aim, achieved after the war, was that electronicequipment should replace existing, less reliable, systems built fromrelays and used in telephone exchanges. Flowers did not investigate theidea of using electronic equipment for numerical calculation, but hasremarked that at the outbreak of war with Germany in 1939 he waspossibly the only person in Britain who realized that vacuum tubescould be used on a large scale for high-speed digital computation. (SeeCopeland 2006 for m more information on Flowers' work.)

Atanasoff

The earliest comparable use of vacuum tubes in the U.S. seems to havebeen by John Atanasoff at what was then Iowa State College (nowUniversity). During the period 1937–1942 Atanasoff developedtechniques for using vacuum tubes to perform numerical calculationsdigitally. In 1939, with the assistance of his student Clifford Berry,Atanasoff began building what is sometimes called the Atanasoff-BerryComputer, or ABC, a small-scale special-purpose electronic digitalmachine for the solution of systems of linear algebraic equations. Themachine contained approximately 300 vacuum tubes. Although theelectronic part of the machine functioned successfully, the computer asa whole never worked reliably, errors being introduced by theunsatisfactory binary card-reader. Work was discontinued in 1942 whenAtanasoff left Iowa State.

Colossus

The first fully functioning electronic digital computer was Colossus,used by the Bletchley Park cryptanalysts from February 1944.

From very early in the war the Government Code and Cypher School(GC&CS) was successfully deciphering German radio communicationsencoded by means of the Enigma system, and by early 1942 about 39,000intercepted messages were being decoded each month, thanks toelectromechanical machines known as ‘bombes’. These weredesigned by Turing and Gordon Welchman (building on earlier work byPolish cryptanalysts).

During the second half of 1941, messages encoded by means of atotally different method began to be intercepted. This new ciphermachine, code-named ‘Tunny’ by Bletchley Park, was brokenin April 1942 and current traffic was read for the first time in Julyof that year. Based on binary teleprinter code, Tunny was used inpreference to Morse-based Enigma for the encryption of high-levelsignals, for example messages from Hitler and members of the GermanHigh Command.

The need to decipher this vital intelligence as rapidly as possibleled Max Newman to propose in November 1942 (shortly after hisrecruitment to GC&CS from Cambridge University) that key parts ofthe decryption process be automated, by means of high-speed electroniccounting devices. The first machine designed and built to Newman'sspecification, known as the Heath Robinson, was relay-based withelectronic circuits for counting. (The electronic counters weredesigned by C.E. Wynn-Williams, who had been using thyratron tubes incounting circuits at the Cavendish Laboratory, Cambridge, since 1932[Wynn-Williams 1932].) Installed in June 1943, Heath Robinson wasunreliable and slow, and its high-speed paper tapes were continuallybreaking, but it proved the worth of Newman's idea. Flowers recommendedthat an all-electronic machine be built instead, but he received noofficial encouragement from GC&CS. Working independently at thePost Office Research Station at Dollis Hill, Flowers quietly got onwith constructing the world's first large-scale programmable electronicdigital computer. Colossus I was delivered to Bletchley Park in January1943.

By the end of the war there were ten Colossi working round the clockat Bletchley Park. From a cryptanalytic viewpoint, a major differencebetween the prototype Colossus I and the later machines was theaddition of the so-called Special Attachment, following a key discoveryby cryptanalysts Donald Michie and Jack Good. This broadened thefunction of Colossus from ‘wheel setting’ — i.e.,determining the settings of the encoding wheels of the Tunny machinefor a particular message, given the ‘patterns’ of thewheels — to ‘wheel breaking’, i.e., determining thewheel patterns themselves. The wheel patterns were eventually changeddaily by the Germans on each of the numerous links between the GermanArmy High Command and Army Group commanders in the field. By 1945 therewere as many 30 links in total. About ten of these were broken and readregularly.

