future wireless technology

The History of Computing in the History of Technology

Michael S. MahoneyProgram in History of Science

Princeton University, Princeton, NJ

(Annals of the History of Computing 10(1988), 113-125)

After surveying the current state of the literature in the history of computing, thispaper discusses some of the major issues addressed by recent work in the history oftechnology. It suggests aspects of the development of computing which are pertinentto those issues and hence for which that recent work could provide models ofhistorical analysis. As a new scientific technology with unique features, computingin turn can provide new perspectives on the history of technology.

Introduction

Since World War II 'information' has emergedas a fundamental scientific and technologicalconcept applied to phenomena ranging fromblack holes to DNA, from the organization ofcells to the processes of human thought, andfrom the management of corporations to theallocation of global resources. In addition toreshaping established disciplines, it hasstimulated the formation of a panoply of newsubjects and areas of inquiry concerned withits structure and its role in nature and society(Machlup and Mansfeld 1983). Theories basedon the concept of 'information' have sopermeated modern culture that it now iswidely taken to characterize our times. Welive in an 'information society', an 'age ofinformation'. Indeed, we look to models ofinformation processing to explain our ownpatterns of thought.

The computer has played the centralro le in that tr ansfo rmat ion, bothaccommodating and encouraging ever broaderviews of 'information' and of how it can betransformed and communicated over time andspace. Since the 1950s the computer hasreplaced traditional methods of accounting and

record-keeping by a new industry of dataprocessing. As a primary vehicle ofcommunication over both space and t ime, ithas come to form the core of moderninformation technology. What theEnglish-speaking world refers to as "computerscience" is known to the rest of westernEurope as informatique (or Informatik orinformatica). Much of the concern overinformation as a commodity and as a naturalresource derives from the computer and fromcomputer-based communications technology.1

Hence, the history of the computer and ofcomputing is central to that of informationscience and technology, providing a thread bywhich to maintain bearing while exploring theever-growing maze of disciplines andsubdisciplines that claim information as theirsubject.

Despite the pervasive presence ofcomputing in modern science and technology,not to mention modern society itself, thehistory of computing has yet to establish asignificant presence in the history of science

1To characterize the unprecedented capabilities ofcomputers linked to telecommunications, Nora andMinc (1978) coined the term télématique.

Mahoney History of Computing in the History of Technology page 2

and technology. Meetings of the History ofScience Society and the Society for theHistory of Technology in recent years haveincluded very few sessions devotedspecifically to history of computing, and fewof the thematic sessions have includedcontributions from the perspective ofcomputing. There is clearly a balance to beredressed here.

The status of the history of computingwithin the history of technology surely reflectson both parties, but the bulk of the task ofredress lies with the former. A look at theliterature shows that , by and large, historiansof computing are addressing few of thequestions that historians of technology arenow asking. It is worth looking at what thosequestions are and what form they might takewhen addressed to computing. The question ishow to bring the history of computing into linewith what should be its parent discipline.Doing so will follow a two-way street: thehistory of computing should use models fromthe history of technology at the same time thatwe use the history of computing to test thosemodels. In some aspects, at least, computingposes some of the major questions of thehistory of technology in special ways. Bothfields have much to learn from the other.

Computing's Present History

Where the current literature in the history ofcomputing is self-consciously historical, itfocuses in large part on hardware and on thepre-history and early development of thecomputer.2 Where it touches on laterdevelopments or provides a wider view, it isonly incidentally historical. A major portion ofthe literature stems from the people involved,

either through regular surveys of the state anddevelopment of various fields (e.g. Rosen1967, Sammet 1969)3 or compilations ofseminal papers (Randell 1982; Yourdon 1979,1982; AT&T 1987),4 or throughreminiscences and retrospectives, eitherwritten directly or transcribed from theircontributions to conferences and symposia.5

Biographies of men or machines --someheroic, some polemical, some both-- are aprominent genre, and one reads a lot about"pioneers". A few corporate histories haveappeared, most notably IBM's EarlyComputers (Bashe et al. 1986), but they tooare in-house productions.

This literature represents for the mostpart "insider" history, full of facts and firsts.While it is first-hand and expert, it is alsoguided by the current state of knowledge andbound by the professional culture. That is, itsauthors take as givens (often technical givens)what a more critical, outside viewer might seeas choices. Reading their accounts makes itdifficult to see the alternatives, as the authorsthemselves lose touch with a time when theydid not know what they now know. In the longrun, most of this literature will become primarysources, if not of the development ofcomputing per se, then of its emerging culture.

From the outset, the computerattracted the at tention of journalists, who bythe late '50s were beginning to recount itshistory. The result is a sizable inventory of

2See Aspray (1984) for a recent, brief survey of thestate of the field.

3Many of the articles in Computing Surveys, begun in1969, include an historical review of the subject.

4The 25th-anniversary issues of the leading journalsalso contain useful collections of importan t articles.

5Wexelblatt (1981), a record of the 1978 ACMConference on the History of Programming Languages,is an excellent example, as is a recent issue of theAnnals of the History of Computing on the BurroughsB5000.


accounts having the virtues and vices of thejournalist's craft. They are vivid, they capturethe spirit of the people and of the institutionsthey portray, and they have an eye for thetelling anecdote. But their immediacy comes atthe price of perspective. Written by peoplemore or less knowledgeable about the subjectand about the history of technology, theseaccounts tend to focus on the unusual and thespectacular, be it people or lines of research,and they often cede to the self-evaluation oftheir subjects. Thus the microcomputer andartificial intelligence have had the lion's shareof attention, as their advocates have roared asuccession of millenia.

The journalistic accounts veer intoanother major portion of the literature oncomputing, namely what may be called "socialimpact statements". Often difficult todistinguish from futurist musing on thecomputer, the discussions of the effects of thecomputer on society and its various activitiestend on the whole to view computing apartfrom the history of technology rather thanfrom its perspective. History here serves thepurpose of social analysis, criticism, andcommentary. Hence much of it comes frompopular accounts taken uncritically andepisodically to support non-historical, oftenpolemical, theses. Some of this literature restson a frankly political agenda; whether itsmodels and modes of analysis provide insightdepends on whether one agrees with thatagenda.

