Grace Murray Hopper
Hopper, Grace Murray
HOPPER, GRACE MURRAY
computer sciences, programming languages, COBOL.
An admiral who never went to sea, Hopper owed her success in the U.S. Navy, as in civilian life, to her mastery of computers and computing. Entering the field at its very beginnings in the 1940s, she spent the next four decades leading the way in the development of computer languages. In 1969, the Data Processing Management Association named her its first Computer Sciences Man of the Year for her contribution to the development of the widely used programming language, COBOL. Her implementation of Standard COBOL in the Navy revolutionized its management information systems. Often referred to as “Amazing Grace” and “Grandma COBOL” by an admiring press, Hopper gave hundreds of speeches annually in the 1970s and 1980s, becoming a nationally recognized advocate for navy computing and for the computer sciences she had helped to establish.
Origins and Education . The oldest of three children, Hopper was raised in New York City in a family that encouraged intellectual curiosity. An early interest in mathematics was stimulated by both her mother, Mary van Horne Murray, the daughter of a civil engineer, and by her father, Walter Murray, an insurance broker. After attending private girls’ schools in New York and New Jersey, Hopper was admitted to Vassar College in September 1924. Vassar, just north of New York City, was a prestigious private college offering a rigorous education to the bright daughters of the affluent. Hopper disregarded the post–World War I trend in women’s education toward subjects considered suitable preparation for marriage and motherhood. Instead, she concentrated on mathematics and physics. Henry Sealy White, her most influential mentor, had been trained in Göttingen, Germany, the center of European mathematics. Europe was still well ahead of the United States in mathematics in the 1920s, and many eminent American as well as European mathematicians and physicists trained at Göttingen, among them J. Robert Oppenheimer, Richard Courant, Max Born, Enrico Fermi, John von Neumann, Edward Teller, and Werner Heisenberg.
In 1928, Hopper graduated from Vassar with a Bachelor of Arts degree in mathematics and physics and was elected to Phi Beta Kappa, a preeminent national honor society. She was awarded a Vassar College Fellowship to pursue graduate study at Yale University, not far away in New Haven, Connecticut. Two years later, Hopper received a Master of Arts degree in mathematics from Yale. In June 1930, she married Vincent Foster Hopper, a graduate of Princeton University. Vincent taught English at New York University and soon enrolled in a PhD program in English. Hopper, meanwhile, continued her graduate studies at Yale, one of only two women in the mathematics doctoral program. In 1931, in addition to her own studies, Hopper began teaching mathematics at Vassar. She was a naturally gifted teacher, and for much of the next fifty years she continued to teach, at least a course or two, whenever she could.
In 1934, Hopper was awarded a PhD in mathematics from Yale, the only woman to receive one of seven doctorates in mathematics granted by the university between 1934 and 1937. She had studied under James Pierpont and with algebraist Oystein Ore and wrote her dissertation on “New Types of Irreducibility Criteria.” Hopper spent the next nine years teaching at Vassar, advancing from assistant to associate professor. In 1941, she used a Vassar faculty fellowship to study under Richard Courant at New York University. Courant came from the same Göttingen tradition as her mentors at Vassar and Yale, and Hopper worked with him on the calculus of variations and on differential geometry. Her work with him on partial differential equations showed up later on her navy job classification card and was, she believed, the reason for her wartime assignment to the computing program at Harvard University.
World War II and the Navy . World War II changed Hopper’s life forever. After Pearl Harbor she was determined to join the navy because, she explained, one of her great grandfathers had been an admiral during the Civil War. Furthermore, she and Vincent had just separated (they were divorced in 1945, and Hopper never remarried), so she was free to serve. In December 1943, at the age of thirty-seven, Hopper was sworn in to the U.S. Navy Reserve. In June 1944, she was commissioned a lieutenant (junior grade) and was sent to the Bureau of Ships Computation Project at Harvard University in Cambridge, Massachusetts.
