Читать книгу Computing and the National Science Foundation, 1950-2016 - William Aspray - Страница 11

21974–1986: CER, CSNET, NSFNET, and the Founding of CISE

W. Richards Adrion

As we discussed in Chapter 1, computing and information programs and activities existed from the beginning of the National Science Foundation. After several major NSF reorganizations, the computer science and engineering research programs in the Office of Computing Activities were transferred to the Research Directorate in 1974 and the Office was renamed the Division of Computer Research (DCR) in 1975. After the Research Directorate was divided into several discipline-based directorates, the DCR programs were moved into the Computer Science Section of the Mathematical and Physical Sciences, and Engineering (MPE) Directorate in 1976.¹Programs for scientific computing resumed in the early 1980s as support for high-performance computing, and then in the 2000s for “cyberinfrastructure.” Educational applications of computing moved to the Education Directorate in 1972 and, following a brief hiatus during the Reagan administration, remained there. The programs in the Office of Science Information Services (OSIS) moved to the Directorate for National and International Programs in 1969, where they suffered substantial reductions in funds and significant changes in staffing. The NSF science information/information science programs evolved to focus on essential technologies for addressing fundamental questions of information science.

By the 1980s, NSF programs supporting computer science, computer engineering, and information science research had moved from the administrative side of NSF to, or were created within, various divisions and sections in the research directorates. Computing research was housed in Mathematics and Physical Sciences (MPS). The Division of Information Science and Technology was moved to the Biological and Behavioral Sciences Directorate in 1978. After an Engineering and Applied Science Directorate was created in 1978 (becoming the Engineering Directorate in 1980), NSF developed explicit programs for computer engineering and housed them in a new Electrical, Computer, and Systems Engineering Division. A new office of Advanced Scientific Computing was created in 1984 to meet the demand for supercomputer centers and associated networking access. The formation of the Computer and Information Science and Engineering (CISE) Directorate in 1986 brought these programs together in a single directorate.

2.1My Background and Perspective on the 1974–86 Period

Much of this chapter is based on my experience and memory of events, augmented with documents and references. I first joined the National Science Foundation in late summer 1976 and for two years was the program director for the Theoretical Computer Science (TCS) program.² I will describe the creation and operation of the Computer Science Section (CSS) within NSF, issues that arose around cryptography research funded from the TCS program, and the roles of the CISE Equipment program and Theorynet in influencing the Coordinated Experiment Research (CER) initiative.

I returned to NSF in January 1980 as the program director for Special Projects in the Computer Science Section (CSS). My responsibilities included the new Coordinated Experimental Research initiatives: CER (facilities), CSNET, a New Faculty Investigators program, and a Postdoctoral program. In FY 1981, CER and CSNET and the New Faculty Investigators program became separate programs, while the Postdoctoral program was terminated. I managed CER and oversaw C. William Kern, the CSNET project manager. I assumed the role of CSNET project manager in 1982 when Kern left for Ohio State. I was also responsible for other programs in Special Projects including research on databases, privacy and security, and social impacts as well as conferences, symposia, and special studies.

In 1984, I joined the Office of Advanced Scientific Computing as program director for Networking while maintaining responsibility for Special Projects programs, CER, and CSNET. I was on an Independent Research and Development (IR&D) assignment at the University of California, Berkeley, for the 1984–1985 academic year, handing over the CER program to Harry Hedges and the Special Projects program to Larry Oliver. I continued to manage the OASC Networking and the CSNET programs until Dennis Jennings took over the OASC Networking program in January 1985 and when CSNET had become more or less independent under management by the University Center for Atmospheric Research (UCAR) and BBN. I will describe the early efforts for “Sciencenet” that led to NSFNET and the successful spinoff of CSNET and its eventual merger with BITNET to form the Corporation for Research and Educational Networking (CREN).

