Computing and the National Science Foundation, 1950-2016. William Aspray
Section with Tom Keenan, John Lehmann, and later Val Tareski as program managers. The concurrent growth in academic computer science programs and researchers led OCA’s computing research portfolio to grow. A discipline of computer science was emerging but was not yet sufficiently well-defined to provide an obvious blueprint for the new Computer Science and Engineering (CS&E) Section. With leadership from Rose and input from the advisory committee, Curtis and his program team began to define a set of programs. As Keenan noted:
Well, computer science had achieved the title computer science without much science in it, early. And I think we—here I have to say that Kent Curtis was a prime person . . . I loved the man very much; he was a great guy—we decided that to be a science you had to have theory, and not just theory itself as a separate program, but everything had to have a theoretical basis. And so, whenever we had a proposal, we encouraged, as much as we could, some kind of a theoretical background for this proposal—not just software, and not just write a program, but there should be some basis for it.98
The CS&E staff worked together to define a set of programs:
. . . we decided that there was a minimum of three—smallest integer greater than two—things that went to make up computing. The first was theory; the second was hardware; the third was software. So, John Lehmann became the hardware person. I became the software person. And in the beginning, I think Val Tareski was the theory person . . . each of these programs had probably something less than a million dollars to spend. I think the section had perhaps a $2 million budget in 1969 or 1970.99
The CS&E portfolio of grants, taken together with support from engineering and information science programs, represented a growing investment in the emerging computing research field. Some of the early OCA research grants were awarded to Niklaus Wirth (Stanford), Michael Harrison (Berkeley), Sam Conte (Purdue), Patrick Fischer (Cornell), Juris Hartmanis (Cornell), and Martin Davis (NYU Courant). Computer science-related facilities awards went to Edward Feigenbaum (Stanford), John Pasta (UIUC), Conte, and Richard Conway (Cornell). The Engineering Section in Mathematics, Physical Sciences, and Engineering (MPE) funded Walter Karplus (UCLA), Melvin Breuer (USC), Edward Coffman (Princeton), and Steve Ungar (Columbia), while the Division of Information Sciences funded Vladimir Slamecka (Georgia Tech) and Naomi Sager (NYU).
When Pasta joined NSF in January 1970, he became extremely important in navigating NSF “politics.” Pasta was respected by the senior NSF staff and other division directors in MPE due to his intellect and background in mathematics, physics, engineering, and computer science. This was essential to the growth of computer science funding in competition with other disciplines for budget. His death in 1981 eventually led to the Computer Science Section (CSS) being split off from the Math Section in 1984,100 as a separate division in Mathematics and Physical Sciences (MPS; by this time, Engineering had become a separate directorate). There was a feeling among many that no one, other than Pasta, had the breadth of background to oversee both mathematics and computer science. His Signal Corps background and his long connection with the Atomic Energy Commission (AEC) and classified projects made it possible for him to play a key role in the conversation between NSF and NSA over cryptography research.
When NSF terminated the computer facilities program, Pasta reorganized OCA into three sections: Computer Science and Engineering, Computer Applications in Research, and Computer Innovations in Education.101 These three new sections reflected the changing nature of computer science and of OCA’s role within NSF.102 The Computer Science and Engineering Section continued to sponsor research in fundamental computer science, the Computer Innovations in Education Section helped bring the power of the computer to bear on the problems of education, and the Computer Applications in Research Section fostered the development of advanced computer techniques to increase science research capability.
In 1972, Pasta recruited Peter G. Lykos, an Illinois Institute of Technology computational chemist, to NSF with the explicit charge to lead a new initiative to address computer impacts on society.103 Lykos was assigned to Aufenkamp’s section until he could get the program started. During his tenure at NSF, Lykos experienced turbulent times. OCA had ended its computing facilities program and the computers in education programs were transferred to the Education Directorate. Lykos recalled104 frustration working with the OCA staff and for the loss of the facilities and later the computers in education programs. He left NSF around the time the Office of Computing Activities was reorganized and transferred to the Research Directorate in November 1973.105
In 1974, OCA was restructured as the Division of Computer Research (DCR) with Pasta as Division Director. The division106 supported research in all areas of computing with a major emphasis on fundamental aspects of computer science and engineering (in Curtis’s section), on research directed toward the development of techniques that increase the responsiveness of the computer to the requirements of scientific disciplines (in Aufenkamp’s section), and on privacy and computer system security, human-machine interface, and societal impacts of computing (in a newly formed section led by Fredrick Weingarten based on Lykos’s initiative).
1.5Summary and Conclusions
The first 24 years of NSF were marked by changing roles and outcomes for its computing and information programs. The Cold War had a strong influence on the science information and computing facilities programs. Interest in foreign intelligence increased the science information budgets. Defense and atomic energy agencies created a rapid growth in the number, capabilities, and providers of computers and computing facilities. Scientists who had limited or no access to Department of Defense (DoD) and AEC laboratories increased their demand on NSF to provide campus facilities. When NSF was given responsibility for applications in science information and computing facilities, the need to provide the underlying technology resulted in NSF investments that advanced fundamental and applied research. Program and office managers in mathematics, engineering, and the OCA began to make grants to the early pioneers in computing research that with DoD support helped establish early computer science programs. OSIS initiated a number of academic information science and systems programs.
By the mid-1970s, OSIS had been greatly weakened and was moved to a nonsupportive directorate. OCA lost its facilities and education programs and had yet to gain the respect of the NSF management. In the 1980s everything would change dramatically.
Much credit for protecting the NSF computing and information programs and building grant portfolios that advanced the underlying technologies is due to a few individuals. Burt Adkinson, the long-term head of OSI and OSIS (1957–1970), was a champion for science information and information science across the government and the discipline. Helen Brownson (1951–1966) was responsible for guiding many of the research efforts funded by OSI and OSIS. Milton Rose (1963–1969), Mathematics Division Director and first head of OCA, recruited to government service a veritable who’s who of computing and was a significant force in the rapid development of computing and computer science in academia. Milt was replaced by John Pasta (1969–1981), who led OCA, DCR, and DMCS through many changes and who with Kent Curtis (1967–1987) established the programs that led to the current strong position of NSF in computer science research.
Notes
1.National Research Council. 1999. Funding a Revolution: Government Support for Computing Research. Washington, DC: National Academies Press. DOI: 10.17226/6323.