Between the two world wars the growth of large-scale businesses required dealing with modern calculation and record-keeping needs. Work progressed in several industrial and academic laboratories on improved calculators. One or two early computer designs were also produced. By the early 1930s, Vannevar Bush and his colleagues at MIT had created a differential analyzer, an electromechanical analog computer, copies of which were widely used during World War II. In 1940, George R. Stibitz, of Bell Telephone Laboratories, created an electromechanical digital computer; at Iowa State College, John V. Atanasoff worked on an electronic digital device that some consider the first electronic digital computer in the United States, though it never worked. In Germany, Konrad Zuse developed various generations of computers named for him, and in England, Alan Turing developed a relay-based machine, called a bombe, to decrypt German enigma codes.
In 1937, Howard Aiken began work at Harvard University on an electromechanical relay machine to perform long mathematical calculations automatically. Engineers at IBM built Aiken’s device, which he named the Mark I when it was set up at Harvard in the spring of 1944. Desperate for gunnery and ballistics calculations, the navy’s Bureau of Ships leased the Mark I from the university for the duration of the war. Aiken, by then a professor of physics and applied mathematics and a naval reserve officer, was assigned to command the Harvard Computation Laboratory for the navy. To run the operation he assembled and trained a small team of mathematicians who, like himself, were serving in the wartime navy. Among them was Grace Hopper.
The Mark I measured 51 feet long, 8 feet tall, and 8 feet deep. It had more than 750,000 parts, used 530 miles of wire, and weighed more than 5 tons. A four-horsepower electric motor and drive shaft drove all the mechanical parts by a system of interlocking gears, counters, switches, and control circuits. Input was entered by punching holes in paper tape that was then fed into the machine. Output was handled automatically by electric typewriters.
What emerged was the first functional, large-scale, automatically sequenced, general-purpose digital computer in the United States. It was electromechanical, destined to dominate the field for only a year or two, briefly bridging the gap between calculators and electronic computers.
The Mark I was reliable, accurate, and speedy—at least one hundred times faster than hand-operated calculators. It was also adaptable and could be set to perform a wide range of different types of numerical calculations. This multipurpose capability of the Mark I set it apart from other contemporary devices, particularly the electronic ENIAC, with which it has been unfavorably compared but which was not fully functional until after the end of the war. Hopper always maintained that the most important thing she learned from Aiken was the notion of a general purpose computer—one that could undertake any task that could be described by a series of logical instructions.
Hopper’s task at the Computation Lab was to write the codes that directed the operation of the Mark I. Because the computer was the first of its kind, she had no experience in coding, and she and her colleagues— arguably the first programmers—had to develop ways to express mathematical problems in digital machine code. Each mathematical process had to be broken down into very small steps of addition, subtraction, multiplication, or division, and put into a sequence. After figuring out how to write the machine instructions, Hopper punched them on tape, put the tape in the computer, and hoped it would run. The navy had new rockets but no firing tables for them. They had new magnetic mines but no tables of effective ranges for optimal mine laying. All these things could be computed on the Mark I, but someone had to translate the problems into terms the machine could understand. This was the challenge for Hopper and her colleagues; in learning how to talk to the Mark I, they learned a process that sixty years later was still the essence of programming.
The first problem that Hopper worked on was how to compute rocket trajectories, which involved finding the interpolation coefficients for applications of the arc tangent series. Another problem she worked on concerned the effective range of magnetic mines that were set off by metal ships. Working by trial and error at first, and learning as she went along, Hopper wrote the code for these and many other problems. The Mark I operated twenty-four hours per day, seven days per week, providing information of great practical value to the war effort.
Hopper also worked on creating codes for the Mark II, Aiken’s second, and still electro-mechanical, computer, commissioned by the navy. In addition to coding, she wrote a manual of operation for the Mark I. Although she had no engineering background, she taught herself to understand the Mark I’s circuits in order to explain its coding procedures and plugging instructions. She compiled 561 pages of detailed descriptions and diagrams of all the computer’s parts and how they operated, as well as samples of all kinds of coding.
While at Harvard, Hopper developed the seeds of what, after the war, became her most creative work. Looking for shortcuts to writing programs, Hopper assembled a collection of subroutines—small bits of code for often-repeated functions. These she could copy into new programs, speeding the whole process. Noting that program-stopping errors often crept into her copying, Hopper recognized that if she could make the computer do the copying, she could avoid human copying errors. This insight eventually led to her creation of compilers.