Finally, there is a small body ofprofessionally historical work, dealing for themost part with the origins of the computer, itsinvention and early development (e.g. Stern1981, Ceruzzi 1982, Williams 1986). It ismeant as no denigration of that work to notethat it stops at the point where computingbecomes a significant presence in science,technology, and society. There historians standbefore the daunting complexity of a subject

that has grown exponentially in size andvariety, looking not so much like an unchartedocean as like a trackless jungle. We pace onthe edge, pondering where to cut in.

The Questions of the History ofTechnology

The state of the literature in history ofcomputing emerges perhaps more clearly bycomparison (and by contrast) with what iscurrently appearing in the history oftechnology in general and with the questionsthat have occupied historians of technologyover the past decade or so. Those questionsderive from a cluster of seminal articles byGeorge S. Daniels, Edwin T. Layton, Jr.,Eugene S. Ferguson, Nathan Rosenberg, andThomas P. Hughes, among others. How hasthe relationship between science andtechnology changed and developed over timeand place? How has engineering evolved,both as an intellectual activity and as a socialrole? Is technology the creator of demand ora response to it? Put another way, doestechnology follow a society's momentum orredirect it by external impulse?6 How far doeseconomics go in explaining technologicalinnovation and development? How do newtechnologies establish themselves in society,and how does society adapt to them? To whatextent and in what ways do societies engendernew technologies? What are the patterns bywhich technology is transferred from oneculture to another? What role do governmentsplay in fostering and directing technological

6George Daniels (1970) put th e question as anassertion (p.6): "... the real effect of technicalinnovation [has been] to help Americans do betterwhat they had already shown a marked inclination todo." The seeming "social lag" in adapting to newtechnology, he argued, is more likely economic innature.


innovation and development? These are someof the "big questions", as George Daniels(1970) once put it. They can be broken downinto smaller, more manageable questions, butultimately they are the questions for whichhistorians of technology bear specialresponsibility within the historical community.They are all of them questions which can shedlight on the development of computing whileit in turn elucidates them.

A few examples from recent literaturemust suffice to suggest the approacheshistorians of technology are taking to thosequestions. Each suggests by implication whatmight be done in the history of computing. Aspate of studies on industrial researchlaboratories has explored the sources,purposes and strategies of organizedinnovation, invention, and patenting in the late19th and early 20th centuries, bringing out thedynamics of technological improvement thatRosenberg (1979) suggested was a majorsource of growth in productivity. In Networksof Power Thomas P. Hughes (1983) hasprovided a model for pursuing anothersuggestion by Rosenberg, namely the need totreat technologies as interactive constituentsof systems. Developments in one subsystemmay be responses to demands in others andhence have their real pay-offs there. Or abreakthrough in one component of the systemmay unexpectedly create new opportunities inthe others, or even force a reorganization ofthe system itself.

In detailed examinations of one of the"really big questions" of the history ofAmerican technology, Merritt Roe Smith(1977) and David A. Hounshell (1984) havetraced the origins of the "American System"and its evolution into mass production and theassembly line. Both have entered theworkshops and factories to reveal the quiteuneven reception and progress of that system,never so monolithic or pervasive as it seemed

then or has seemed since. Daniel Nelson(1975) and Stephen Meyer (1981) haveentered the factory floor by another door tostudy the effects of mass production on theworkers it organized.

Looking at technology in othercontexts, Walter McDougall (1985) hasanatomized the means and motivation ofgovernment support of research anddevelopment since World War II, revealingstructures and patterns that extend wellbeyond the space program. Behind his studystands the ongoing history of NASA and of itsindividual projects. From another perspective,David F. Noble (1984) has examined the"command technology" that lay behind thedevelopment of numerically controlled tools.At a more mundane level, Ruth Cowan (1983)has shown how "progress is our mostimportant product" often translated into MoreWork for Mother, while her own experimentsin early nineteenth-century domestictechnology have brought out the intimaterelationship between household work andfamily relations.

In the late 1970s Anthony F.C.Wallace (1978) and Eugene Ferguson (1979b)recalled our attention to the non-verbal modesof thought that seem more characteristic of theinventor and engineer than does thelanguage-based thinking of the scientist.7

Brooke Hindle's (1981) study of Morse'stelegraph and Reese Jenkins's (1987) recentwork on the iconic patterns of Edison'sthought provide examples of the insightshistorians can derive from artifacts read as theconcrete expressions of visual and tactilecognition, recognizing that, as Henry Fordonce put it,

7See in particular Wallace's "Thinking AboutMachinery" (Wallace 1978, pp.237ff.).


There is an immense amountto be learned simply bytinkering with things. It is notpossible to learn from bookshow everything is made --anda real mechanic ought to knowhow nearly everything is made.Machines are to a mechanicwhat books are to a writer. Hegets ideas from them, and if hehas any brains he will applythose ideas (Ford 1922, p.24).8

The renewed emphasis on the visual hasreinforced the natural ties between thehistorian of technology and the museum, at thesame time that it has forged links betweenhistory of technology and the study of materialculture.

The Tripartite Nature of Computing

Before trying to translate some of the abovequestions and models into forms specific to thehistory of computing, it may help to reflect abit on the complexity of the object of ourstudy. The computer is not one thing, butmany different things, and the same holds trueof computing. There is about both terms adeceptive singularity to which we fall victimwhen, as is now common, we prematurelyunite its multiple historical sources into asingle stream, treating Charles Babbage's

analytical engine and George Boole's algebraof thought as if they were conceptually relatedby something other than 20th-centuryhindsight. Whatever John von Neumann'sprecise role in designing the "von Neumannarchitecture" that defines the computer for theperiod with which historians are properlyconcerned, it is really only in von Neumann'scollaboration with the ENIAC team that twoquite separate historical strands came together:the effort to achie ve high-speed,high-precision, automatic calculation and theeffort to design a logic machine capable ofsignificant reasoning.9

The dual nature of the computer isreflected in its dual origins: hardware in thesequence of devices that stretches from thePascaline to the ENIAC, software in the seriesof investigations that reaches from Leibniz'scombinatorics to Turing's abstract machines.Until the two strands come together in thecomputer, they belong to different histories,the electronic calculator to the history oftechnology, the logic machine to the history ofmathematics,10 and they can be unfoldedseparately without significant loss of fullnessor texture. Though they come together in thecomputer, they do not unite. The computerremains an amalgam of technological deviceand mathematical concept, which retainseparate identities despite their influence onone another.