While I was at Berkeley, I was hired as the Deputy Division Director for the new Division of Computer Research (DCR), which had split off from Mathematics but remained in MPS. When I returned to Washington in the fall of 1985, I had mostly administrative duties in DCR, including upgrading the computing infrastructure within DCR and working with Connie McLindon, the NSF CIO, on NSF-wide technology.

During fall 1985, I also began working with Chuck Brownstein, Division Director of Information Science and Technology, to assist Director Erich Bloch with plans to develop a full-blown computing directorate. In March 1986, he announced that Gordon Bell would be joining NSF to lead the effort. A week or so earlier, Bell had requested that Brownstein take on the role of Executive Officer of the new directorate, Jerry Daen be added as the Planning and Administrative Officer, and I join half-time on loan from DCR. Eventually, I became the Senior Scientist for the Computer and Information Science and Engineering (CISE) Directorate. I will describe the negotiations and planning that went into the first nine months of the CISE Directorate.

I returned to NSF in January 2000 as Division Director for Experimental and Integrative Activities (EIA). Chapter 8 includes a description of the President’s Information Technology Advisory Committee (PITAC) report that led to the government wide initiative on Information Technology for the 21st Century (IT²), the designation of NSF as lead agency, and the planning and experiences that led to the NSF implementation of IT², the Information Technology Research (ITR) program.

In addition, I served on a number of advisory committees and was involved in three more reorganization efforts: chairing the NSF/CISE Committee on CISE Organization in 1995–1997 for Paul Young, chairing the divisional NSF/CISE/EIA Reorganization Working Group in 1997–1998 for Juris Hartmanis, and—as a part-time CISE senior advisor—chairing a committee that advised Peter Freeman on his 2003 reorganization.

2.2Making the Case for NSF’s Computing Research Programs

NSF provided funding for computing, communications and information infrastructure, applications, and fundamental research from its beginning. The physical scientists who ran the NSF were not quite sure there was a “discipline” of computer science, but they clearly appreciated the growing importance of computing, communications and information infrastructure, and applications. Scientific and engineering disciplines typically turned to related professional societies or the National Academies to describe the field, its accomplishments, and its future promise. An influential report was needed to define computer science, its value to the nation, and the need for investment and support.

The professional societies—ACM, IEEE-CS, AFIPS, SIAM, and AAAI—established the conferences and journals in this new field. None of them adequately represented academic computer science research in Washington, DC. This gap led to the creation of the Computer Science Board in 1972, later renamed the Computing Research Association (CRA), which created a Washington presence in 1988. Ever since, CRA has played an important role in advocating for the computing research community.

From 1978 to 1986, the National Academy Board on Telecommunications and Computer Applications primarily published reviews of information technology issues and challenges experienced by federal agencies such as the Social Security Administration, the Internal Revenue Service, NASA, and the Departments of Defense and Commerce. One exception was a 1982 report from an ad hoc committee on the roles of industry and the university in computer research and development.³ The National Research Council created the Computer Science and Telecommunications Board (CSTB) in 1986 to replace the Board on Telecommunications and Computer Applications.

Earlier in the 1960s, a number of individuals attempted to define computer science as a discipline. In addition to Louis Fein’s⁴ efforts described in Chapter 1, Saul Gorn of the University of Pennsylvania wrote in 1963 that “a new basic discipline is emerging which might be called ‘The Computer and Information Sciences’ [that] makes application of concepts from the traditional fields of mathematics, philosophy, linguistics, psychology, engineering, management science, library science, etc.”⁵ George Forsythe,⁶ the founder of Stanford’s computer science department and ACM President, commented on Gorn’s analysis, suggesting that computer scientists are concerned with the pragmatics of the applications of mathematics. In 1967 Allen Newell, Alan Perlis, and Herbert Simon⁷ defined computer science as the study of phenomena related to computers. Donald Knuth’s definition⁸ of computer science as the study of algorithms appeared in 1968. Curriculum 68⁹ defined computer science as the study of information structures. Edsger Dijkstra defined computer science as the study and management of complexity.¹⁰ Historian Janet Abbate observed that computer scientists, in arguing for scientific status of their field, drew on “three distinct meanings of science (sometimes in combination)”¹¹: (1) science as the study of natural phenomena (information in this instance),¹² (2) science as the derivation of abstract ideas from concrete phenomena,¹³ and (3) the experimental method as the defining characteristic of science.¹⁴

