Grace Hopper and the Invention of the Information Age

by Kurt W Beyer

  But COBOL did flourish on an unprecedented scale, despite the warnings of the experts. In the 10 years after its introduction, COBOL became the most widely used programming language throughout the world. By the turn of the millennium, it was estimated, 240 billion of the 300 billion lines of computer code (about 80 percent of all code worldwide) was written in COBOL. Because 95 percent of finance and insurance data was processed in COBOL, the fear of computer glitches on a global scale mounted as the year 2000 approached. “Y2K” anxieties were based on the fact that the CODASYL committee, for reasons probably connected to memory conservation, adopted a two-digit convention to represent that year. Hence, as year fields worldwide switched from “99” to “00,” the potential threat of an array of arithmetic, comparison, and sorting errors emerged. The efforts of thousands of COBOL programmers, may of whom came out of retirement in the years leading up to 2000, fixed the vast majority of the Y2K bugs, and the millennium began with little cyber-fanfare.51

  As with all technologies that proliferate, COBOL’s long-term success was far from inevitable and rested on a combination of interrelated factors. The most obvious were specific actions by Charles Phillips, chairman of the CODASYL committee. In effect, Phillips used the economic weight of the Department of Defense to put pressure on uncooperative companies. On 12 August 1960, Phillips wrote a personal letter to computer manufacturers asking for a commitment to support, develop, and implement the common business language created by the CODASYL committee. This request was given financial gravity when Phillips required all future DOD computer procurements to be bundled with a working COBOL compiler.

  In view of the purchasing power of the Department of Defense in the early 1960s, even the largest manufacturers were forced to fulfill the DOD’s requirement.52 This did not mean, however, that IBM put other initiatives on hold. Shortly after Phillips’s DOD policy announcement, a senior IBM executive wrote to inform Phillips that IBM intended to complete COBOL processors for its line of computers by the third quarter of 1961. He also mentioned that IBM’s customers were still interested in using COMTRAN, and that IBM had distributed Commercial Translator manuals 3 months earlier.

  Phillips’s actions may have created a temporary demand for COBOL compilers, but, as IBM’s actions proved, DOD policy in no way guaranteed the sustained use and subsequent spread of the contentious language. For Betty Holberton and its other creators, COBOL’s successful proliferation could be attributed in part to the design of the technology itself . For all of the slights about COBOL’s capabilities, Holberton was a true believer in the language. She had attended the first CODASYL meeting, had been a member of the short-term committee, and had edited the final COBOL specifications. Her fondness for the language was personal as well as professional. Having dedicated a year of her life to the creation of the common business language, she was convinced that the committee had done its best to find a technological solution to the future needs of users and manufacturers.

  Holberton’s 15 years of programming experience, matched only by Hopper’s longevity in the field, contributed to her belief that COBOL’s positive design aspects outweighed the technical limitations identified by a variety of other computer experts. She defended the English-based pseudo-code (criticized by many as verbose and inefficient) on the ground that the language had been developed for future users, not for present programming experts. The next generation of computer users needed a language that could be learned quickly. The self-documenting nature of the English pseudo-code permitted a programmer to easily understand the work of another and build upon it, which made COBOL ideal for collaborative programming efforts in which multiple people worked on the same data-processing problem. The flexible file nomenclature did not require the user to change the definitions on preexisting data files, only the style of writing the program. This was important when converting older programs to the new COBOL format. Holberton’s experience in managing data-processing installations, combined with her deep knowledge of programming, helped her to appreciate these more subtle COBOL design characteristics. During the 1950s an elite of mathematics-oriented programmers dominated the field, but with the exponential spread of computers during the 1960s others gained access to computers. The new users, less interested in the “art” of programming, just wanted to solve data-processing problems. They were the first generation of programmers who knew little about the workings of computer hardware and who therefore appreciated a computer language that served as an intermediary between them and the machine.

  GRACE HOPPER AND THE SPREAD OF COBOL

  No matter how well designed a new technology appears to be, it cannot spread spontaneously. A new technology spreads by means of the continuous actions of hundreds and thousands of users, manufacturers, and intermediaries. But not everyone associated with a new product or idea is equal in his or her ability to influence the expanding technological system. Some individuals are connected in special ways throughout the technological system. They have the ability to influence both micro and macro outcomes, and their actions have ripple effects. Despite Betty Holberton’s accolades for the technology itself , she and others continually acknowledged Grace Hopper as the person most responsible for the success of COBOL during the 1960s.

  How can a single person influence widespread technical change? Modern society appears too large and complex for individual actions to carry much weight. This is especially true in the realm of technology, where new products and practices seem to appear and proliferate of their own accord. Technological progress is something that simply happens. It is the inevitability of society focused on the future.