Thus the computer in itself embodiesone of the central problems of the history oftechnology, namely the relation of science and

8In The Sciences of the Artificial Herbert Simon (1981;cf. Newell and Simon 1976) argues forcefully for theempirical nature of computer research th at under liesits mathematical trappings. The thinking of computerdesigners and programmers is embodied in the waytheir machines and programs work, and th e languagesthey use to specify how things are to work arethemselves artifacts. The models they use are fil ledwith images difficult or distractingly tedious totranslate into words; cf. Bolter (1984).

9I do not make this claim in ignorance of KonradZuse's Z4 or Alan Turing's ACE, which realizedroughly the same goals as von Neumann's alongindependent paths. Clearly the computer was "in theair" by the 1940s. But it was the 1940s, not the 1840s.

10I am including the history of mathematical logic inthe history of mathematics


technology.11 Computing as an enterprisedeepens the problem. For not only are finiteautomata or denot at iona l semanticsindependent of integrated circuits, they arealso linked in only the most tenuous anduncertain way to programs and programming,that is, to software and its production. Sincethe mid-1960s experience in this realm hasrevealed a third strand in the nature of thecomputer. Between the mathematics thatmakes the device theoretically possible andthe electronics that makes it practically feasiblelies the programming that makes itintellectually, economically, and sociallyuseful. Unlike the extremes, the middleremains a craft, technical rather thantechno logical, mathematical only inappearance. It poses the question of therelation of science and technology in a veryspecial form.

That tripartite structure shows up inthe three distinct disciplines that are concernedwith the computer: electrical engineering,computer science, and software engineering.Of these, the first is the most well established,since it predates the computer, even though itscurrent focus on microelectronics reflects itsbasic orientation toward the device. Computerscience began to take shape during the 1960s,as it brought together common concerns frommathematical logic (automata, proof theory,recursive function theory), mathematicallinguistics, and numerical analysis (algorithms,computational complexity), adding to themquestions of the organization of information

(data structures) and the relation of computerarchitecture to patterns of computation.Software engineering, conceived as adeliberately provocative term in 1967 (Naurand Randell 1969), has developed more as aset of techniques than as a body of learning.Except for a few university centers, such asCarnegie-Mellon University, University ofNorth Carolina, Berkeley, and Oxford, itremains primarily a concern of military andindustrial R&D aimed at the design andimplementation of large, complex systems, andthe driving forces are cost and reliability.

History of Computing as History ofTechnology

Consider, then, the history of computing inlight of current history of technology. Severallines of inquiry seem particularly promising.Studies such as those cited above offer apanoply of models for tracing the patterns ofgrowth and progress in computing as atechnology. It is worth asking, for example,whether the computing industry has movedforward more by big advances of radicalinnovation or by small steps of improvement.Has it followed the process described byNathan Rosenberg, whereby "... technologicalimprovement not only enters the structure ofthe economy through the main entrance, aswhen it takes the highly visible form of majorpatentable technological breakthroughs, butthat it also employs numerous and less visibleside and rear entrances where its arrival isunobtrusive, unannounced, unobserved, anduncelebrated" (Rosenberg 1979, p.26)? Todetermine whether that is the case will requirechanges in the history of computing as it iscurrently practiced. It will mean lookingbeyond "firsts" to the revisions andmodifications that made products work andthat account for their real impact. Given thecorporate, collaborat ive structure of modern

11It should sharpen the question for the history ofscience as well, if only by giving special force to thereciprocal influence of scientific theory and scientificinstrumentation. But up to now at least it has notattracted the same attention. The computer may wellchange that as the shaping of scientific concepts andthe pursuit of scientific inquiry come to depend on thesta te of computer technology.


R&D, historians of computing must follow theadmonition once made to historians oftechnology to stop "substituting biography forcareful analysis of social processes". Withoutdenigrat ing the role of heroes and pioneers, weneed more knowledge of computing'sequivalent of "shop practices, [and of] theactivities of lower-level technicians infactories" (Daniels 1970, p.11). The questionis how to pursue that inquiry across thevariegated range of the emerging industry.

Viewing computing both as a systemin itself and as a component of a variety oflarger systems may provide important insightsinto the the dynamics of its development andmay help to distinguish between its internaland its external history. For example, itsuggests an approach to the question of therelation between hardware and software, oftencouched in the antagonistic form of onedriving the other, a form which seems toassume that the two are relatively independentof one another. By contrast, linking them in asystem emphasizes their mutual dependence.One expects of a system that the relationshipamong its internal components and theirrelationships to external components will varyover time and place but that they will do so ina way that maintains a certain equilibrium orhomeostasis, even as the system itself evolves.Seen in that light, the relation betweenhardware and software is a question not somuch of driving forces, or of stimulus andresponse, as of constraints and degrees offreedom. While in principle all computers havethe same capacities as universal Turingmachines, in practice different architectures areconducive to different forms of computing.Certain architectures have technical thresholds(e.g. VSLI is a prerequisite to massivelyparallel computing), others reflect consciouschoices among equally feasible alternatives;some have been influenced by the needs andconcerns of software production, others by

the special purposes of customers. Early on,programming had to conform to the narrowlimits of speed and memory set byvacuum-tube circuitry. As largely exogenousfactors in the electronics industry made itpossible to expand those limits, and at thesame time drastically lowered the cost ofhardware, programming could take practicaladvantage of research into programminglanguages and compilers. Researchers' ideas ofmultiuser systems, interactive programming, orvirtual memory required advances in hardwareat the same time that they drew out the fullpower of a new generat ion of machines. Justas new architectures have challengedestablished forms of programming, so tootheoretical advances in computat ion andartificial intelligence have suggested new waysof organizing processors (e.g. Backus 1977).