Colossus I contained approximately 1600 vacuum tubes and each of thesubsequent machines approximately 2400 vacuum tubes. Like the smallerABC, Colossus lacked two important features of modern computers. First,it had no internally stored programs. To set it up for a new task, theoperator had to alter the machine's physical wiring, using plugs andswitches. Second, Colossus was not a general-purpose machine, beingdesigned for a specific cryptanalytic task involving counting andBoolean operations.

F.H. Hinsley, official historian of GC&CS, has estimated thatthe war in Europe was shortened by at least two years as a result ofthe signals intelligence operation carried out at Bletchley Park, inwhich Colossus played a major role. Most of the Colossi were destroyedonce hostilities ceased. Some of the electronic panels ended up atNewman's Computing Machine Laboratory in Manchester (see below), alltrace of their original use having been removed. Two Colossi wereretained by GC&CS (renamed GCHQ following the end of the war). Thelast Colossus is believed to have stopped running in 1960.

Those who knew of Colossus were prohibited by the Official SecretsAct from sharing their knowledge. Until the 1970s, few had any ideathat electronic computation had been used successfully during thesecond world war. In 1970 and 1975, respectively, Good and Michiepublished notes giving the barest outlines of Colossus. By 1983,Flowers had received clearance from the British Government to publish apartial account of the hardware of Colossus I. Details of the latermachines and of the Special Attachment, the uses to which the Colossiwere put, and the cryptanalytic algorithms that they ran, have onlyrecently been declassified. (For the full account of Colossus and theattack on Tunny see Copeland 2006.)

To those acquainted with the universal Turing machine of 1936, andthe associated stored-program concept, Flowers' racks of digitalelectronic equipment were proof of the feasibility of using largenumbers of vacuum tubes to implement a high-speed general-purposestored-program computer. The war over, Newman lost no time inestablishing the Royal Society Computing Machine Laboratory atManchester University for precisely that purpose. A few months afterhis arrival at Manchester, Newman wrote as follows to the Princetonmathematician John von Neumann (February 1946):

I am … hoping to embark on a computing machinesection here, having got very interested in electronic devices of thiskind during the last two or three years. By about eighteen months ago Ihad decided to try my hand at starting up a machine unit when I gotout. … I am of course in close touch with Turing.

Turing's Automatic Computing Engine

Turing and Newman were thinking along similar lines. In 1945 Turingjoined the National Physical Laboratory (NPL) in London, his brief todesign and develop an electronic stored-program digital computer forscientific work. (Artificial Intelligence was not far from Turing'sthoughts: he described himself as ‘building a brain’ andremarked in a letter that he was ‘more interested in thepossibility of producing models of the action of the brain than in thepractical applications to computing’.) John Womersley, Turing'simmediate superior at NPL, christened Turing's proposed machine theAutomatic Computing Engine, or ACE, in homage to Babbage's DifferenceEngine and Analytical Engine.

Turing's 1945 report ‘Proposed Electronic Calculator’gave the first relatively complete specification of an electronicstored-program general-purpose digital computer. The report isreprinted in full in Copeland 2005.

The first electronic stored-program digital computer to be proposedin the U.S. was the EDVAC (see below). The ‘First Draft of aReport on the EDVAC’ (May 1945), composed by von Neumann,contained little engineering detail, in particular concerningelectronic hardware (owing to restrictions in the U.S.). Turing's‘Proposed Electronic Calculator’, on the other hand,supplied detailed circuit designs and specifications of hardware units,specimen programs in machine code, and even an estimate of the cost ofbuilding the machine (£11,200). ACE and EDVAC differedfundamentally from one another; for example, ACE employed distributedprocessing, while EDVAC had a centralised structure.

Turing saw that speed and memory were the keys to computing.Turing's colleague at NPL, Jim Wilkinson, observed that Turing‘was obsessed with the idea of speed on the machine’[Copeland 2005, p. 2]. Turing's design had much in common with today'sRISC architectures and it called for a high-speed memory of roughly thesame capacity as an early Macintosh computer (enormous by the standardsof his day). Had Turing's ACE been built as planned it would have beenin a different league from the other early computers. However, progresson Turing's Automatic Computing Engine ran slowly, due toorganisational difficulties at NPL, and in 1948 a ‘very fedup’ Turing (Robin Gandy's description, in interview withCopeland, 1995) left NPL for Newman's Computing Machine Laboratory atManchester University. It was not until May 1950 that a small pilotmodel of the Automatic Computing Engine, built by Wilkinson, EdwardNewman, Mike Woodger, and others, first executed a program. With anoperating speed of 1 MHz, the Pilot Model ACE was for some time thefastest computer in the world.