Postwar Work at Harvard . At the end of the war, the Computation Lab was returned to Harvard, with Aiken, a civilian again, retaining control. Vassar College offered Hopper a full professorship, but she preferred to continue working in the innovative new field of computing and never taught full-time again. Instead, she remained at the Computation Lab as a research fellow in engineering systems and applied physics under a three-year, navy-funded contract. As Aiken’s deputy she had charge of the operation of the Mark I, which continued to churn out answers to mathematical problems in fields as varied as atomic physics, radio research, optics, electronics, and astronomy. She also worked on Aiken’s first electronic computer, the Mark III, commissioned by the navy and equipped with an innovative magnetic storage drum.
Under Aiken’s direction, Hopper organized a series of international conferences to promote open discussion of computers and computing. The first of these, the Symposium of Large Scale Digital Calculating Machinery, was held at Harvard in 1947 and attracted many of the big names in computing, including Norbert Wiener, George Stibitz, Jay Forrester, Richard Courant, Wassily Leontief, John Mauchly, Herman Goldstine, and John von Neumann. These conferences introduced Hopper to the growing computer community, as did a number of articles she coauthored with Aiken on their wartime work at the Computation Lab.
After the war, Hopper maintained her naval reserve status and was promoted to lieutenant in 1946. With the passage of the Women’s Armed Services Integration Act in 1948, she hoped to transfer to the regular navy, but she was over the age limit and had to settle for remaining in the reserves. With the expiration of her three-year contract at Harvard in 1949, she moved out of academe and into industry.
In Industry . Commercial development of electronic digital computers was in its infancy when Hopper left Harvard. In the late 1940s, even many involved in the design and construction of computers thought that their use was limited to providing mathematical solutions for scientific projects and that a very few would be sufficient to satisfy the world’s need. Looking ahead more clearly, Hopper joined the Eckert-Mauchly Computer Corporation (EMCC) in Philadelphia. She chose EMCC because she believed its Universal Automatic Computer (Univac) would be operational before those in development at other fledgling companies, including IBM. Established by J. Presper Eckert and John Mauchly—the creators of the ENIAC—EMCC’s high-speed Univac proved very successful. Hopper remained with the Univac division through Eckert-Mauchly’s acquisition by Remington
Rand in 1950 and then through the merger creating Sperry Rand (later Unisys) in 1955.
Hired as a senior mathematician, Hopper continued her focus on methods to simplify the laborious and labor-intensive task of writing coding instructions. She was one of the original visionaries to recognize that computers could serve as the prime vehicle for their own programming and was a key player in conceiving, developing, and implementing the concept of compilers. In 1951, she began work on her first compiler, and by 1953, she had produced a commercially successful version, the A-2. Many mathematical problems, even with different objectives and results, used some of the same sets of instructions, or subroutines. The subroutines might solve certain classes of equations, or extract roots, and they were identical no matter what program they were inserted into. Yet for each job the programmer had to write pages and pages of duplicate code. Hopper’s solution was to design a tape with a three-letter call sign for each set of subroutines. The programmer only had to enter the appropriate call sign for the computer to retrieve the subroutine from its library of instruction tapes. What had previously taken a month of programming time could now be accomplished in five minutes.
By the mid-1950s, as Hopper had predicted, computers were being used by a wide range of businesses. When they began to come off the assembly line, programming— because it was complex and time-consuming—became a bottleneck. Hopper had placed herself in the vanguard of those addressing this problem. She understood that, because it was impossible to teach everyone how to write computer code, there would have to be an interface that would accept what Hopper called “people-oriented” data. This the computer itself would translate into machine code. Pursuing this reasoning, Hopper led the development of a compiled programming language, Flow-Matic, which by the end of 1956 could translate instructions written in a limited vocabulary of English sentences into a computer program ready to run on the Univac.
Contrary to the impression given by some accounts of her accomplishments, Hopper was not the only one working on computer languages. That same year IBM came out with FORTRAN, for use on scientific and mathematical problems, and then developed Comtran, while Honeywell created Fact. And everyone drew on the brilliant early work of British mathematician Alan Turing. Nevertheless, Flow-Matic was a big breakthrough. Apart from a slight difference in terminology, business compilers from that point followed the pattern that it had established, including the use of the imperative form of verbs.