These assertions about computer science as a science did not persuade NSF management that computing was or was beginning to be a mature scientific discipline. Abbate notes: “. . . organizational control wielded by the established disciplines, as well as NSF’s emphasis on basic research, put the emerging field of computer science at a disadvantage. In this context, the notion of computing as a ‘science’ and the appropriateness of NSF funding for computing researchers were both contested.”¹⁵

After NSF moved the Office of Computing Activities into the Research Directorate, renaming it the Division of Computing Research (DCR) in 1974, the weak support for computer science as a discipline resulted in DCR programs being placed in a section (CSS) within a Mathematical and Computer Sciences Division in the Mathematical and Physical Sciences, and Engineering Directorate in 1976. When DCR was created, Gordon Bell, then with Carnegie Mellon University and Digital Equipment Corporation, was “concerned about funding for computer science within the National Science Foundation and that we [the computer science community] lack representation on the National Science Board.”¹⁶ Saunders MacLane, a Chicago algebraist on the National Science Board (NSB), was a good supporter of computer science but not a true representative of the discipline. NSF provided a 12.2% increase for Computer Science research for FY 1976, while MPE overall was increased 6.3%. The $13.22 Computer Science research budget, however, was only 6.6% of the total MPE budget.

To offset the perception that computing research was well-served by industry, Bell argued that funding for basic research in computing should be directed to universities and not industry. Bell added that while mission agencies, such as ARPA, played a significant role, NSF had the role of supporting basic computer science research. Bell also suggested that NSF funding of basic computer science research introduce a “question of scale” and that NSF consider investments of an ARPA-like magnitude in several non-ARPA-funded, leading computer science programs.¹⁷

Facing skepticism from NSF leadership about the emerging field of computer science and its core research questions, John Pasta and Kent Curtis mobilized influential scientists. In 1974, they funded the Computer Science and Engineering Research Study (COSERS) under the direction of Bruce W. Arden of Princeton University. “For the first time in its quarter century of activity . . . this discipline will be given a comprehensive examination by researchers in the field. . . . The report will define what computer science and engineering is, describe major research problems now under investigation, and point out future educational and research opportunities.”¹⁸ Apart from a brief progress report¹⁹ in 1976, the massive 1000+ page report, What Can Be Automated?: Computer Science and Engineering Research Study,²⁰ unfortunately did not appear until 1980. By that time, other influential reports had appeared and diminished its impact.

By the late 1970s, Curtis and Pasta were working with leading members of the computer science community to address a serious concern about the health of academic computer science. Academic salaries were falling behind industry salaries, there was a significant lack of computing equipment except at the ARPA-funded departments, undergraduate enrollments were rising, and many scientists, engineers, and policymakers still viewed computer science as consisting only of programming, computing applications, and hardware development. Faculty, new PhDs, and promising graduate students were leaving academia for industry in large numbers.

The NSF sponsored a workshop in Washington, DC, on November 2, 1978, led by Jerome Feldman (Rochester) and including Gordon Bell (DEC), Bernard Galler (Michigan), Patricia Goldberg (IBM), John Hamlin (Missouri), Eliot Pinson (Bell Labs), Ivan Sutherland (CalTech), and William Robert “Bert” Sutherland (Xerox PARC). The “Feldman Report,”²¹ also published in the Communications of the ACM, called for universities to recognize the special resource needs of experimental computer science, use appropriate criteria in evaluating experimental computer science programs and faculty, and encourage cooperative programs. While it called for industry to exchange and share people and technology with universities and provide funds and equipment, it looked to government to modernize tax and patent policies, develop funding of adequate scale and time-horizon for experimental computer science, and identify a lead agency responsible for computing. The report proposed large, 5-year capital resources to produce 25 well-equipped university laboratories for a total cost of about $15 million yearly. This recommendation led to the Coordinated Experimental Research (CER) program described below.

