Computing and the National Science Foundation, 1950-2016. William Aspray

the U.S. Office of Education, and the Carnegie Corporation of New York supported Suppes’s research projects.

In 1963 at Dartmouth, John Kemeny and Thomas Kurtz transformed the role of computers in education from primarily a research activity to an academic one. They did not like the idea that students had to stand in long lines with punch cards for batch processing. So they adopted the recently demonstrated concept of timesharing, which enabled many students to interact directly with the computer. The university developed its own time-shared system and expanded it into a regional computing center for colleges and schools. Kemeny, a mathematician who later became Dartmouth’s president, had applied for an NSF grant to bring a GE-225 computer to campus and to build the first fully functional general-purpose timesharing system.⁸¹ He received the funding despite reviewers’ serious doubts about his plan to employ undergraduates as his research team. Together, Kemeny, Kurtz, and their undergraduate students built a time-sharing system at Dartmouth. At the same time, they developed a new programming language, BASIC (Beginner’s All-purpose Symbolic Instruction Code). It turned out to be ideal for introducing beginners to programming and nevertheless was powerful enough to be used for most applications. BASIC worked on any computer. It spread rapidly and was used for the creation of computer-based instructional materials for a wide variety of subjects at all levels of education.

In the early seventies, Seymour Papert at MIT set out to develop a new and different approach to computers in education. He developed a programming language, Logo, to encourage rigorous thinking about mathematics. He wanted it to be accessible to children and be easy to use to express procedures for simple, non-numerical tasks familiar to children. He used it for mathematics education by teaching it in a wide variety of interesting “micro world” environments such as music and physics. Papert insisted that one should not teach mathematics but instead should teach children to be mathematicians. Logo soon became the language of the elementary school computer literacy movement. After OCA was created, the Logo group wanted to do more testing in schools in collaboration with Wally Feurzeig at Bolt, Baranek and Newman (BBN).⁸² The joint project did receive NSF funding, but only following extensive arguments and considerable reservations. NSF was concerned with giving research funding to a private company such as BBN. At the time, NSF preferred a non-profit, research-oriented institute or university such as MIT. “BBN was a suspect as being a money-grabbing kind of place rather than pure as a drift of snow like universities. So, he [the head of OCA, Dr. Milton Rose] said: ‘Why should I fund you? You are not a university.’ ”⁸³ However, Feurzeig’s group at BBN was the only group then doing this type of research, and so the NSF obliged. Because of differing viewpoints between the Logo Group’s goal to revolutionize mathematics teaching and NSF’s focus on educational applications, NSF cut the project’s funding in 1977. “These cuts succeeded in allowing the NSF to better control Logo’s development as an educational tool rather than a revolution.”⁸⁴

In October 1972, OCA’s Computer Innovations in Education Section⁸⁵ was transferred to the Education Directorate where funds for research and education started to tighten. To bolster support for their programs, the group decided to support two demonstration projects: PLATO IV⁸⁶ and the Time-shared Interactive Computer Controlled Information Television system (TICCIT),⁸⁷ directed by John Volk of the MITRE Corporation. While PLATO was a large centralized system, TICCIT used a minicomputer and two-way television in a more distributed system.⁸⁸ The National Science Board, at first skeptical, was impressed with the demonstrations and the result slowed budget reductions temporarily.

President Ronald Reagan’s fiscal year 1982 budget for NSF included major reductions for education and social science funding. As a result, all funding in the Education Directorate, except for graduate fellowships, was slashed.⁸⁹ Molnar was left to close out all of the existing grants. However, he was able to find ways to fund computers in education researchers. He and Dorothy Deringer, an information scientist from Case Western Reserve serving as an NSF program officer, recruited vendors to donate equipment to NSF and this equipment was made available to researchers. The Education Directorate was eventually restored, and Molnar remained there. Attempts failed to move the computers in education programs into the Computer and Information Science and Engineering Directorate when it was created in 1986. Molnar continued to interact with CISE staff and was involved in the MOSIS VLSI fabrication facility and worked with DARPA and CISE staff members John Lehmann and Bernard Chern to provide access to that system.⁹⁰

1.4Finding a Home for Computer Science Research

By the late 1950s, the Mathematical Sciences Section was making computer research grants, for example to Delaware, Harvard, Kansas, Michigan State, Michigan, Princeton, Syracuse, and Yale as well as for computing facilities at Northwestern.⁹¹ Grants were later awarded to Oregon State University, Columbia, Delaware, and Rice.⁹²

Under the leadership of Donald Laird, program director for Computer Sciences, and Milton Rose, program director and, later, section head for Mathematics, the NSF program in the mathematical sciences began in the early 1960s to include grants for theoretical symbolic logic, computer sciences, artificial intelligence, and pattern recognition.⁹³ In 1965, 10% of the NSF fellowships in mathematics went to computer scientists; by 1974, the percentage grew to 20%.⁹⁴ The computing facilities and research activities and program managers were transferred from the Mathematical Sciences Section to the Office of Computing Activities when it was created in 1967, with Rose as its head.⁹⁵

The NSF leadership’s view lingered that computer science was primarily a form of scientific infrastructure, rather than a discipline in its own right, but OCA fulfilled the hopes of ACM activists by bringing computer science out from under the shadow of mathematics, where its status as a research field had always been in question.⁹⁶

      The shift also kept computer science out of the Engineering Division, which had been lobbying since 1965 for control over computing activities. The placement of the Office of Computing Activities under the NSF Director, and its emphasis on education rather than engineering, was a disappointment to NSF’s engineers.

The Office of Computing Activities’ initial budget was $22 million, a 73% increase from the $12.7 million allocated for computer education and research in mathematics and other NSF offices in the previous year. OCA had three sections: the Institutional Computing Services Section (for funding universities to purchase computers as a tool for scientists), led by Kent Curtis; the Special Projects/ Computer Innovations in Education Section, led by Arthur Melmed; and the Computer Science Education, Research, and Training Section, led by Fredrick Weingarten. The initial OCA Advisory Committee included a number of leading figures in the developing discipline.⁹⁷

      The primary initial role of OCA was to support computing facilities, computers in education, and training of computing professionals. In 1968, Donald Aufenkamp assumed management of the facilities programs and Curtis moved over to head