At present, the evolution of computingas a system and of its interfaces with othersystems of thought and action has yet to betraced. Indeed, it is not clear how manyidentifiable systems constitute computingitself, given the diverse contexts in which it hasdeveloped. We speak of the computer industryas if it were a monolith rather than a networkof interdependent industries with separateinterests and concerns. In addition tohistorically more analytical studies ofindividual firms, both large and small, we needanalyses of t heir interaction andinterdependence. The same holds forgovernment and academia, neither of whichhas spoken with one voice on matters ofcomputing. Of particular interest here may bethe system-building role of the computer inforging new links of interdependence amonguniversities, government, and industry afterWorld War II.

Arguing in "The Big Questions" thatcreators of the machinery underpinning theAmerican System worked from a knowledgeof the entire sequence of operations in


production,12 Daniels (1970) pointed to PeterDrucker's suggestion that "the organization ofwork be used as a unifying concept in thehistory of technology." The recent volume byCharles Bashe et al. on IBM's EarlyComputers illustrates the potential fruitfulnessof that suggestion for the history ofcomputing. In tracing IBM's adaptation to thecomputer, they bring out the corporatetensions and adjustments introduced into IBMby the need to keep abreast of fast-breakingdevelopments in science and technology and inturn to share its research with others.13 Thecomputer reshaped R&D at IBM, defining newrelations between marketing and research,introducing a new breed of scientific personnelwith new ways of doing things, and creatingnew roles, in particular that of theprogrammer. Whether the same holds true of,say, Bell Laboratories or G.E. ResearchLaboratories, remains to be studied, as doesthe structure of the R&D institutionsestablished by the many new firms thatconstituted the growing computer industry ofthe '50s, '60s, and '70s. Tracy Kidder's (1981)frankly journalistic account of development atData General has given us a tantalizingglimpse of the patterns we may find. Equallyimportant will be studies of the emergence ofthe data-processing shop, whether as anindependent computer service or as a newelement in established institutions.14 More

than one company found that the computerreorganized de facto the lines of effectivemanagerial power.

The computer seems an obvious placeto look for insight into the question of whethernew technologies respond to need or create it.Clearly, the first computers responded to thefelt need for high-speed, automat ic calculation,and that remained the justification for theirearly development during the late '40s.Indeed, the numerical analysts evidentlyconsidered the computer to be their baby andresented its adoption by "computerologists" inthe late '50s and early '60s (Wilkinson 1971).But it seems equally clear that the computerbecame the core of an emergentdata-processing industry more by creatingdemand than by responding to it. Much asHenry Ford taught the nation how to use anautomobile, IBM and its competitors taughtthe nation's businesses (and its government)how to use the computer. How much of thetechnical development of the computeroriginated in the marketing division remains anuntold story central to an understanding ofmodern technology.15 Kidder's Soul of a NewMachine again offers a glimpse of what thatstory may reveal.

One major factor in the creation ofdemand seems to have been the alliance

12Elting E. Morison (1974) has pursued this pointalong slight ly different but equally revealing lines.

13Lundstrom (1987) has recently chronicled the failureof some companies to make the requisite adjustments.

14The obvious citations here are Kraft (1977) andGreenbaum (1979), but both works are concerned morewith politics than with computing, and the focus oftheir political concerns, the "deskilling" ofprogrammers through the impos ition of methods ofstructured programming, has proved ephemeral, as

subsequent experience and data show thatprogrammers have made the transition with nosignificant loss of control over their work; cf. Boehm(1981).

15See, for example, Burke (1970): "Thus technologicalinnovation is not the product of society as a whole butemanates rather from certain segments within oroutside of it; the men or institutions responsible for theinnovation, to be successful, must 'sell' it to the generalpublic; and innovation does have the effect of creatingbroad social change.(p.23)" Ferguson (1979a) hasmade a similar observation about selling newtechnology.


between the computer and the nascent field ofoperations research/management science. Asthe pages of the Harvard Business Review for1953 show, the computer and operat ionsresearch hit the business stage together, eacha new and untried tool of management, bothclothed in the mantle of science. Against thefanciful backdrop of Croesus' defeat bycamel-riding Persians, an IBM advertisementproclaimed that "Yesterday ... 'The Fates'Decided. Today ... Facts Are What Count". Appealing to fact-based strides in "militaryscience, pure science, commerce, andindustry", the advertisement pointed beyonddata processing to "'mathematical models' ofspecific processes, products, or situations, [bywhich] man today can predetermine probableresults, minimize risks and costs." In less vividterms, Cyril C. Herrmann of MIT and John F.Magee of Arthur D. Little introduced readersof HBR to "'Operations Research' forManagement" (1953), and John Diebold(1953) proclaimed "Automation - The NewTechnology". As Herbert Simon (1960, p.14)later pointed out, operations research wasboth old and new, with roots going back toCharles Babbage and Frederick W. Taylor. Itsnovelty lay precisely in its claim to provide'mathematical models' of business operationsas a basis for rational decision-making.Depending for t heir sensit ivity oncomputationally intensive algorithms and largevolumes of data, those models required thepower of the computer.