  Hopper understood that technology did not exist apart from society. A technology’s logical design would not guarantee its ultimate acceptance. A multitude of cultural, social, political, and economic factors figured in the acceptance or rejection of a given technology, be it a car, a camera, or a computer language. “Because of early difficulties . . . prejudices, inertia, normal reluctance to change procedures, some remaining technological imperfections,” Hopper wrote in a detailed report to the Navy explaining why the institution’s leadership needed to make a concerted effort to push COBOL as the service’s business language standard, “it will be necessary to supply motivation for the use of COBOL.” COBOL would not sell itself; the concerted efforts of a variety of people were needed. This brings up an interesting point concerning the process of invention. Invention does not stop with the creation of a prototype. Prototypes must be actively marketed by their creators, or by proxies for their creators, in order to gain acceptance. The marketing process, ideally, feeds back into the technology’s design, embedding in the physical artifact social, economic, and political attributes that further its chance of survival. In effect, Hopper’s support of COBOL did not begin in 1960; it began in 1951 with the invention of the A-0 compiler. Since then Grace Hopper had been automatic programming’s most energetic proponent, and COBOL was the next iteration of that grander vision.53

  What had changed by 1960 was Hopper’s status within the maturing computer industry. The prominent pioneer had levers that few within the computing community possessed during the 1960s. Her position as director of automated programming at the number-two computer manufacturer conferred on her influence over senior management within Sperry Rand and over other senior computer executives. She was also one of the most senior women in the Naval Reserve, serving as a technical advisor to the largest consumer of computers, the Department of Defense.54

  Within the world of computers, Hopper was a connector, a role that few people were capable of playing during the 1960s. Malcolm Gladwell, a science writer for the Washington Post and The New Yorker, eloquently describes the importance of “connectors” in technological revolutions in his 2002 book The Tipping Point. Gladwell employs the biological theory of epidemics as his chief metaphor to relate how new technologies propagate rapidly. Epidemics are a function of the particular attributes of the infectious agent, of the behavior of the carriers of the
infectious agents, and of the environment in which the infectious agent is introduced. The three elements interact to determine how quickly the epidemic spreads.

  For instance, the AIDS virus spreads via exchange of fluids. Activities that facilitate such exchange, such as unprotected sex and the sharing of needles, are the chief vehicles for the virus. At the beginning of the epidemic, it was not the actions of all those infected but the activities of a handful of “connectors” who engaged repeatedly in high-risk behaviors that influenced the early progression of the virus.

  To shine a more positive light on the epidemic metaphor, Gladwell substitutes for the spread of disease the flow of information, and tells the story of the most honored “connector” of the American Revolution: Paul Revere. During the night of 18 April 1775, America’s famous equestrian was not the only person riding into the Massachusetts night to warn that the British were coming. William Dawes, Samuel Prescott, and more than 60 others also got on their horses to spread the word. Why then do we remember Revere’s name but not the names of the many others who rode? According to Gladwell, it is because, like Grace Hopper, Paul Revere was a connector. His prominence in local Bostonian politics and in various revolutionary organizations allowed him to construct the groundwork for spreading the message in the months and years leading up to the fateful night.55

  The epidemic metaphor is relevant to the dissemination of COBOL as well as to Grace Hopper’s role as a connector. Instead of a virus with unique biological characteristics, there existed in 1960 a prototype of a common business language with certain design attributes. In the post-Sputnik, Cold War-era United States, the overall demand for computers was on the rise, as was the need for programmers. Into this environment stepped Hopper as a connector. With the ability to operate in a variety of different subcultures, connecting users with programmers and industry executives with senior military officers, she translated her computing vision into a language that each of these groups understood, communicating complex technology to technically savvy people and nontechnical people alike. Her special gifts and unique position allowed her to drive the COBOL “epidemic” forward.

  COBOL STANDARDIZATION AND THE NAVY

  On 1 January 1967, Commander Grace Hopper was placed on the Naval Reserve’s retirement list. Two years earlier she had stepped down as Director of Automatic Programming in the UNIVAC Division of Sperry Rand. Though she remained with UNIVAC as a senior staff scientist, from all appearances the 60-year-old Hopper was slowly making the transition to retirement.