The Feldman Report was enthusiastically expanded upon by the ACM Executive Committee,²² the Computer Science Board (sponsor of a 1980 meeting of computer science department chairs²³ now known as the biennial Snowbird Meeting), the 1979 NSF Computer Science Advisory Committee,²⁴ and a series of ACM letters and articles by Peter Denning that began with his well-known “eating our seed corn” letter.²⁵ For example, a panel at the 1980 Snowbird Meeting²⁶ addressed the nature of computer science, advances in computer technology, and how computer scientists might address societal implications. These panels and reports changed the perception of computer science within the NSF and across the federal government.

2.3The Importance of Computing Research and Infrastructure to the Nation

For a time after the Computer and Information Science and Engineering (CISE) Directorate was created in 1986, the computer science advisory committees remained associated with the CISE division that Kent Curtis directed. At the request of the NSF Advisory Committee for Computer Research, a subcommittee appointed by Curtis submitted a preliminary report²⁷ at the committee’s meeting held on December 5–6, 1986. This report was revised, published in 1989, and became known as the “Hopcroft-Kennedy Report.”²⁸ Kent Curtis passed away December 17, 1987, and the Hopcroft-Kennedy Report was completed after Peter Freeman had replaced Curtis as DD/CCR.

Frank Press, then president of the National Academy of Sciences and chairman of the National Research Council, created in 1986 what came to be called the Computer Science and Telecommunications Board (CSTB), with Joseph Traub as chair and Marjory Blumenthal as executive director. Under Traub, Blumenthal, and their successors, CSTB published many influential (and occasionally controversial)²⁹ reports. Their first efforts did not try to identify the achievements and opportunities of computing research as did the Hopcroft-Kennedy Report. Traub noted, “CSTB decided that beginning with a report on the nature of the field would be self-serving. We wanted first to build a record of reports dealing with critical national issues.”³⁰ The Board published Toward a National Research Network³¹ in July 1988 and The National Challenge in Computer Science and Technology³² in September 1988. Among the many CSTB reports are ones the Academy characterizes as “explaining how information technology evolves, the role of R&D, and the role of different contributors, public and private, to that process.” These include³³ Innovation in Information Technology, Making IT Better, Funding a Revolution, Evolving the HPCCI to Support the Nation’s Information Infrastructure, and Computing the Future. While clearly influential, one of the criticisms of the Academy and the National Research Council, as voiced by Ed Feigenbaum (perhaps a bit sharply), is that they are “extremely slow and conservative organizations, unwilling to say things that make anyone bristle. So, a lot of what CSTB might try to do is either squashed or squashed in advance by this elaborate structure.”³⁴ Until the CRA was chosen to create the Computing Community Consortium in 2006, the options for fast response “blue ribbon” reports remained limited.³⁵

Beginning in the 1970s when NSF reduced its support for computing facilities, concern grew in the scientific community that future scientific advances would be impeded by the lack of advanced computers. Moreover, a number of countries were developing “supercomputers”³⁶ and programs to increase access³⁷ for their scientists. An interagency study group, led by Peter Lax of NYU, made the case³⁸ (known as the “Lax Report”) for a program that would increase access to supercomputers via high bandwidth networks; increase research on computation, software, and algorithms; train personnel; and increase R&D on new supercomputer systems. Ken Wilson, then at Cornell and a recent Nobel laureate, was one of the leading proponents of a program in supercomputers and a national network to connect them. In an attachment to the Lax Report, Wilson stated that “the lack of large scale scientific computing resources for basic university research has become a major problem for U.S. science.”³⁹ In this, he advocated for a national network linking all scientists and support for a collaborative program in large-scale scientific computing hardware, software, and algorithms led by the science, computer science, and electrical engineering communities and industry. As part of his advocacy, Wilson coined the term grand challenges.