Sales of DEUCE, the production version of the Pilot Model ACE, werebuoyant — confounding the suggestion, made in 1946 by theDirector of the NPL, Sir Charles Darwin, that ‘it is verypossible that … one machine would suffice to solve all theproblems that are demanded of it from the whole country’[Copeland 2005, p. 4]. The fundamentals of Turing's ACE design wereemployed by Harry Huskey (at Wayne State University, Detroit) in theBendix G15 computer (Huskey in interview with Copeland, 1998). The G15was arguably the first personal computer; over 400 were sold worldwide.DEUCE and the G15 remained in use until about 1970. Another computerderiving from Turing's ACE design, the MOSAIC, played a role inBritain's air defences during the Cold War period; other derivativesinclude the Packard-Bell PB250 (1961). (More information about theseearly computers is given in [Copeland 2005].)

The Manchester Machine

The earliest general-purpose stored-program electronic digital computerto work was built in Newman's Computing Machine Laboratory atManchester University. The Manchester ‘Baby’, as it becameknown, was constructed by the engineers F.C. Williams and Tom Kilburn,and performed its first calculation on 21 June 1948. The tiny program,stored on the face of a cathode ray tube, was just seventeeninstructions long. A much enlarged version of the machine, with aprogramming system designed by Turing, became the world's firstcommercially available computer, the Ferranti Mark I. The first to becompleted was installed at Manchester University in February 1951; inall about ten were sold, in Britain, Canada, Holland and Italy.

The fundamental logico-mathematical contributions by Turing andNewman to the triumph at Manchester have been neglected, and theManchester machine is nowadays remembered as the work of Williams andKilburn. Indeed, Newman's role in the development of computers hasnever been sufficiently emphasised (due perhaps to his thoroughlyself-effacing way of relating the relevant events).

It was Newman who, in a lecture in Cambridge in 1935, introducedTuring to the concept that led directly to the Turing machine: Newmandefined a constructive process as one that amachine can carryout (Newman in interview with Evans, op. cit.). As a result of hisknowledge of Turing's work, Newman became interested in thepossibilities of computing machinery in, as he put it, ‘a rathertheoretical way’. It was not until Newman joined GC&CS in1942 that his interest in computing machinery suddenly becamepractical, with his realisation that the attack on Tunny could bemechanised. During the building of Colossus, Newman tried to interestFlowers in Turing's 1936 paper — birthplace of the stored-programconcept - but Flowers did not make much of Turing's arcane notation.There is no doubt that by 1943, Newman had firmly in mind the idea ofusing electronic technology in order to construct a stored-programgeneral-purpose digital computing machine.

In July of 1946 (the month in which the Royal Society approvedNewman's application for funds to found the Computing MachineLaboratory), Freddie Williams, working at the TelecommunicationsResearch Establishment, Malvern, began the series of experiments oncathode ray tube storage that was to lead to the Williams tube memory.Williams, until then a radar engineer, explains how it was that he cameto be working on the problem of computer memory:

[O]nce [the German Armies] collapsed … nobody wasgoing to care a toss about radar, and people like me … weregoing to be in the soup unless we found something else to do. Andcomputers were in the air. Knowing absolutely nothing about them Ilatched onto the problem of storage and tackled that. (Quoted inBennett 1976.)