It seems crucial for the developmentof the computer industry that the businesscommunity accepted the joint claims of ORand the computer long before either couldvalidate them by, say, cost-benefit analysis.The decision to adopt the new methods of"rational decision-making" seems itself to havebeen less than fully rational:

As business managers we are

revolutionizing the proceduresof our factories and officeswith automation, but whatabout out decision making? Inother words, isn't there adanger that our thoughtprocesses will be left in thehorse-and-buggy stage whileour operations are being runin the age of nucleonics,electronics, and jet propulsion?... Are the engineering andscientific symbols of our agesignificant indicators of a needfor change? (Hurni 1955, p.49)

Even at this early stage, the computer hadacquired symbolic force in the businesscommunity and in society at large. We need toknow the sources of that force and how itworked to weave the computer into theeconomic and social fabric.16

The government has played adetermining role in at least four areas ofcomputing: microelectronics; interactive,real-time systems; artificial intelligence; andsoftware engineering. None of these storieshas been told by an historian, although eachpromises deep insight into the issues raisedabove. Modern weapons systems and thespace program placed a premium onminiaturization of circuits. Given the costs ofresearch, development, and tooling forproduction, it is hard to imagine that the

16Along these lines, historians of computing would dowell to remember that a line of writings on the nature,impact, and even history of computing stretch ing fromEdmund C. Berkeley's (1949) Giant Brains throughJohn Diebold's several volumes to EdwardFeigenbaum's and Pamela McCorduck's (1983) TheFifth Generation stems from people with a product tosell, whether management consulting or expertsystems.

Mahon ey History of Comput ing in the History of Technology page 10

integrated circuit and the microprocessorwould have emerged --at least as quickly asthey did-- without government support. AsFrank Rose (1984) put it in Into the Heart ofthe Mind, "The computerization of society ...has essentially been a side effect of thecomputerization of war.(p.36)" More isinvolved than smaller computers. Architectureand software change in response to speed ofprocessor and size of memory. As a result, therapid pace of miniaturization tended to placealready inadequate methods of softwareproduction under the pressure of risingexpectations. By the early 1970s theDepartment of Defense, as the nation's singlelargest procurer of software, had declared amajor stake in the development of softwareengineering as a body of methods and tools forreducing the costs and increasing the reliabilityof large programs.

As Howard Rheingold (1985) hasdescribed in Tools for Thought thegovernment was quick to seize on the interestof computer scientists at MIT in developingthe computer as an enhancement and extensionof human intellectual capabilities. In general,that interest coincided with the needs ofnational defense in the form of interactivecomputing, visual displays of both text andgraphics, mult i-user systems, andinter-computer networks. The AdvancedResearch Projects Agency (later DARPA),soon became a source of almost unlimitedfunding for research in these areas, a sourcethat bypassed the usual procedures of scientificfunding, in particular peer review. Much of theearly research in artificial intelligence derivedits funding from the same source, and itsdevelopment as a field of computer sciencesurely reflects that independence from theagenda of the discipline as a whole.

Although we commonly speak ofhardware and software in tandem, it is worthnoting that in a strict sense the notion of

software is an artifact of computing in thebusiness and government sectors during the'50s. Only when the computer left the researchlaboratory and the hands of the scientists andengineers did the writing of programs becomea question of production. It is in that light thatwe may most fruitfully view the developmentof programming languages, programmingsystems, operating systems, database and filemanagement systems, and communications andnetworks, all of them aimed at facilitating thework of programmers, maintaining managerialcontrol over them, and assuring the reliabilityof their programs. The Babel of programminglanguages in the '60s tends to distract attentionfrom the fact that three of the most commonlyused languages today are also among theoldest: FORTRAN for scientific computing,COBOL for data processing, and LISP forart ificial intelligence. ALGOL might haveremained a laboratory language had it and itsoffspring not become the vehicles of structuredprogramming, a movement addressed directlyto the problems of programming as a form ofproduction.17

Central to the history of software isthe sense of "crisis" that emerged in the late'60s as one large project after another ranbehind schedule, over budget, and belowspecifications. Though pervasive throughoutthe industry, it posed enough of a strategicthreat for the NATO Science Committee toconvene an international conference in 1968

17An effort at international cooper ation in establishinga standard programming language, ALGOL from itsinception in 1956 to its final (and, some argued,over-refined) form in 1968 provides a multileveledview of computing in the '60s. While contributingrichly to the conceptual development of programminglanguages, it also has a political history which carriesdown to the present in differing directions of research,both in computer science and, perhaps most clearly, insoftware engineering.


to address it. To emphasize the need fo r aconcerted effort along new lines, thecommittee coined the term "softwareengineering", reflecting the view that theproblem required the combination of scienceand management thought characteristic ofengineering. Efforts to define that combinationand to develop the corresponding methodsconstitute much of the history of computingduring the 1970s, at least in the realm of largesystems, and it is the essential background tothe story of Ada in the 1980s. It also revealsapparently fundamental differences betweenthe formal, mathematical orientation ofEuropean computer scientists and thepractical, industrial focus of their Americancounterparts. Historians of science andtechnology have seen those differences in thepast and have sought to explain them. Canhistorians of computing use those explanationsand in turn help to articulate them?

The effort to give meaning to"software engineering" as a discipline and todefine a place for it in the training of computerprofessionals should call the historian'sattention to the constellation of questionscontained under the heading of "disciplineformation and professionalization". In 1950computing consisted of a handful of speciallydesigned machines and a handful of speciallytrained programmers. By 1955 some 1000general-purpose computers required theservices of some 10,000 programmers. By1960, the number of devices had increasedfivefold, the number of programmers sixfold.And so the growth continued. With it cameassociations, societies, journals, magazines,and claims to professional and academicstanding. The development of theseinstitut ions is an essential part of the the socialhistory of computing as a technologicalenterprise. Again, one may ask to what extentthat development has followed historicalpatterns of institutionalization and to what

extent it has created its own. The question of sources illustrates

part icularly well how recent work in thehistory of technology may provide importantguidance to the history of computing, at thesame time that the latter adds new perspectivesto that work. As noted above, historians oftechnology have focused new attention on thenon-verbal expressions of engineering practice.Of the three main strands of computing, onlytheoretical computer science is essentiallyverbal in nature. Its sources come in the formmost familiar to historians of science, namelybooks, articles, and other less formal pieces ofwriting, which by and large encompass thethinking behind them. We know pretty wellhow to read them, even for what they do notsay explicitly. Similarly, at the level ofinstitutional and social history, we seem to beon familiar ground, suffering largely from anembarrassment of wealth unwinnowed by time.