A four-part federal program was proposed⁴⁰ and, in mid-1983, an internal NSF working group, led by Marcel Bardon and Kent Curtis, laid out specific actions they recommended that NSF take (the “Bardon-Curtis Report”).⁴¹ These actions included providing “supercomputer services for academic research and science education . . . ” and supporting “networks linking universities and laboratories with each other. . . . ” Following a report⁴² from a panel on “Computer Architecture,” led by Jack Schwartz of the NYU Courant Institute on behalf of the National Academies’ Committee on Science, Engineering, and Public Policy (COSEPUP), NSF Director Erich Bloch asked the engineering and physical sciences directorates and the newly formed Office of Advanced Scientific Computing to comment on Schwartz’s suggestion that NSF “might strive for a position of higher importance and impact” in high-performance computing.⁴³ The response recommended that reaching the Schwartz panel’s recommendation would “require a well coordinated federal effort among at least the following agencies: DOD (including DARPA, ONR, AFOSR, and ARO), DOE, NASA, NBS, and NSF . . . [and] it is appropriate that NSF provide more leadership because of its independence from mission criteria in selecting research projects for support and because of the excellent technical judgment it can bring to bear.”⁴⁴

The emphasis on networking in the Lax and Bardon-Curtis reports led to a series of reports on networking including the Sciencenet⁴⁵ proposal and the initial ideas for NSFNET.⁴⁶ These and CSTB reports provided background for the Federal High-Performance Computing program and NSF’s programs in advanced scientific computing and networking.

2.4Computing and Information Research in NSF, 1974–1978

By the late 1970s, the programs that would be joined to form the Computer and Information Science and Engineering (CISE) Directorate were in place but divided among several NSF research directorates. In 1974, NSF transferred the Office of Computing Activities to the Research Directorate and renamed it the Division of Computer Research (DCR). In 1976 DCR merged its sections and programs into the Computer Science Section (CSS) of the Division of Mathematical and Computer Sciences (DMCS) within the new Mathematics, Physical Sciences and Engineering (MPE) Directorate.

DCR had two sections from 1974 to 1975: computer science and engineering, and computer applications in research. The former ran programs in theory, programming languages and systems, and systems design. The latter ran programs in techniques and systems, software quality research, and networking for science. The FY75 NSF Annual Report includes these comments:

The discipline of computer science is barely 10 years old, only vaguely defined, and mushrooming. . . . In a field as new and as rich as computer science it is not surprising that new areas appear, create a flurry of activity, and then level off or stagnate; automata theory, mechanical translation, and theory of formal languages are a few such . . . researchers in computer science are anxious to follow new leads into uncharted regions. This kind of process of extension to new areas and pruning of less productive ones partly accounts for the lack of definition of the field.⁴⁷

The report goes on to suggest that the Arden COSERS initiative, described above, was a necessary disciplinary self-examination. By the time of my arrival at NSF,⁴⁸ toward the end of Transition Quarter 1976,⁴⁹ the Assistant Director for the Mathematical and Physical Sciences and Engineering (MPE) Directorate, Ed Creutz, had decided to merge computer research with mathematics.