Newman learned of Williams' work, and with the able help of PatrickBlackett, Langworthy Professor of Physics at Manchester and one of themost powerful figures in the University, was instrumental in theappointment of the 35 year old Williams to the recently vacated Chairof Electro-Technics at Manchester. (Both were members of the appointingcommittee (Kilburn in interview with Copeland, 1997).) Williamsimmediately had Kilburn, his assistant at Malvern, seconded toManchester. To take up the story in Williams' own words:

[N]either Tom Kilburn nor I knew the first thing aboutcomputers when we arrived in Manchester University. We'd had enoughexplained to us to understand what the problem of storage was and whatwe wanted to store, and that we'd achieved, so the point now had beenreached when we'd got to find out about computers … Newmanexplained the whole business of how a computer works to us. (F.C.Williams in interview with Evans [1976])

Elsewhere Williams is explicit concerning Turing's role and givessomething of the flavour of the explanation that he and Kilburnreceived:

Tom Kilburn and I knew nothing about computers, but a lotabout circuits. Professor Newman and Mr A.M. Turing … knew a lotabout computers and substantially nothing about electronics. They tookus by the hand and explained how numbers could live in houses withaddresses and how if they did they could be kept track of during acalculation. (Williams [1975], p. 328)

It seems that Newman must have used much the same words withWilliams and Kilburn as he did in an address to the Royal Society on4th March 1948:

Professor Hartree … has recalled that all theessential ideas of the general-purpose calculating machines now beingmade are to be found in Babbage's plans for his analytical engine. Inmodern times the idea of a universal calculating machine wasindependently introduced by Turing … [T]he machines now beingmade in America and in this country … [are] in certain generalrespects … all similar. There is provision for storing numbers,say in the scale of 2, so that each number appears as a row of, say,forty 0's and 1's in certain places or "houses" in the machine.… Certain of these numbers, or "words" are read, one afteranother, as orders. In one possible type of machine an order consistsof four numbers, for example 11, 13, 27, 4. The number 4 signifies"add", and when control shifts to this word the "houses" H11 and H13will be connected to the adder as inputs, and H27 as output. Thenumbers stored in H11 and H13 pass through the adder, are added, andthe sum is passed on to H27. The control then shifts to the next order.In most real machines the process just described would be done by threeseparate orders, the first bringing [H11] (=content of H11) to acentral accumulator, the second adding [H13] into the accumulator, andthe third sending the result to H27; thus only one address would berequired in each order. … A machine with storage, with thisautomatic-telephone-exchange arrangement and with the necessary adders,subtractors and so on, is, in a sense, already a universal machine.(Newman [1948], pp. 271–272)

Following this explanation of Turing's three-address concept (source1, source 2, destination, function) Newman went on to describe programstorage (‘the orders shall be in a series of houses X1, X2,…’) and conditional branching. He then summed up:

From this highly simplified account it emerges that theessential internal parts of the machine are, first, a storage fornumbers (which may also be orders). … Secondly, adders,multipliers, etc. Thirdly, an "automatic telephone exchange" forselecting "houses", connecting them to the arithmetic organ, andwriting the answers in other prescribed houses. Finally, means ofmoving control at any stage to any chosen order, if a certain conditionis satisfied, otherwise passing to the next order in the normalsequence. Besides these there must be ways of setting up the machine atthe outset, and extracting the final answer in useable form. (Newman[1948], pp. 273–4)

In a letter written in 1972 Williams described in some detail whathe and Kilburn were told by Newman:

About the middle of the year [1946] the possibility of anappointment at Manchester University arose and I had a talk withProfessor Newman who was already interested in the possibility ofdeveloping computers and had acquired a grant from the Royal Society of£30,000 for this purpose. Since he understood computers and Iunderstood electronics the possibilities of fruitful collaboration wereobvious. I remember Newman giving us a few lectures in which heoutlined the organisation of a computer in terms of numbers beingidentified by the address of the house in which they were placed and interms of numbers being transferred from this address, one at a time, toan accumulator where each entering number was added to what was alreadythere. At any time the number in the accumulator could be transferredback to an assigned address in the store and the accumulator clearedfor further use. The transfers were to be effected by a stored programin which a list of instructions was obeyed sequentially. Orderedprogress through the list could be interrupted by a test instructionwhich examined the sign of the number in the accumulator. Thereafteroperation started from a new point in the list of instructions. Thiswas the first information I received about the organisation ofcomputers. … Our first computer was the simplest embodiment ofthese principles, with the sole difference that it used a subtractingrather than an adding accumulator. (Letter from Williams to Randell,1972; in Randell [1972], p. 9)