With the proliferation of computers and computing languages came a move among manufacturers to introduce some sort of standardization and portability of languages. Hopper was a key player in this from the very beginning. She was instrumental in the creation of the Committee on Data Systems Languages (CODASYL), tasked with putting together a standard business language. What emerged in 1960 was the first version of COBOL, Common Business Oriented Language, which was frequently updated but not replaced. Although Hopper was only indirectly involved in the work of the committee, the bulk of COBOL, including its format, was based on Hopper’s Flow-Matic. When Hopper ran COBOL on a Univac and then on an RCA computer, demonstrating that one data-processing language could be run on different machines produced by different manufacturers, COBOL quickly became the most widely used programming language for mainframe computers.
In 1961, Hopper was named director of research in systems and programming for the Remington Rand division of Sperry Rand, responsible for broad-scale systems and programming research for all Remington Rand divisions. In 1964, she was appointed senior staff scientist at the Univac Division and traveled widely representing Univac interests in Europe and Japan as well as the United States and Canada. She worked on the basic techniques of creating, recording, retrieving, processing, and transmitting information and developed new concepts in systems design and information processing for such computer fields as medical science, language translation, and simulation processes.
During her eighteen years with the Univac Division, Hopper remained active in the naval reserve. Assigned to an ordnance unit in the Philadelphia Navy Yard, she played a part in the extension of computing to different navy tasks, making a name for herself in the navy as she was doing in the civilian world. Finally, in 1967, having reached the mandatory navy retirement age of sixty, Hopper was directed to retire. A few months later the navy recalled her to active duty for six months. She stayed for nineteen years.
Return to the Navy . Taking a military leave of absence from the Univac division, Hopper was assigned to the Pentagon to standardize computer-programming languages for all navy computers not part of weapons systems. She had been handpicked for the job by Norman Ream, special assistant to the Secretary of the Navy, who was tasked with imposing order on the navy’s automatic data processing. By 1967, different and incompatible versions of COBOL had been proliferating almost unchecked to accommodate machine-dependent or vendor-unique features as well as the different tasks undertaken at different navy installations. Without a uniform language, there was no portability; programs had to be revised and rewritten every time they were transferred from one machine to another and across generations. Increasing demand for computerized systems in response to the navy’s expansion in the Vietnam War made the restructuring of computing essential. Ream turned to Hopper to head an effort mandating the use of Standard COBOL in the navy and developing procedures for validating compliance.
Until 1970, Hopper headed a small group in the Office of Information Systems Planning and Development. Three-quarters of the navy’s work was ashore, and COBOL was widely employed for inventory, payroll, personnel, and other management information systems. There were millions of lines of COBOL code to deal with, as well as the usual resistance to change and to the imposition of uniformity that bedevil any bureaucracy. Hopper’s task was broad, far-reaching, and difficult, and it required her considerable diplomatic as well as technical skills. Her success in developing Standard Navy COBOL, persuading users throughout the navy to adopt it, creating test routines to check COBOL compilers for compatibility and standards, and providing technical support for its use made her a legend in the navy and beyond. Eventually, Hopper’s Standard Navy COBOL was adopted by the entire Department of Defense; it influenced an entire industry as well, because manufactures could not sell their computers to the military unless they were compatible with Hopper’s COBOL.
In 1970, Hopper was honored by the American Federation of Information Processing Societies with their prestigious Harry Goode Memorial Award for her leadership in the development of computer software and her influence on the computing profession. Three years later the navy acknowledged her accomplishments, promoting her to the rank of captain. Also in 1973, Hopper became the only woman, and the only American, to be named a Distinguished Fellow of the British Computer Society. She spent her last nine years in the navy as special staff to the commanding admiral of the Naval Data Automation Command. She not only advised him on the state of computing in the navy but became a roving ambassador for navy computing, giving over two hundred speeches annually, across the country and abroad. In 1983, her work was recognized by her promotion to the rank of commodore and, two years later, at the age of seventy-nine, she became a rear admiral.
Life’s Work . In 1986, after forty-three years in the navy (and its oldest serving officer), Hopper reluctantly retired. She worked as a consultant for Digital for several years until her health failed. She died in 1992 at the age of eighty-five and was buried with full military honors in Arlington National Cemetery. Tributes poured in from around the world, and the navy named an Aegis guided missile destroyer for her. The USS Hopper(DDG 70) was commissioned in San Francisco in 1997. Fittingly, its motto is “Dare and Do.”