John Pasta became Division Director for the Division of Mathematical and Computer Sciences (MCS). The three sections in DCR (Computer Science and Engineering, Computer Applications, and Computer Impact on Society) became one section within MCS. Kent Curtis, who had been section head for Computer Science and Engineering (CS&E), became section head for the Computer Science Section (CSS). William H. Pell led the Mathematics Section. Don Aufenkamp, who had been section head for applications, took over the NSF US-USSR program. The program directors in the DCR Computer Science and Engineering Section—Bruce Barnes (Theory), Thomas Keenan (Programming Languages and Systems), and John Lehman (Computer Architecture)—moved with their programs into CSS. Sally Sedelow from Techniques and Systems became the Intelligent Systems program director in CSS. Fredrick Weingarten became Special Projects program director. Walter Sedelow came over from the applications section, where he had overseen computer networking-related grants, to join Weingarten in Special Projects. While Kent had recruited me for the Software Engineering program, he decided to have Bruce Barnes head that program because of his experience and interest. I was assigned instead to the Theoretical Computer Science program. The Sedelows left in 1977, and Sally was replaced by Eamon Barrett (from ESL Inc.)⁵⁰ and Walter by Larry Oliver (from NSF Education).

Engineering, which also was a division in MPE in 1976, had an Electrical Sciences and Analysis Section, which funded research on digital systems and communications, and information theory. Later, after a possibility that a separate National Engineering Foundation might be created, NSF merged applied research and engineering to create a new Engineering Directorate with an Electrical, Computer, and Systems Engineering (ECSE) Division. Steve Kahne, Thelma Estrin, and others served as ECSE division directors. The Division of Science Information in the Scientific, Technological, and International Affairs (STIA) Directorate supported fundamental research on information sciences and applied research on information access and user requirements. This division would later be renamed the Division of Information Science and Technology and moved to the Biological and Behavioral Sciences (BBS) Directorate.

The new Computer Science Section had six programs—Theoretical Computer Science, Software Systems Science, Software Engineering, Intelligent Systems, Computer Systems Design, and Special Projects—each described in the NSF Guide to Programs as shown in Figure 2.1 (before Software Engineering was added). The programs had no deadlines, target dates, or solicitations; and all proposals were essentially “unsolicited” without restrictions on page length, format, font size, etc. Prospective principal investigators were encouraged to submit proposals in the fall if they wanted summer funding for the following year.

William Aspray writes in Chapter 7 , “Foundation staff did not generally set a research agenda for funding. They relied instead on the scientific community to set the agenda, both through the proposals individual scientists submitted and the reviews the scientific community gave to these proposals.” I would argue that, while we placed no constraints on what could be submitted and solicited no proposals, the program directors, Kent Curtis, and John Pasta were very proactive in encouraging people to submit and in publicizing the programs. The proposals the section funded and the people we encouraged, in effect, defined an agenda.

Figure 2.1FY 1977 NSF guide to programs: Computer science research. (Source: National Science Foundation (1976) Guide to programs, FY 1977.) https://hdl.handle.net/2027/mdp.39015043526683; last accessed 6 June 2018.

Before FastLane⁵¹ made web-based submissions possible, proposals were mailed to NSF with approximately 25 copies arriving in the one office that processed all arriving proposals. After the CSS administrative officer picked up the proposals from central processing and distributed copies to the program officers, they would do a quick check on the appropriateness and redistribute if needed. Since the volume of applications was modest,⁵² program directors took time reading each proposal in detail and consulting colleagues for suggested reviewers. One also could walk down the street to the George Washington University Library (or use the much smaller NSF library) to read related or cited papers to help in understanding the proposals and selecting reviewers.

Typically, a program director needed three to four reviews to support a recommendation. Proposals were sent to six to eight reviewers, given the low response rate. These reviews were carried out as “mail reviews,” that is, copies of the proposals along with a review form and check boxes for an “adjective” review (poor, fair, good, very good, excellent). Proposals were triaged: the clearly fundable proposals were recommended as soon as possible, the clearly non-fundable proposals were declined, and the remaining were held for discussion in weekly meetings with all six program directors, Kent Curtis, and often John Pasta. In these meetings, we discussed the status of our programs and the awards and declinations we were planning. These were often lively discussions about priorities and high-risk proposals.