Turing's early input to the developments at Manchester, hinted at byWilliams in his above-quoted reference to Turing, may have been via thelectures on computer design that Turing and Wilkinson gave in Londonduring the period December 1946 to February 1947 (Turing and Wilkinson[1946–7]). The lectures were attended by representatives ofvarious organisations planning to use or build an electronic computer.Kilburn was in the audience (Bowker and Giordano [1993]). (Kilburnusually said, when asked from where he obtained his basic knowledge ofthe computer, that he could not remember (letter from Brian Napper toCopeland, 2002); for example, in a 1992 interview he said:‘Between early 1945 and early 1947, in that period, somehow orother I knew what a digital computer was … Where I got thisknowledge from I've no idea’ (Bowker and Giordano [1993], p.19).)

Whatever role Turing's lectures may have played in informingKilburn, there is little doubt that credit for the Manchester computer— called the ‘Newman-Williams machine’ in acontemporary document (Huskey 1947) — belongs not only toWilliams and Kilburn but also to Newman, and that the influence onNewman of Turing's 1936 paper was crucial, as was the influence ofFlowers' Colossus.

The first working AI program, a draughts (checkers) player writtenby Christopher Strachey, ran on the Ferranti Mark I in the ManchesterComputing Machine Laboratory. Strachey (at the time a teacher at HarrowSchool and an amateur programmer) wrote the program with Turing'sencouragement and utilising the latter's recently completedProgrammers' Handbook for the Ferranti. (Strachey later became Directorof the Programming Research Group at Oxford University.) By the summerof 1952, the program could, Strachey reported, ‘play a completegame of draughts at a reasonable speed’. (Strachey's program formed thebasis for Arthur Samuel's well-known checkers program.) The firstchess-playing program, also, was written for the Manchester Ferranti,by Dietrich Prinz; the program first ran in November 1951. Designed forsolving simple problems of the mate-in-two variety, the program wouldexamine every possible move until a solution was found. Turing startedto program his ‘Turochamp’ chess-player on the FerrantiMark I, but never completed the task. Unlike Prinz's program, theTurochamp could play a complete game (when hand-simulated) and operatednot by exhaustive search but under the guidance of heuristics.

ENIAC and EDVAC

The first fully functioning electronic digital computer to be built inthe U.S. was ENIAC, constructed at the Moore School of ElectricalEngineering, University of Pennsylvania, for the Army OrdnanceDepartment, by J. Presper Eckert and John Mauchly. Completed in 1945,ENIAC was somewhat similar to the earlier Colossus, but considerablylarger and more flexible (although far from general-purpose). Theprimary function for which ENIAC was designed was the calculation oftables used in aiming artillery. ENIAC was not a stored-programcomputer, and setting it up for a new job involved reconfiguring themachine by means of plugs and switches. For many years, ENIAC wasbelieved to have been the first functioning electronic digitalcomputer, Colossus being unknown to all but a few.

In 1944, John von Neumann joined the ENIAC group. He had become‘intrigued’ (Goldstine's word, [1972], p. 275) withTuring's universal machine while Turing was at Princeton Universityduring 1936–1938. At the Moore School, von Neumann emphasised theimportance of the stored-program concept for electronic computing,including the possibility of allowing the machine to modify its ownprogram in useful ways while running (for example, in order to controlloops and branching). Turing's paper of 1936 (‘On ComputableNumbers, with an Application to the Entscheidungsproblem’) wasrequired reading for members of von Neumann's post-war computer projectat the Institute for Advanced Study, Princeton University (letter fromJulian Bigelow to Copeland, 2002; see also Copeland [2004], p. 23).Eckert appears to have realised independently, and prior to vonNeumann's joining the ENIAC group, that the way to take full advantageof the speed at which data is processed by electronic circuits is toplace suitably encoded instructions for controlling the processing inthe same high-speed storage devices that hold the data itself(documented in Copeland [2004], pp. 26–7). In 1945, while ENIACwas still under construction, von Neumann produced a draft report,mentioned previously, setting out the ENIAC group's ideas for anelectronic stored-program general-purpose digital computer, the EDVAC(von Neuman [1945]). The EDVAC was completed six years later, but notby its originators, who left the Moore School to build computerselsewhere. Lectures held at the Moore School in 1946 on the proposedEDVAC were widely attended and contributed greatly to the disseminationof the new ideas.