But the computers themselves and theprograms that were written for them constitutea quite different range of sources and thuspose the challenge of determining how to readthem. As artifacts, computers present theproblem of all electrical and electronic devices.They are machines without moving parts. Evenwhen they are running, they display no internalaction to explain their outward behavior. Yet,Tracy Kidder's (1981) portrait of Tom Westsneaking a look at the boards of the new Vaxto see how DEC had gone about its workreminds us that the actual machines may holdtales untold by manuals, technical reports, andengineering drawings. Those sources toodemand our attention. When imaginativelyread, they promise to throw light not only onthe designers but also on those for whom theywere designing. Through the hardware and itsattendant sources one can follow the changingphysiognomy of computers as they made theirway from the laboratories and largeinstallations to the office and the home.


Today's prototypical computer iconically linkstelevision to typewriter. How that formemerged from a roomful of tubes and switchesis a matter of both technical and culturalhistory.

Though hard to interpret, thehardware is at least tangible. Software bycontrast is elusively intangible. In essence, it isthe behavior of the machines when running. Itis what converts their architecture to act ion,and it is constructed with action in mind; theprogrammer aims to make something happen.What, then, captures software for the historicalrecord? How do we document and preserve anhistorically significant compiler, operatingsysem, or database? Computer scientists havepointed to the limitations of the static programtext as a basis for determining the program'sdynamic behavior, and a provocative article(DeMillo et al. 1979) has questioned howmuch the writ ten record of programming cantell us about the behavior of programmers.Yet, Gerald M. Weinberg (1971, Chapter 1)has given an example of how programs maybe read to reveal the machines and peoplebehind them. In a sense, historians ofcomputing encounter from the oppositedirection the problem faced by the softwareindustry: what constitutes an adequate andreliable surrogate for an actually runningprogram? How, in particular, does thehistorian recapture, or the producer anticipate,the component that is always missing from thestatic record of software, namely the user forwhom it is written and whose behavior is anessential part of it?

Placing the history of computing in thecontext of the history of technology promisesa peculiarly recursive benefit. Althoughcomputation by machines has a long history,computing in the sense I have been using heredid not exist before the late 1940s. There wereno computers, no programmers, no computerscientists, no computer managers. Hence those

who invented and improved the computer,those who determined how to program it,those who defined its scientific foundations,those who established it as an industry in itselfand introduced it into business and industry allcame to computing from some otherbackground. With no inherent precedents fortheir work, they had to find their ownprecedents. Much of the history of computing,certainly for the first generat ion, but probablyalso for the second and third, derives from theprecedents these people drew from their pastexperience. In that sense, the history oftechnology shaped the history of computing,and the history of computing must turn to thehistory of technology for initial bearings.

A specific example may help toillustrate the point. Daniels (1970) stated asone of the really big questions thedevelopment of the 'American System' and itsculmination in mass production. It is perhapsthe central fact of technology in 19th-centuryAmerica, and every historian of the subjectmust grapple with it. So too, though Danielsdid not make the point, must historians of20th-century technology. For mass productionhas become an historical touchstone formodern engineers, in the area of software aswell as elsewhere.

For instance, in one of the majorinvited papers at the NATO SoftwareEngineering Conference of 1968, M.D.McIlroy of Bell Telephone Laboratorieslooked forward to the end of a "preindustrialera" in programming. His metaphors andsimiles harked back to the machine-toolindustry and its methods of production.

We undoubtedly producesoftware by backwardtechniques. We undoubtedlyget the short end of the stick inconfrontations with hardwarepeople because they are the


industrialists and we are thecrofters. Software productiontoday appears in the scale ofindustrialization somewherebelow the more backwardconstruction industries. I thinkits proper place is considerablyhigher, and would like toinvestigate the prospects formass-production techniques insoftware.(McIlroy, 1969)

What McIlroy had in mind was not replicationin large numbers, which is trivial for thecomputer, but rather programmed modulesthat might serve as standardized,interchangeable parts to be drawn from thelibrary shelf and inserted in larger productionprograms. A quotation from McIlroy's paperserved as leitmotiv to the first part of PeterWegner's series on "Capital Intensive SoftwareTechnology" in the July 1984 number of IEEESoftware, which was richly illustrated byphotographs of capital industry in the 1930sand included insets on the history oftechnology.18 By then McIlroy's equivalent tointerchangeable parts had become "reusablesoftware" and software engineers haddeveloped more sophisticated tools forproducing it. Whether they were (or now are)any closer to the goal is less important to thehistorian than the continuing strength of themodel. It reveals historical self-consciousness.

We should appreciate thatself-consciousness at the same time that weview it critically, resisting the temptation toaccept the comparisons as valid. An activity'schoice of historical models is itself part of the

history of the activity. McIlroy was notdescribing the state or even the direction ofsoftware in 1968. Rather, he was proposingan historical precedent on which to base itsfuture development. What is of interest to thehistorian of computing is why McIlroy chosethe model of mass production as thatprecedent. Precisely what model of massproduction did he have in mind, why did hethink it appropriate or applicable to software,why did he think his audience would respondwell to the proposal, and so on? The historyof technology provides a critical context forevaluating the answers, indeed for shaping thequestions. For historians, too, the evolvingtechniques of mass production in the 19thcentury constitute a model, or prototype, oftechnological development. Whether it is onemodel or a set of closely related models is amatter of current scholarly debate, but somefeatures seem clear. As a system it rested onfoundat ions established in the early andmid-19th century, among them in particularthe development of the machine-tool industry,which, as Nathan Rosenberg (1963) hasshown, itself followed a characteristic andrevealing pattern of innovation and diffusion ofnew techniques. Even with the requisiteprecision machinery, methods of massproduction did not transfer directly or easilyfrom one industry to another, and itsintroduction often took place in stages peculiarto production process involved (Hounshell1984). Software production may prove to bethe latest variation of the model, or criticalhistory of technology may show how it has notfit.

Conclusion: The Real ComputerRevolution

We can take this example a step farther. Fromvarious perspectives, people have been drawnto compare the computer to the automobile.

18One has to wonder about an article on softwareengineering that envisions progress on an industrialmodel and uses photographs taken from the GreatDepression.


