Adam Thornton

March 26, 1997

SHARE is the users’ group founded in August 1955 by the prospective users of the IBM 704, at that time the largest of IBM’s computers. SHARE grew out of informal talks between the RAND Corporation, Lockheed Aircraft Corporation, and North American Aviation, Inc., all of whom were preparing to migrate from the IBM 701 to the 704 as it became available. Eighteen sites were represented at the first meeting of SHARE, and it grew rapidly thereafter: by the ninth meeting of SHARE in October 1957, its membership included Bell Labs, Boeing, Dow Chemicals, Esso, General Motors, General Electric, Lockheed, Los Alamos, MIT, NSA, Rand, SAC, Standard Oil, and Union Carbide. Clearly, all three sectors of the "Military-Industrial-Academic Complex" were represented; indeed, the SHARE membership list reads much like a Who’s Who of Cold War Big Science and Engineering.

It is not immediately clear why SHARE should have come to exist at all; its stated function was "to eliminate, as much as possible, redundant effort expended in using the 704," by maintaining a library of software written by and available to SHARE members (Paul Armer, from the Rand Corporation, was one of SHARE’s founding members). However, the companies sending representatives to SHARE were often in precisely the same line of business, and thus certainly bidding on many of the same contracts—for example, Boeing, Lockheed, Grumman Aircraft, and Hughes Aircraft. It is not obvious that these companies should feel they had more to gain by sharing software techniques than to lose by giving their solutions to their competition. However, Armer’s 1956 address to the Electronic Business Systems Conference suggests the magnitude of the savings to an institution participating in SHARE: "by [the introduction of the IBM 701] the amount of redundant effort was horrendous; the cost of developing a system for using the machine, and a set of routines to go with that system, was usually in excess of a year’s rental for the equipment." The huge savings promised by using already-developed and tested systems through SHARE may have thus overshadowed the potential loss incurred by letting a competitor share software. However, there is a second motive at work as well. Although IBM could have afforded to have lost any one or two of even its largest accounts, it could not lose a dozen. Thus, SHARE could act as a bargaining collective, giving the companies a position of power from which to make demands of IBM. As Armer put it in 1956, "SHARE provides collective ‘feedback’ from the customers to the manufacturer," and IBM—at that point—felt obligated to act on this feedback.

I have not yet acquired the earliest SHARE proceedings; however, the first three meetings’ proceedings were reprinted in the 1980s and will be relatively easy to obtain (unlike the proceedings of meetings IV-VIII, which have so far remained elusive). One assumption I make is that SHARE’s initial membership list will, like the membership by the ninth meeting in 1957, prove to be composed of the largest corporations and research laboratories in postwar America. I assume this to be a reasonably safe bet, since it is precisely those largest and richest companies that had the capital to invest in an IBM 704, the largest commercially available scientific computer at that time.

The relationship between SHARE and IBM changed radically between 1958 and 1964: the SHARE Operating System was built with the understanding that SHARE would deliver the specifications to IBM, and IBM would produce the software. In 1964, IBM announced System/360, whose operating system was to be OS/360, which would be specified, written, and maintained by IBM alone. By then, there seems to have been no pretense that the members of SHARE, if dissatisfied with the product, would—or, indeed, could—take their business elsewhere: that is, by 1964, the companies involved in SHARE were dependent on IBM and could not function without IBM computers.

In this respect at least, it is clear that a fundamental change in the nature of these companies had taken place between 1958 and 1964. An assumption that I am initially making is that dependence on the computer—whether for business or scientific computing, or both (SHARE has traditionally been concerned with the use of computers in research, while GUIDE, founded in 1956, has focussed on the application of computers to commercial data processing; another way to view this is that SHARE was initially made up of 701 and 704 customers and GUIDE of 702 and 705 users)—initiated changes that rippled throughout the company and will be visible in such areas as creation of permanent programming and computer operations staff positions, as well as the creation of divisions whose responsibility is primarily the computer. These companies can hardly fail to have an effect—though this will be difficult to show—on the entire American economic landscape. Somewhere between 1955 and 1965, there exists a watershed, after which the computer became a necessity for a company engaged in certain kinds of research. I hope to locate this watershed more precisely and to draw a picture of the shift making a computer system an integral part of the company.

I believe that much of the answer is to be found in the model of industry proposed in Galbraith’s The New Industrial State. Even those SHARE members that nominally were part of the private sector found themselves relying increasingly on governmental money, whether as direct research support or in the form of contracts. Large amounts of money were earmarked precisely for the solution of calculation-intensive engineering problems that required extensive computing. For example, it is surely no accident that many of the early SHARE members were aerospace engineering companies. To maintain and increase their funding, these companies had a need for ever more computing power; thus, they (as SHARE members) continually spent more money with IBM, and found themselves increasingly dependent on IBM. According to Galbraith it is precisely the high-tech industries that demonstrate the breakdown of the traditional model of laissez-faire capitalism. This feedback loop, which serves to cement IBM’s hold over its customers through indirect influx of governmental funds, is, I think, an excellent example of the fundamental change in the nature of business occurring in postwar America.