  Hopper’s 1966 report to the Navy recommending the full implementation of COBOL began with a two-page discussion of a general theory of language. She explained that a primary characteristic of intelligent life is the ability to communicate: “Man’s brain enables him to learn an elaborate system of rules both for communication with others and for control of his own mental processes. Man, having created rules, added a generative potential—rules can produce consistent but flexible (i.e., contingent) systems—making languages possible.” The elements of code that make up a language, following a system of rules, combine in almost unlimited combinations to creating meaning. But the key, according to Hopper, is to have orators and listeners agree on both codes and rules. “Surely,” she writes, “the story of Babel is the lesson of the destruction of standards.”56

  For Hopper, computer languages were no different than any other human language. COBOL, in particular, was an artificial dialect, made up of English words and some parts of English syntax, that followed agreed-upon rules. The artificial language permitted humans to communicate an exact description of data and the algorithms or procedures to be followed in the solution of business problems by computer. What made it unique was that the language could also communicate by means of a compiler with a variety of computer hardware and translate exact descriptions into machine instructions.57 Linguistic history shows that over time languages tend to standardize in order to increase the efficiency of communication between people. Hopper put it this way: “Thus it seems that COBOL is but one of many efforts in science and technology to standardize terminology and syntax in order to provide better communications between men concerning a particular class of activities.”58

  COBOL AND TECHNOLOGICAL CLOSURE

  COBOL is distinctive within the history of technology, for it represents a technological artifact consciously and deliberately designed and selected by a number of social groups with competing interests. Although manufacturers, users, administrators, and engineers typically participate in what social constructivists refer to as the “closure” of artifact design,59 technological stabilization is usually a much more random process, with standards organically emerging from the struggles between heterogeneous organizations and individuals.

  The story of COBOL has some elements of randomness, but the intention of Hopper, Phillips, and the other CODASYL members was to create a forum to rationalize the design process and generate consensus apart from the marketplace. Not only was a standardized technology blueprint produced and approved by January 1960; the intermediate-term and long-term committees had grander visions. In effect, the intermediate-term committee was charged with inventing the next generation following the “state-of-the-art” design of the short-term committee. Even more ambitious, the envisioned long-term committee was supposed to then design the “ultimate” state of the art that would unify all programming languages into one.

  In the end, the short-term committee’s design, which was meant to be temporary, became the de facto COBOL standard. Though individual manufacturers made changes to the language as the 1960s progressed, COBOL did not degenerate into a large number of dialects, as occurred with FORTRAN.60 The intermediate-term committee had little influence over future COBOL developments, while the long-term committee was never even established. By the late 1970s, COBOL was the most extensively used computer language, and more than 80 percent of all business applications were written in the FLOW-MATIC-based language.

  12 INVENTING THE INFORMATION AGE

  As the 1950s came to a close, Sperry Rand’s director of programming research Grace Hopper conducted an “experiment” to demonstrate how far the field of software development had advanced in the 15 years since she wrote her first code. Hopper would attempt to turn a “trim, attractive blonde” into a computer programmer. Hopper’s subject—Marilyn Mealey, a 19-year-old high school graduate from the Mayfair section of Philadelphia—was “prettier than average” and liked to swim and listen to records. According to Hopper, Marilyn was much like other young women who “window shop during their lunch hours and look forward to the evening’s dates and dancing.”1

  The experiment’s results were published in Popular Electronics, a magazine with an overwhelmingly male audience of gadget lovers who must have been amused at the thought of a young woman barely out of high school programming a computer. Anyone with even a limited knowledge of computers knew that communicating with multi-million-dollar mechanical brains was a task best left to highly educated professionals. Only the sharpest mathematical and scientific minds could write computer code, and there was no place in the growing industry for pretty blondes from the Mayfair section of Philadelphia.

  Grace Hopper believed otherwise. In fact, Marilyn Mealey embodied Hopper’s 15-year crusade to democratize the computer industry. Hopper invented the fundamental technologies that permitted humans to efficiently communicate with complex calculating machines. Subroutine libraries, pseudo-code (source code), decision branching, debugging routines, and compilers were the building blocks for all future high-end computer languages. But it was Hopper’s relentless promotion of a “computer age” that gave her technological creations contextual meaning. If computers were merely a tool for the scientific elite, as her former boss Howard Aiken believed, then there was no need for automatic programming. Hopper envisioned a future wherein computers would be indispensable to a diverse set of people and organizations. In such a world, both elite mathematicians and high school graduates would
need to converse with the machines.

  Hopper’s example runs counter to the well-known adage that necessity is the mother of invention. Hopper “invented” not because of a glaring need in her immediate present, but because of a “potential” need in a “possible” future. To complicate matters, the probability of success of that possible future was dependent on not only the technologies Hopper invented but also on her relentless promotion of a particular future that incorporated those technologies. An inventor not only invents a technology but also formulates a future need for that technology. In this light, Hopper’s skills in writing and in oratory were as important as her technical and mathematical acumen.

  Marilyn Mealey also represented Hopper’s continued struggle to keep the new technical field gender neutral. Computers were not only for men, nor were they just for exceptional women such as Hopper. By purposefully picking a young, attractive blonde, Hopper aimed to shatter stereotypes. “New opportunities for women in electronics have been created by the wide scale application of computers to business and scientific work,” she wrote in the Popular Electronics article.2 By connecting the growing electronics industry with new opportunities for women, Hopper aimed to expand a legacy that she had worked so hard to create.