Apple, Atari, and others have boasted ofcreat ing the Model T of microcomputers,clearly intending to convey the image of a carin every garage, an automobile that everyonecould drive, a machine that reshaped Americanlife. The software engineers who invoke theimage of mass production have it inseparablylinked in their minds to the automobile and itsinterchangeable variations on a standardtheme.

The two analogies serve different aimswithin the computer industry, the first lookingto the microcomputer as an object of massconsumption, the second to software systemsas objects of mass production. But they sharethe vision of a society radically altered by anew technology. Beneath the comparison liesthe conviction that the computer is bringingabout a revolution as profound as thattriggered by the automobile. The comparisonbetween the machines is fascinating in itself.Just how does one weigh the PC against thePT (personal transporter)?19 For that matter,which PC is the Model T: the Apple ][, theIBM, the Atari ST, the Macintosh? Yet thequestion is deeper than that. What would itmean for a microcomputer to play the role ofthe Model T in determining new social,economic, and political patterns? Thehistorical term in that comparison is not theModel T, but Middletown (Lynd and Lynd1929), where in less than forty years"high-speed steel and Ford cars" hadfundamentally changed the nature of work andthe lives of the workers. Where is theMiddletown of today, similarly transformed bythe presence of the microcomputer? Wherewould one look? How would one identify thechanges? What patterns of social andint e ll ec tua l behav io r mark suchtransformation? In short, how does one

compare technological societies? That is oneof the "big questions" for historians oftechnology, and it is only in the context of thehistory of technology that it will be answeredfor the computer.

From the very beginning, thecomputer has borne the label "revolutionary".Even as the first commercial machines werebeing delivered, commentators were extollingor fretting over the radical changes thewidespread use of computers would entail, andfew doubted their use would be widespread.The computer directed people's eyes towardthe future, and a few thousand bytes ofmemory seemed space enough for the solutionof almost any problem. On that bothenthusiasts and critics could agree. Computingmeant unprecedented power for science,industry, and business, and with the powercame difficulties and dangers that seemedequally unprecedented. By its nature as well asby its youth, the computer appeared to haveno history.

Yet, "revolution" is an essentiallyhistorical concept (Cohen 1986). Even whenturning things on their head, one can onlydefine what is new by what is old, andinnovation, however imaginative, can onlyproceed from what exists. The computer hada history out of which it emerged as a newdevice, and computing took shape from other,continuing activities, each with its ownhistorical momentum. As the world of thecomputer acquired its own form, it remainedembedded in the worlds of science,technology, industry, and business whichstructured computing even as they changed inresponse to it . In doing so they linked thehistory of computing to their own histories,which in turn reflected the presence of afundamentally new resource.

What is truly revolutionary about thecomputer will become clear only whencomputing acquires a proper history, one that19The latter designation stems from Frand (1983).


ties it to other technologies and thus uncoversthe precedents that make its innovationssignificant. Pursued within the larger enterpriseof the history of technology, the history ofcomputing will acquire the context of placeand time that gives history meaning.

Acknowledgements

This article is a slightly revised version of aposition paper prepared for the Seminar onInformation Technologies in HistoricalContext, held at the National Museum ofAmerican History, 11 September 1987. Itbenefitted at that time from the criticalcomments of David K. Allison, WilliamAspray, I. Bernard Cohen, and ArthurNorberg. The research from which it stems hasbeen generously supported by the Alfred P.Sloan, Jr. Foundation under its New LiberalArts Program.

References

Aspray, William. 1984. "Literature and Institutions inthe History of Computing." ISIS 75, pp. 162-170.

AT&T Bell Laboratories. 1987. UNIX SystemReadings and Applications. 2 vols. Englewood Cliffs,N.J., Prentice-Hall.

Backus, John. 1977. "Can Programming Be Liberatedfrom the von Neumann Style? A Functional Style andIts Algebra of Programs." (ACM Turing AwardLecture for 1977). Communications of the ACM, 21,8,pp. 613-641.

Bashe, Charles J., Lyle R. Johnson, John H. Palmer,and Emerson W. Pugh. 1986. IBM's Early Computers.Cambridge, MA, MIT Press.

Berkeley, Edmund C. 1949. Giant Brains or MachinesThat Think. New York, John Wiley & Sons.

Bolter, J. David. 1984. Turing's Man. Chapel Hill,University of North Carolina Press.

Boehm, Barry. 1981. Sof tware EngineeringEconomics. Englewood Cliffs, N.J., Prentice-Hall.

Burke, John G. 1970. "Comment: The complex natureof explanation in the historiography of technology."Technology and Culture, 11, pp. 22-26.

Buxton, J.N. and Brian Randell (eds.). 1970 SoftwareEngineering Techniques: Report on a conferencesponsored by the NATO Science Committee, Rome,Italy, 27th to 31st October 1969. Brussels, ScientificAffairs Department, NATO.Cf. Naur et al. (1976).

Ceruzzi, Paul E. 1982 Reckoners: The Prehistory ofthe Digital Computer, From Relays to the StoredProgram Concept, 1935-1945. Westport, CT,Greenwood Press.

Cohen, I. Bernard. 1986. Revolutions in Science.Cambridge, MA, Harvard University Press.

Cowan, Ruth. 1983. More Work for Mother: TheIronies of Household Technology from the OpenHearth to the Microwave. New York, Basic Books.

Daniels, George. 1970. "The Big Questions in theHistory of American Technology." Technology andCulture, 11, pp. 1-21.

DeMillo, Richard, Richar d J. Lipton, and Alan J.Perlis. 1979. "Social Processes and Proofs of Theoremsand Programs." Communications of the ACM, 22, 5,pp. 271-280.

Diebold, John. November-December 1953."Automation - the new technology. Harvard BusinessReview. pp. 63-71.

Feigenbaum, Edward and Pamela McCorduck. 1983The Fifth Generation: Artificial Intelligence andJapan's Computer Challenge to the World. Reading,MA, Addison-Wesley.