The primary issue delaying recommendations was the time it took to get solid reviews. “We read the comments very carefully, used our best judgement, and did not really put much weight on the adjective ratings.”⁵³ The directorate, however, did consider the ratings and compared our recommendations against the other programs in MPE/MPS. The field was young and the “shooting inward”⁵⁴ phenomenon was at its height. Our first strategy was to plead with the researchers to evaluate proposals fairly and to understand that there were risks that could be overcome with good new approaches. Our second strategy to address both response rate and review quality was to employ John Lehmann’s skill at mining the NSF databases. We gathered data for every reviewer on the time to review, the number of reviews, and the average review, and compared their performance with other reviewers of the same proposals. So, if Mary Smith seldom gave “excellent” ratings and typically gave ratings below those of other reviewers, we could use that in the recommendation. Our next strategy was to remove the adjective ratings from the review forms entirely. This had two good outcomes: it left interpretation more to the program directors rather than depending on scoring, and the lack of the option to just check a box resulted in longer and more thoughtful reviews. In the long run and because of a desire to have uniform measures across the Foundation, however, we were asked to return to using adjective reviews.

Computing research funding rose relatively slowly over the period 1974–1980 (see Figure 2.2) with the first significant increase coming with the establishment of the Coordinated Experimental Research Programs (described below) in 1980. There were several ways in which we managed our program portfolios.

Once an adequate number of reviews arrived, we would seek out other program managers in computer science, mathematics, engineering, or information sciences to discuss “split funding” if appropriate. Typically, there was little budget gain from these transactions inasmuch as we might co-fund as many proposals in other programs and divisions as we would get co-funding from them. It did contribute, however, to a broader understanding of computing and computer science and, as I will discuss, some quite important joint-funded grants were made. One important feature of the 1800 G Street NW NSF headquarters building was a “senior staff lunchroom” on the 12th floor, where program officers would grab lunch and join program officers from other offices and directorates at the few available tables. These casual meetings led to collaborations, joint funding, and collegiality. Unfortunately, due to its size, entrance to the lunchroom was limited by grade level, thus barring junior program officers and clerical and administrative staff. Erich Bloch closed the lunchroom for just this reason.

Figure 2.2Computing research funding FY 1974–1980.

I believe the process I have described led to thorough and thoughtful reviews and recommendations, which corrected the perception that computing research proposals were of comparatively lower quality. The number, breadth, and quality of the research the Computer Science Section supported under its constraints and with limited funds demonstrates an effective stewardship of NSF investments in a growing field.

2.5Funding the Innovators in Computer Science

It is not easy to measure the impact of individual funding decisions on the health and growth of computer science. One indicator might be the role NSF played in the careers of Turing awardees. The A. M. Turing Award is the oldest and most prestigious award⁵⁵ in computing. It is presented annually by the Association for Computing Machinery (ACM)⁵⁶ “to an individual who has made lasting contributions of a technical nature to the computing community.”

Many of the Turing Award winners from the 1960s and 70s were in industry (Maurice Wilkes, Richard Hamming, Charles Bachman, John Backus, and Kenneth Iverson), Europe/Israel (Wilkes, Jim Wilkinson, Edsger Dijkstra, Michael Rabin), or the (D)ARPA-funded universities (Alan Perlis, Marvin Minsky, John McCarthy, Don Knuth, Allen Newell, Herb Simon, Dana Scott, and Bob Floyd). However, Don Knuth received significant NSF funding for the work that went into his The Art of Computer Programming⁵⁷ series and the development of TEX.⁵⁸ Both Alan Perlis and John McCarthy were involved with NSF facilities grants in the 1960s, which provided them with an environment for their early work on programming languages and operating systems. John McCarthy was funded by multiple NSF programs during the later 1970s.