In addition to her many other awards—including the 1991 National Medal of Technology—Hopper was the recipient of thirty-four honorary degrees. She remained a teacher at heart, influencing generations of students both at work and through the computer courses she taught for many years at various universities. She also published more than fifty papers and articles on computer software and programming languages. Her genius lay in making accessible this new and highly specialized field. In her own person, she represented the entire history of programming languages, from Flow-Matic in the 1950s to COBOL for minicomputers in the 1980s. Yet whenever she gave public addresses, she invariably asked to be introduced simply as one of the programmers on the first large-scale digital computer, the Mark I.
For a complete list of publications by Grace Hopper, see Charlene W. Billings, Grace Hopper: Navy Admiral and Computer Pioneer (Hillside, NJ: Enslow Publishers, 1989). Most of Grace Hopper’s surviving papers are to be found in the Grace Murray Hopper Collection at the Archives Center, National Museum of American History, Smithsonian Institution.
WORKS BY HOPPER
A Manual of Operation for the Automatic Sequence Controlled Calculator. Cambridge, MA: Harvard University Press, 1946.
With Howard H. Aiken. “The Automatic Sequence Controlled Calculator, I–III.” Electrical Engineering 65 (August–September 1946): 384–391, 449–454, 522–528.
“The Education of a Computer.” Symposium of Industrial Applications of Automatic Computing Equipment (January 1953): 139–44.
“Compiling Routines.” Computers and Automation (May 1953): 1–5.
With John Mauchly. “Influence of Programming Techniques on the Design of Computers.” Proceedings of the IRE 41 (October 1953): 1250–1254.
“Standardization and the Future of Computers.” Data Management(April 1970): 32–35.
With Steve Mandell. Understanding Computers. St. Paul. MN: West Publishing, 1984.
Beyer, Kurt. “Grace Hopper and the Early History of Computer Programming: 1944–1960.” PhD diss., University of California, Berkeley, 2002.
Billings, Charlene W. Grace Hopper: Navy Admiral and Computer Pioneer. Hillside, NJ: Enslow Publishers, 1989.
Cohen, I. Bernard. New Foreword to A Manual of Operation for the Automatic Sequence Controlled Calculator. Cambridge: MIT Press, 1985.
———. Howard Aiken: Portrait of a Computer Pioneer. Cambridge, MA: MIT Press, 1999.
———, and Gregory W. Welch, eds. Makin’ Numbers: Howard Aiken and the Computer. Cambridge, MA: MIT Press, 1999.
“Harvard Computation Laboratory.” Journal of Applied Physics 17, no. 10 (October 1946): 856.
Light, Jennifer S. “When Computers Were Women.” Technology and Culture 40, no. 3 (July 1999): 455–483.
Mason, John F. “Grand Lady of Software.” Electronic Design 22 (25 October 1976): 86.
Mitchell, Carmen Lois. “The Contribution of Grace Murray Hopper to Computer Science and Computer Education.” PhD diss., University of North Texas: University Microfilms, 1994.
Prokop, Jan, ed. Computers in the Navy. Annapolis, MD: Naval Institute Press, 1976.
Randell, Brian. The Origin of Digital Computers: Selected Papers. New York: Springer-Verlag, 1973.
Rodgers, William. Think: A Biography of the Watsons and IBM. New York: Stein and Day, 1969.
Spencer, Donald P. Great Men and Women of Computing. Ormond Beach, FL: Camelot Publishing Company, 1996.
Stern, Nancy. From Eniac to Univac: An Appraisal of the EckertMauchly Computers. Bedford, MA: Digital Press, 1981.
Whitelaw, Nancy. Grace Hopper: Programming Pioneer. New York: W. H. Freeman, 1995.
Williams, Kathleen Broome. Grace Hopper: Admiral of the Cyber Sea. Annapolis, MD: Naval Institute Press, 2004.
———. Improbable Warriors: Women Scientists and the U.S. Navy in World War II. Annapolis, MD: Naval Institute Press, 2001.
Yost, Edna. Women of Modern Science. New York: Dodd, Mead, 1959.
Kathleen Broome Williams