Von Neumann was a prestigious figure and he made the concept of ahigh-speed stored-program digital computer widely known through hiswritings and public addresses. As a result of his high profile in thefield, it became customary, although historically inappropriate, torefer to electronic stored-program digital computers as ‘vonNeumann machines’.

The Los Alamos physicist Stanley Frankel, responsible with vonNeumann and others for mechanising the large-scale calculationsinvolved in the design of the atomic bomb, has described von Neumann'sview of the importance of Turing's 1936 paper, in a letter:

I know that in or about 1943 or ‘44 von Neumann waswell aware of the fundamental importance of Turing's paper of 1936… Von Neumann introduced me to that paper and at his urging Istudied it with care. Many people have acclaimed von Neumann as the"father of the computer" (in a modern sense of the term) but I am surethat he would never have made that mistake himself. He might well becalled the midwife, perhaps, but he firmly emphasized to me, and toothers I am sure, that the fundamental conception is owing to Turing,in so far as not anticipated by Babbage … Both Turing and vonNeumann, of course, also made substantial contributions to the"reduction to practice" of these concepts but I would not regard theseas comparable in importance with the introduction and explication ofthe concept of a computer able to store in its memory its program ofactivities and of modifying that program in the course of theseactivities. (Quoted in Randell [1972], p. 10)

Other Notable Early Computers

Other notable early stored-program electronic digital computers were:

EDSAC, 1949, built at Cambridge University by Maurice Wilkes
BINAC, 1949, built by Eckert's and Mauchly's Electronic ControlCo., Philadelphia (opinions differ over whether BINAC ever actuallyworked)
Whirlwind I, 1949, Digital Computer Laboratory, MassachusettsInstitute of Technology, Jay Forrester
SEAC, 1950, US Bureau of Standards Eastern Division, WashingtonD.C., Samuel Alexander, Ralph Slutz
SWAC, 1950, US Bureau of Standards Western Division, Institute forNumerical Analysis, University of California at Los Angeles, HarryHuskey
UNIVAC, 1951, Eckert-Mauchly Computer Corporation, Philadelphia(the first computer to be available commercially in the U.S.)
the IAS computer, 1952, Institute for Advanced Study, PrincetonUniversity, Julian Bigelow, Arthur Burks, Herman Goldstine, vonNeumann, and others (thanks to von Neumann's publishing thespecifications of the IAS machine, it became the model for a group ofcomputers known as the Princeton Class machines; the IAS computer wasalso a strong influence on the IBM 701)
IBM 701, 1952, International Business Machine's first mass-producedelectronic stored-program computer.

High-Speed Memory

The EDVAC and ACE proposals both advocated the use of mercury-filledtubes, called ‘delay lines’, for high-speed internalmemory. This form of memory is known as acoustic memory. Delay lineshad initially been developed for echo cancellation in radar; the ideaof using them as memory devices originated with Eckert at the MooreSchool. Here is Turing's description:

It is proposed to build "delay line" units consisting ofmercury … tubes about 5′ long and 1″ in diameter incontact with a quartz crystal at each end. The velocity of sound in… mercury … is such that the delay will be 1.024 ms. Theinformation to be stored may be considered to be a sequence of 1024‘digits’ (0 or 1) … These digits will be representedby a corresponding sequence of pulses. The digit 0 … will berepresented by the absence of a pulse at the appropriate time, thedigit 1 … by its presence. This series of pulses is impressed onthe end of the line by one piezo-crystal, it is transmitted down theline in the form of supersonic waves, and is reconverted into a varyingvoltage by the crystal at the far end. This voltage is amplifiedsufficiently to give an output of the order of 10 volts peak to peakand is used to gate a standard pulse generated by the clock. This pulsemay be again fed into the line by means of the transmitting crystal, orwe may feed in some altogether different signal. We also have thepossibility of leading the gated pulse to some other part of thecalculator, if we have need of that information at the time. Making useof the information does not of course preclude keeping it also. (Turing[1945], p. 375)

Mercury delay line memory was used in EDSAC, BINAC, SEAC, PilotModel ACE, EDVAC, DEUCE, and full-scale ACE (1958). The chief advantageof the delay line as a memory medium was, as Turing put it, that delaylines were "already a going concern" (Turing [1947], p. 380). Thefundamental disadvantages of the delay line were that random access isimpossible and, moreover, the time taken for an instruction, or number,to emerge from a delay line depends on where in the line it happens tobe.

In order to minimize waiting-time, Turing arranged for instructionsto be stored not in consecutive positions in the delay line, but inrelative positions selected by the programmer in such a way that eachinstruction would emerge at exactly the time it was required, in so faras this was possible. Each instruction contained a specification of thelocation of the next. This system subsequently became known as‘optimum coding’. It was an integral feature of everyversion of the ACE design. Optimum coding made for difficult and untidyprogramming, but the advantage in terms of speed was considerable.Thanks to optimum coding, the Pilot Model ACE was able to do a floatingpoint multiplication in 3 milliseconds (Wilkes's EDSAC required 4.5milliseconds to perform a single fixed point multiplication).

In the Williams tube or electrostatic memory, previously mentioned,a two-dimensional rectangular array of binary digits was stored on theface of a commercially-available cathode ray tube. Access to data wasimmediate. Williams tube memories were employed in the Manchesterseries of machines, SWAC, the IAS computer, and the IBM 701, and amodified form of Williams tube in Whirlwind I (until replacement bymagnetic core in 1953).

Drum memories, in which data was stored magnetically on the surfaceof a metal cylinder, were developed on both sides of the Atlantic. Theinitial idea appears to have been Eckert's. The drum providedreasonably large quantities of medium-speed memory and was used tosupplement a high-speed acoustic or electrostatic memory. In 1949, theManchester computer was successfully equipped with a drum memory; thiswas constructed by the Manchester engineers on the model of a drumdeveloped by Andrew Booth at Birkbeck College, London.

The final major event in the early history of electronic computationwas the development of magnetic core memory. Jay Forrester realisedthat the hysteresis properties of magnetic core (normally used intransformers) lent themselves to the implementation of athree-dimensional solid array of randomly accessible storage points. In1949, at Massachusetts Institute of Technology, he began to investigatethis idea empirically. Forrester's early experiments with metallic coresoon led him to develop the superior ferrite core memory. DigitalEquipment Corporation undertook to build a computer similar to theWhirlwind I as a test vehicle for a ferrite core memory. The MemoryTest Computer was completed in 1953. (This computer was used in 1954for the first simulations of neural networks, by Belmont Farley andWesley Clark of MIT's Lincoln Laboratory (see Copeland and Proudfoot[1996]).

Once the absolute reliability, relative cheapness, high capacity andpermanent life of ferrite core memory became apparent, core soonreplaced other forms of high-speed memory. The IBM 704 and 705computers (announced in May and October 1954, respectively) broughtcore memory into wide use.

