A Superb Manager IndeedJuly 6, 2005
In a conversation with Philip Davis, Kathleen Broome Williams, author of a new biography of Grace Hopper ("The Author and Her Subject," SIAM News, May 2005, page 6; http://www.siam.org/news/news.php?id=57), clarified Hopper's role in the creation of COBOL. Neither Williams nor Davis ever met Hopper. SIAM founder I.E. Block, who did know Hopper, thought that readers might enjoy a fuller version of the story.
One afternoon in early March 1959, I received a call from Mary Hawes, manager of software development at Electrodata, in Pasadena. At that time, I was manager of Burroughs' Philadelphia Computer Center. Burroughs had acquired Electrodata several years earlier.
Mary wanted me to organize a volunteer group that would initiate the development of a common language for business data processing. With such a language, a single program could be written for a business problem that could be compiled to run for each of the computers of the various manufacturers. The group Mary envisioned would represent users of computers for business purposes, along with computer manufacturers. The users would need to accept the language as a standard for programming business problems. The manufacturers would need to develop compilers for their respective computers that could accommodate programs in the common language.
According to Mary, a common language was urgently needed. But she cautioned that the project was a "political hot potato" in that it would require the support of both users and manufacturers of computers.
Mary said that she would be handicapped in organizing such an effort from her office in California, which was not central to business software development. She considered Philadel-phia an ideal place to start the effort because of the amount of software activity in the Delaware Valley--at Burroughs, Philco, RCA, Remington Rand Univac, the University of Pennsylvania--and its close proximity to IBM in New York and to the federal agencies in Washington.
We decided that we needed a small representative group to formulate objectives and to identify a larger group that would develop specifications for the language and promulgate the results to government and industry. I agreed to take on the project.
I first called my friend Saul Gorn, a professor of computer and information sciences at the Moore School of Electrical Engineering at the University of Pennsylvania. After some discussion, he suggested holding the meeting in his office, which would be politically neutral territory. I then called Grace Hopper (SIAM vice president for planning in 1953–54), who was director of software development for Remington Rand Univac. Grace and her group had developed FLOWMATIC, Univac's English-language business compiler. She had already received a letter about the proposed project from Mary, and she agreed to meet with us. Among the others I invited were representatives of IBM, Philco, and RCA, and Gene Smith, director of the computing branch of the Navy's Bureau of Ships.
We met in Saul's office in late March. Grace sent her assistant, Robert Rossheim. I recall that we had many ideas about what the proposed working group should do and managed to identify some candidates for it. We came to no conclusions but agreed to meet again about two weeks later. Grace came to the later meeting. We agreed on some activities for the working group and expanded our list of candidates. But we were troubled about how to handle the politics of the many diverse, competitive groups.
As I recall, Grace contributed little until, near the end of the meeting, she said that the only way to proceed was to convince Charlie Philips, director of the data systems research staff at the Office of the Assistant Secretary of Defense, to convene a meeting of the users and manufacturers. As a politically neutral leader, he would implicitly bring pressure on the manufacturers to cooperate. Grace suggested that Gene Smith plant the idea in Philips's mind and then excused herself, leaving the rest of us impressed by her acuity in handling such a sensitive situation.
Gene agreed to call Philips. Philips grabbed the idea. By mid-May, Philips had sent letters inviting key people among the users, including Defense Department representatives, and manufacturers to a meeting that would be held May 28–29, 1959, to "consider development of specifications for a common business language for automatic digital computers." More than 40 attended. That meeting marked the beginning of the development of COBOL.
In fingering Philips for the job, Grace illustrated her leadership and prowess in getting management support for a difficult project that did not have an evident payoff. Grace was a good manager, one who was especially adept in getting her way on controversial projects.--Ed Block