Jack Minker of the University of Maryland, who was co-chair of the 1980–1981 computer science advisory subcommittee, asked John Guttag of MIT to head an ad hoc committee to review current NSF policy regarding cryptographic research. At the May 1981 Advisory Subcommittee,84 a three-and-one-half-hour discussion was held on the “Role of the NSF in Supporting Cryptological Research.” Guttag was asked to prepare a final version of the report, have it approved by the subcommittee chairs (Minker and Thomas Pyke) and the section heads (Curtis and William Rosen), and transmitted to Langenberg. The report urged:
No agency or part of the government should be allowed to bypass the normal means of controlling information by using the National Science Foundation to threaten the funding of those producing the information. Most of the recommendations made in this report have as their implicit goal promoting the clean separation of the procedures for funding and otherwise promoting basic research from the procedures for handling national security and other non-scientific considerations. We believe that the applicability of most of the recommendations contained within this report is by no means limited to the area of cryptology. . . .NSF must continue to support, as Dr. Langenberg put it, “the best research it can find in all areas of science and engineering, with the fewest possible restrictions on investigators.”85
Sometime after I had returned to NSF in 1980, John Cherniavsky, the new program director for the Theory program, and I made several trips to the NSA headquarters at Fort Meade to help them design an open and unclassified basic research program. I believe our work with NSA was in the same time period as the delivery of the Guttag Report. NSA subsequently established an unclassified research grants program, which made its first award in FY 1982. The NSF cooperated in this program and made joint awards with the NSA.
2.8The Computer Science Section, 1979–1984
I left NSF in the fall of 1978 for a position in the Institute for Computer Science and Technology (ICST) at the National Bureau of Standards (now the National Institute for Standards and Technology). George Davida had replaced me as program director for Theoretical Computer Science and, after a year, he was followed by Meera Blattner. Anil Jain had come in to replace Eamon Barrett in the Intelligent Systems program and later was followed by Y. T. Chien. Before I left, Frederick Weingarten left, eventually joining the Congressional Office of Technology Assessment.
The period from 1977 to 1984 saw many changes in the NSF management and, eventually, growing support, if not budget, for computing research. At the director level, in 1976 Guy Stever became Gerald Ford’s Science Advisor. Richard Atkinson replaced Stever as NSF Director through the early NSA discussions. Both Atkinson and his deputy, Donald Langenberg, were supportive of computer science. John Slaughter’s term as Director (1980–1982) was short, but he recruited Thelma Estrin86 of the UCLA computer science department to head the Electrical, Computer, and Systems Engineering Division. Slaughter was also supportive of the CER and CSNET programs. Ed Knapp arrived from Los Alamos in 1982 and served as Assistant Director for MPS for only two months before being named NSF Director. He was very supportive of computer science, CSNET, and NSF’s role in future networking and high-performance computing. When he returned to Los Alamos, Erich Bloch became NSF Director in September 1984. Soon after, Computer Science became a separate division again and, in just two years, part of the new Computer and Information Science and Engineering Directorate.
Within MPS, Ed Creutz retired from his role as Assistant Director for MPS in 1977 and was replaced by Jim Krumhansl from Cornell. Krumhansl was much more supportive of computer science but left in 1979. Bill Klemperer came from Harvard to serve as AD from 1979 to 1981. He was supportive of computing research but skeptical about the section’s leadership. When asked to create a separate Division of Computing Research after Pasta’s death, he brought Jim Infante from Brown University back in87 to head the Mathematical and Computer Sciences Division, delaying the creation of a separate computing research division until 1984. After Klemperer left, he was replaced in MPS by Marcel Bardon on an acting basis. At some point in the fall of 1984, with support from Bardon, Infante, and Bloch, the Mathematical and Computer Sciences Division was split and the Division of Computing Research (DCR) was created.
After the NSF, with the backing of Klemperer, Atkinson, Langenberg, and the National Science Board, decided to allow a new set of programs to address the crisis in experimental computer science research, Kent Curtis contacted me to see if I would be interested in returning to NSF. I was able to retain a visiting research position at NBS while having a chance to make a difference for computer science nationally at NSF. In January 1980 I returned as the program director for Special Projects, which included the Coordinated Experimental Research program and CSNET, as well as the Computer Science Section programs in databases, security and privacy, and social impacts of computing. After almost four years with the Special Projects program, I left in August 1984 for an Independent Research and Development (IR&D) assignment to the University of California, Berkeley.
Before I left for Berkeley, I was assigned part time to the new Office of Advanced Scientific Computing (OASC) to direct the networking programs. While I was at Berkeley, the Division of Computing Research (DCR) was created. Kent Curtis became DCR Division Director and immediately opened a search for a deputy division director (DDD). I applied and was hired into that position beginning in September 1985. My duties were largely administrative as DDD/DCR and were quickly overtaken by my role as a senior scientist for CISE.
In the following sections, I will describe the Coordinated Experimental Research program and its successors, CSNET, the OASC Networking program and the beginning of NSFNET, and the creation of CISE. One goal is to describe the roles of the many people within NSF who helped the computing-related programs begin to grow, thrive, and assume the significant leadership position that CISE holds today.
2.9Addressing the Need for Academic Experimental Computer Science
The Coordinated Experimental Research (CER) program was created in response to a perceived “crisis” in academic computer research. The NSF heard from the Computer Science advisory committees, the Feldman and Snowbird reports, and Peter Denning’s articles in ACM Communications that serious problems were arising in the field. This drumbeat of reports began to have a significant impact on the perception of computer science within the NSF and across the federal government.
In these reports, members of the computing research community pointed to the rapid deterioration of research facilities and the flight of faculty and graduate students to industrial laboratories. Many felt that only three institutions, Carnegie Mellon, MIT, and Stanford, were adequately capitalized to perform experimental research. Only researchers associated at these three universities and, to a lesser extent, other departments and labs with specialized DARPA/IPTO projects, had adequate experimental infrastructure. Even at the major DARPA centers, access was often limited. Remote access to these facilities could be obtained via ARPANET in some cases, but ARPANET access was also limited. As a result, computer scientists at many of the major research universities were engaged primarily in theoretical research and training, graduating fewer Ph.D. computer scientists, and failing to meet the growing demand for experimental computer science faculty.
At the May 1979 Computer Science Advisory Committee,88 Jim Krumhansl, AD/MPS, cited the beneficial effect of recent reports. He noted that Frank Press, the President’s Science Advisor, used the Feldman Report as the basis for recent remarks. Krumhansl, however, admitted that computer science would not be a part of any special initiative. The Office of Science and Technology Policy was said to be considering the “general area as one of national concern and in this is dealing with DARPA, OMB, and any other agencies involved.”89
The Advisory Committee in May 1979 warned of “the eroding research position of the United States in experimental computer science,” applauding “the recent action by the National Science Board to place special emphasis on computer science in FY 1981,” and encouraging “the Foundation to continue that initiative throughout the budgeting and appropriation process.”90 The Committee placed a high priority on five-year Centers of Excellence Grants and a Computer Network for Research. The Centers of Excellence program could