Bibliography

Works Cited

Babbage, C. (ed. by Campbell-Kelly, M.), 1994,Passages fromthe Life of a Philosopher, New Brunswick: Rutgers UniversityPress
Bennett, S., 1976, ‘F.C. Williams: his contribution to thedevelopment of automatic control’, National Archive for theHistory of Computing, University of Manchester, England. (This is atypescript based on interviews with Williams in 1976.)
Bowker, G., and Giordano, R., 1993, ‘Interview with TomKilburn’,Annals of the History of Computing,15: 17–32.
Copeland, B.J. (ed.), 2004,The Essential Turing OxfordUniversity Press
Copeland, B.J. (ed.), 2005,Alan Turing's Automatic ComputingEngine: The Master Codebreaker's Struggle to Build the ModernComputer Oxford University Press
Copeland, B.J. and others, 2006,Colossus: The Secrets ofBletchley Park's Codebreaking Computers Oxford UniversityPress
Copeland, B.J., and Proudfoot, D., 1996, ‘On Alan Turing'sAnticipation of Connectionism’Synthese,108: 361–377
Evans, C., 197?, interview with M.H.A. Newman in ‘ThePioneers of Computing: an Oral History of Computing’, London:Science Museum
Fifer, S., 1961,Analog Computation: Theory, Techniques,Applications New York: McGraw-Hill
Ford, H., 1919, ‘Mechanical Movement’,OfficialGazette of the United States Patent Office, October 7, 1919:48
Goldstine, H., 1972,The Computer from Pascal to vonNeumann Princeton University Press
Huskey, H.D., 1947, ‘The State of the Art in ElectronicDigital Computing in Britain and the United States’, in [Copeland2005]
Newman, M.H.A., 1948, ‘General Principles of the Design ofAll-Purpose Computing Machines’Proceedings of the RoyalSociety of London, series A,195 (1948):271–274
Randell, B., 1972, ‘On Alan Turing and the Origins of DigitalComputers’, in Meltzer, B., Michie, D. (eds),MachineIntelligence 7, Edinburgh: Edinburgh University Press, 1972
Smith, B.C., 1991, ‘The Owl and the ElectricEncyclopaedia’,Artificial Intelligence,47: 251–288
Thomson, J., 1876, ‘On an Integrating Machine Having a NewKinematic Principle’Proceedings of the Royal Society ofLondon,24: 262–5
Turing, A.M., 1936, ‘On Computable Numbers, with anApplication to the Entscheidungsproblem’Proceedings of theLondon Mathematical Society, Series 2,42(1936–37): 230–265. Reprinted inThe EssentialTuring (Copeland [2004]).
Turing, A.M, 1945, ‘Proposed Electronic Calculator’, inAlan Turing's Automatic Computing Engine (Copeland[2005])
Turing, A.M., 1947, ‘Lecture on the Automatic ComputingEngine’, inThe Essential Turing (Copeland [2004])
Turing, A.M., and Wilkinson, J.H., 1946–7, ‘TheTuring-Wilkinson Lecture Series (1946-7)’, inAlan Turing'sAutomatic Computing Engine (Copeland [2005])
von Neumann, J., 1945, ‘First Draft of a Report on theEDVAC’, in Stern, N.From ENIAC to UNIVAC: An Appraisal ofthe Eckert-Mauchly Computers Bedford, Mass.: Digital Press (1981),pp. 181–246
Williams, F.C., 1975, ‘Early Computers at ManchesterUniversity’The Radio and Electronic Engineer,45 (1975): 237–331
Wynn-Williams, C.E., 1932, ‘A Thyratron "Scale of Two"Automatic Counter’Proceedings of the Royal Society ofLondon, series A,136: 312–324

Academic Tools

How to cite this entry.
Preview the PDF version of this entry at theFriends of the SEP Society.
Look up topics and thinkers related to this entry at the Internet Philosophy Ontology Project (InPhO).
Enhanced bibliography for this entry atPhilPapers, with links to its database.

Other Internet Resources

Open access to the SEP is made possible by a world-wide funding initiative.
The Encyclopedia Now Needs Your Support
Please Read How You Can Help Keep the Encyclopedia Free

Browse

About

Support SEP

Mirror Sites

View this site from another server:

USA (Main Site)Philosophy, Stanford University

Info about mirror sites

Library of Congress Catalog Data: ISSN 1095-5054

Movatterモバイル変換