Ferguson, Eugene S. 1979a. "The American-ness ofAmerican Technology." Technology and Culture, 20,.pp. 3-24.

Ferguson, Eugene S. 1979b. "The Mind's Eye:Nonverbal Thought in Technology." Science, 197, pp.827-836.


Ford, Henry. 1922. My Life and Times. Garden City,N.Y., Doubleday.

Frand, Erwin. July 1983. "Thoughts on ProductDevelopment: Remembrance of Things Past."Industrial Research and Development. p. 23.

Greenbaum, Joan. 1979. In the Name of Efficiency:Management Theory and Shopfloor Practice in DataProcessing Work. Philadelphia, Temple UniversityPress.

Herrmann, Cyril C. and John F. Magee. July-August1953. "'Operat ions Research ' for Management."Harvard Business Rev iew. pp. 100-112.

Hindle, Brooke. 1981. Emulation and Invention. NewYork, Basic Books.

Hounshell, David A. 1984. From American System toMass Production: The Development of ManufacturingTechnology in the United States. Baltimore, JohnsHopkins University Press.

Hughes, Thomas P. 1983. Networks of Power:Electrification in Western Society, 1880-1930.Baltimore, Johns Hopkins University Press.

Hurni, Melvin L. September-October 1955."Decision Making in the Age of Automation."Harvard Business Review 34, pp. 49-58.

Jenkins, Reese V. 1987. "Words, Images, Artifacts andSound: Documents for the History ofTechnology." British Journal for the History ofScience 20, pp. 39-56.

Kidder, Tracy. 1981. The Soul of a New Machine. NewYork, Little, Brown & Co.

Kraft, Philip. 1977. Programmers and Managers: TheRoutinization of Computer Programming in the UnitedStates. New York, Springer Verlag.

Lundstrom, David E. 1987. A Few Good Men FromUnivac. Cambridge, MA, MIT Press.

Lynd, Robert S. and Helen Merrell Lynd. 1929.Middletown. New York, Harcourt, Brace & World.

Machlup, Fritz and Una Mansfeld. 1983. The Study ofInformation: Interdisciplinary Messages. New York,

John Wiley & Sons.

McDougall, Walter A. 1985. ... The Heavens and theEarth: A Political History of the Space Age.New York, Basic Books.

McIlroy, M. Douglas. 1969. "Mass-ProducedSoftware." In Naur and Randell 1969.

Meyer, Stephen III. 1981. The Five-Dollar Day.Albany, State University of New York Press.

Morison, Elting E. 1974. From Know-How toNowhere: The Development of American Technology.New York, Basic Books.

Naur, Peter and Brian Randell (eds.). 1969. SoftwareEngineering: Report on a conference sponsored bythe NATO Science Committee, Garmisch, Germany,7th to 11th October, 1968. Brussels, Scientific AffairsDivision, NATO. See Naur et al. 1976.

Naur, Peter, Brian Randell and J.N. Buxton (eds.).1976. Software Engineering: Concepts andTechniques. New York, Petrocelli/Charter.

Nelson, Daniel. 1975. Managers and Workers:Origins of the New Factory System in the UnitedStates, 1880-1920. Madison, University of WisconsinPress.

Newell, Allen and Simon, Herbert A. 1976."Computer Science as Empirical Inquiry."Communications of the ACM 19, pp.113-126.

Noble, David. 1984. Forces of Production: A SocialHistory of Industrial Automation. New York, AlfredKnopf.

Nora, Simon and Alain Minc. 1978. L'Informatisationde la société. Paris, La documentation française.Translated into English under the t itle TheComputerization of Society. Cambridge, MA, MITPress, 1980.

Randell, Brian (ed.). 1982. Origins of DigitalComputers: Selected Papers. Berlin/Heidelberg/NewYork, Springer Verlag. 3rd ed.

Rheingold, Howard. 1985. Tools for Thought: ThePeople and Ideas Behind the Next ComputerRevolution. New York, Simon and Schuster.


Rose, Frank. 1984. Into the Heart of the Mind: AnAmerican Quest for Artificial Intelligence. New York,Random House.

Rosen, Saul. 1967. "Programming Systems andLanguages - A Historical Survey." In his ProgrammingSystems and Languages. New York, McGraw-Hill.

Rosenberg, Nathan. 1963. "Technological Change inthe Machine Tool Industry, 1840-1910." Journal ofEconomic History 23, pp. 414-443.

Rosenberg, Nathan. 1979. "Technologicalinterdependence in the American Economy."Technology and Culture 20, pp. 25-50.

Sammet, Jean. 1969. Programming Languages.History and Fundamentals. Englewood Cliffs, N.J.,Prentice-Hall.

Simon, Herbert A. 1960. The New Science ofManagement Decision. New York, Harper & Row.

Simon, Herbert. 1981. The Sciences of the Artificial.Cambridge, Mass., MIT Press. 2nd ed.

Smith, Merritt Roe. 1977. Harper's Ferry Armory andthe New Technology. Ithaca, Cornell University Press.

Stern, Nancy. 1981. From ENIAC to UNIVAC: AnAppraisal of the Eckert-Mauchly Computers. Billerica,MA, Digital Press.

Wallace, Anthony F.P. 1978. Rockdale: The Growth ofan American Village in the Early IndustrialRevolution. New York, Alfred A. Knopf. (Paperbacked. New York, W.W. Norton, 1980).

Weinberg, Gerald M. 1971. The Psychology ofComputer Programming. New York, Van NostrandReinhold.

Wexelblatt, Richard L. 1981. History of ProgrammingLanguages. New York, Academic Press.

Wilkinson, James H. 1971. "Some Comments from aNumerical Analyst." (1970 Turing Award Lecture).Journal of the ACM 18,2(1971), 137-147;

Williams, Michael. 1986. A History of ComputingTechnology. Englewood Cliffs, N.J., Prentice-Hall.

Yourdon, Edward. 1979. Classics of SoftwareEngineering. New York, Yourdon Press.

Yourdon, Edward. 1982. Papers of the Revolution.New York, Yourdon Press.