Presently I have no real evidence for this hypothesis. To keep my research relatively focussed I intend to narrow the timeframe considerably. Assuming that I am able to acquire the first three SHARE proceedings and some of the early SHARE Secretary’s Distributions (the usual form of missives sent to the SHARE community as a whole), the first few years of SHARE seem to be the most profitable, although it may be difficult to prove the dependence of SHARE members on IBM before the System/360 era. I will in addition select just a few SHARE members to trace through my period. These will be private industrial companies: whereas NSA, SAC, and Rand will not be considered, General Electric, General Motors, and Lockheed are three attractive possibilities.. However, the final choice of companies to be examined will be largely determined according to the survival, suitability, maintenance, and accessibility of their archives.

It is tempting to expand the scope farther; having looked at the influence of computerization on these companies, it is more than inviting to look at the influence both of computerization on the popular culture and media of the late 1950s and early 1960s, and of these companies on that same culture and media. However, I am presently ill-prepared for this expanded project; at the very least it should be put on hold until I am fairly confident in my ability to turn out a credible and well-grounded account of the corporations in question. While it is certainly plausible that the incredible growth of parts of the economy between 1950 and 1965 played itself out in the creation of a consumer culture dependent on widespread affluence in precisely those classes likely to be working in newly-computerized companies—which themselves grew at such a rate as to require recruitment into these classes—this hypothesis would be (I suspect) well-nigh impossible to prove.

Several other questions become immediately apparent. The most glaring: how are companies convinced that they need a computer in the first place? After all, a computer was an expensive and untried new technology. I suspect that this is a more interesting question for those locales doing primarily data processing rather than scientific calculation: it is probably easier to justify the necessity for an electronic computer to calculate airflow over a wing than to justify replacing a set of mechanical accounting machines with an expensive electronic machine that performs the same functions.

There are, however, other questions which must be addressed. At what point did companies realize that programmers were an integral part of their workforce? To what degree was IBM’s realization that software as big business per se a factor in its decision to control system software for its hardware? What large, early IBM computer customers (3M being one example) decided not to enter SHARE but to continue proprietary development in-house? Did this strategy work for them, or did they eventually join SHARE, and when? Did SHARE members usually use only IBM computers, or were they likely to have a Univac (or some other computer) as well? Did they initially attempt to develop their own computer systems? Is the answer different in 1957 and 1962?

In short, I intend to use SHARE as a window through which to observe the change in US technological industry between roughly 1955 and 1963. This change is one that, I believe, has fundamental (though certainly not exclusive) roots in computerization and in the nature of increasing corporate dependence on the state, supported by the pressures and needs of the defense industry during the Cold War.

Every business history of the computer that I have seen has focused on the computer as an item of production rather than an item of consumption. The most useful book I have found is Kenneth J. Flamm’s work Creating The Computer: Government, Industry, and High Technology. However, Flamm’s book essentially tells the story of direct federal funding of the major computer companies and its decline during the 1960s; it is my contention that although federal funding of IBM and its competitors dropped off during the 1960s, the money with which they were being paid by their customers was overwhelmingly governmental in origin

Creating The Computer, however, is not especially concerned with the uses of the computer within the private sector. Although Flamm recognizes that the first major commercial users of the IBM 701 computer were, in fact, the California aerospace companies, he does not draw out the implication that, while direct federal funding of IBM research and development declined, the money being poured into IBM by its customers was supplied to those customers by the government either in the form of research grants or contracts. Flamm does point to an interesting feature of the California high-tech business culture: the extreme fluidity of its workforce: "[t]he key element in these efforts was the fluid movement of engineers from one aerospace patron to another and into start-ups that often survived only long enough to be acquired by a larger firm." This becomes particularly interesting in light of Ross Bassett’s analysis of the development of the Silicon Valley CMOS transistor industry, which is said to have been driven by precisely the same type of heightened worker motility. In agreement with these claims, I suspect that I will find rapid turnover of a skilled engineering workforce to be characteristic of the businesses I include in my study. Flamm’s companion volume, Targeting the Computer: Government Support and International Competition makes the case that governmental funding of computing technology is justified by the high rate of return in terms of the declining costs of data processing. The two books are meant to be considered as a set, with the first primarily historical, and the second primarily economic.

James W. Cortada has analyzed the deployment of the computer as a commercial product in The Computer in the United States: From Laboratory to Market, 1930 to 1960. However, his analysis rests on his consideration of the computer as an office tool; that is, the computer as used in what have traditionally been called "business applications" rather than in "scientific computing." Although a similar story could certainly be told of the data processing industry with a focus on GUIDE, my choice of limiting the scope of my research to SHARE will effectively constrains my analysis to "scientific computing. Princeton has strong historical ties to SHARE, which it lacks with respect to GUIDE. In some sense, too, the story of corporate dependence on computers to do data processing is less challenging, as IBM and Remington Rand explicitly marketed their business computers as the next logical step beyond the familiar punched-card electromechanical accounting machines, which they had been supplying for decades and for which businesses had already demonstrated a need; therefore, inertia is a plausible reason for the computerization of the data processing industry.

No one has yet analyzed how the introduction of the computer as a scientific/engineering tool affected those companies that invested in computing technology. While studies have shown much fundamental computer research to have been governmentally funded (and only later spun off into commercial products), the converse question regarding computer consumption has not yet been examined: to what extent was the consumption of commercial computers governmentally funded? The initial answer appears to be that consumption was overwhelmingly funded through governmental sources; this answer, however, still awaits its supporting evidence.

A seminal text describing the change in the American corporate landscape after WWII is John Kenneth Galbraith’s The New Industrial State. The theme of the high-tech corporation's increasing dependence on public funding is, of course, central to my project. From this theme, it follows that precisely those engineering activities that required extensive computation were the ones most likely to attract governmental funding during the 1950s and 1960s. A second theme contributing to my project lies in Alfred Chandler’s Scale and Scope: large corporations were reorganized along structural-functional lines, and this new structure of the corporation becomes particularly interesting in the early days of a company’s computerization. Although the changes Chandler details in his work are, in general, finished before my story begins, Scale and Scope is still immensely helpful in generating a picture of a typical structure for a large, technology-intensive industrial company immediately before its computerization. How that structure is changed by the addition of a computer is the focus of some interesting questions: if a company was recently restructured into an amalgam of semi-autonomous units, then was the computer, when acquired, immediately granted its own division of the company? Would the computer remain under the control of the unit that initially requested it? Who would manage the computer when more than one division wanted to use it (one can, for instance, imagine a computer whose services were in demand from both the propulsion and the airframe group at an aerospace company)? These questions will probably require archival research to be satisfactorily answered.

One anomalous feature is that, although the period of my study is precisely the time the large corporation shifted from being an entrepreneurial company, led by a charismatic visionary—for instance, the Ford Motor Company in the days of the Model T—to a diversified, impersonal behemoth run by a Board of nameless and faceless Directors, IBM itself very strongly bore the stamp both of Thomas J. Watson, Sr. and his son, Thomas Jr. (often credited with driving IBM’s commitment to computer research). Such a corporate profile would not, perhaps, be unusual in a computer company that was, in essence, an entrepreneurial startup (for instance, the Eckert-Mauchly computer venture, or, more recently, Microsoft); it is very strange for a company as huge and diversified as IBM. I do not yet know whether this observation has any relevance to my project, but it is glaring and difficult to ignore.

The story of IBM’s entry into the computing market is authoritatively told by Bashe, Johnson, Palmer and Pugh in IBM’s Early Computers. Emerson Pugh’s Memories That Shaped an Industry: Decisions Leading to IBM System/360 is also a useful view of the technical considerations informing the development of IBM’s early computer systems. Both of these, of course, focus exclusively on the computer as the object of production, and only address its use insofar as answering who bought what, when.

David Noble’s The Forces Of Production offers a leftist view of the introduction of numerical control methods into the machining industry. His book is primarily a labor history, describing the de-skilling and disempowering of the shop floor worker as the machines he used became more and more automated. This, then, is the story of the assembly line rewritten for the 1950s and 1960s; it provides a view of a use of computers easily overlooked. Noble’s book convincingly describes how computers were useful in the aerospace industry as tools for both production and design. The major effect—and primary intention—of the introduction of numerical control methods to the workplace was to transfer authority and control from the machine operator to the engineer who wrote the program to control the machine. However, the computing problems being talked about at SHARE were not issues of machine control of tools. My guess—as yet wholly unverified—is that the computers controlling tools were, at least in the 1950s and 1960s, much smaller machines than those discussed by SHARE members, and thus are of ancillary importance, at best, to my project.

There are still a number of published works I need to read, even before I begin to use archival sources. Cortada claims that Fisher, McKie, and Mancke’s IBM and The U.S. Data Processing Industry: An Economic History paints a detailed picture of the commercial computing landscape. I need to revisit The Annals of the History of Computing and compile an index of relevant articles, beginning with the reprint of Armer’s 1956 speech about SHARE and notes on the occasion of SHARE’s 25^th anniversary. John Varick Wells’s Yale Ph.D. dissertation, "The Origins of the Computer Industry: A Case Study in Radical Technological Change," will be included in my research. Pugh’s Building IBM offers a more business-centered view of IBM’s entry into the computer industry than his other, technology-centered books.

The place to start my archival research is in the collection of the Charles Babbage Institute in Minneapolis, Minnesota. CBI possesses 15.5 cubic feet of documents covering the early years of SHARE. These ought to provide evidence about the concerns of SHARE members in its formative years. Also, I expect to learn which companies were the most active in SHARE and, therefore, are worth being investigated further as possible candidates for my case studies. CBI’s Oral History collection will also provide useful information. I have already requested transcripts of the two interviews conducted with Paul Armer in their collection. These two transcripts will help me determine which of the other interviews in the CBI collection may contain information pertinent to my project. Those interviews, in turn, will point me towards companies in SHARE that ought to be considered in detailed case studies.