During the period from 1976 to 1978, the Computer Science Section launched the research careers of many future Turing Award winners. Of the winners from the 1980s through 2017, again some spent most or all of their careers in industry/government or outside the United States.⁵⁹ From 1976 to 1978, the Theoretical Computer Science (TCS) program made grants to Richard Karp, John Hopcroft, Robert Tarjan, Juris Hartmanis, Manuel Blum, Amir Pnueli (as a visitor at University of Pennsylvania), Andrew Yao, Leonard Adelman, Ronald Rivest, Adi Shamir, Judea Pearl (with Intelligent Systems), Martin Hellman, and Whitfield Diffie (Hellman and Diffie with Engineering Systems). Michael Stonebreaker was funded from the Special Projects program in 1980. At a later time, the TCS program funded Leslie Valliant and Shafi Goldwasser. Edward Clarke, Alan Emerson, Barbara Liskov, and Leslie Lamport all received NSF grants from the Software Systems Science program.

Many of the Turing Award winners, including those in industry, played significant roles in advising and advocating for computer science within NSF, including John Hopcroft and Fredrick Brooks, who both served on the National Science Board (NSB). Vinton Cerf and Robert Kahn were important to the development of CSNET and NSF, and Cerf recently served on the NSB.

In addition to the Turing awardees, there were other important grants made in the 1976 to 1980 period. I discuss cryptography and security below, including the work of Hellman and Diffie (Stanford); Rivest, Adelman, and Shamir (MIT); George Davida (Wisconsin Milwaukee); and Dorothy Denning (then at Purdue and SRI). Lawrence Landweber’s Theorynet project and the concurrent analysis of collaboration over networks by Starr Roxanne Hiltz were important in building support for CSNET and later NSFNET. The Coordinated Experiment Research (CER) program addressed the national issue described by the Feldman and Snowbird reports, but the successful grantees would not have succeeded without the equipment grants (typically VAXes and PDP-11s), which initiated experiment work in the grantee departments and established their credibility as potential centers of experimental research.

In Chapter 7, William Aspray describes research done by some of the most respected, NSF-supported scientists. During 1976–1978, most of these people were funded by the Computer Science Section. Mary Shaw and Barbara Liskov were among many influential women researchers funded by CSS in the late 70s—a group that included Sue Graham, Sherry Turkle, Irene Grief, Lori Clarke, Anita Jones, Mary Jane Irwin, Ruzena Bajcsy, Nancy Lynch, Diane O’Leary, Shari Pfleeger, Elaine Cohen, Sheila Griebach, Dorothy Denning, and Naomi Sager.

Additional grants from this period illustrate the impact of the Computer Science Section. The work of Arthur Burks and John Holland became the basis of classifiers used in machine learning. The Stanford AI lab (with John McCarthy, Edward Feigenbaum, and Cordell Green) moved artificial intelligence ahead. The Ingres Relational Database developed by Michael Stonebreaker, Lawrence Rowe, and Eugene Wong was arguably the first practical research relational database. Concurrently, the Division of Information Science and Technology funded early work in Information Retrieval by Gerald Salton, Naomi Sager, Michael McGill, and several others.

2.6Facilities

In his book on applications of case study research, Robert Yin took David Gries’s abstract from the final report on the first Coordinated Experimental Research (CER) grant to Cornell and analyzed it as a case study. Yin quotes Gries’s final report and identifies the outcomes:

From 1980 to 1986, the Computer Science Department at Cornell was radically transformed from a theoretical, pencil-and-paper research operation to one with a high degree of experimental computing. The departmental computing facility grew from a VAX780 and a PDP11/60 to an integrated complex of almost 100 workstations and UNIX mainframes. All faculty and graduate students now use computing daily, and much research that was hitherto impossible for us is now being performed.

The CER grant enabled the department to attract bright young faculty who would not have joined a department with inadequate facilities. As a result, the department has been able to branch out into new areas, such as VLSI, parallel architectures and code optimization, functional programming, and artificial intelligence. The CER program did what it set out to do: It made it possible for the department to expand its research activity, making it far more experimental and computing intensive while still maintaining strong theoretical foundations.⁶⁰

The CER program was transformative in the ways that Gries describes. Earlier support for the VAX780 and a PDP11/60 likely came from the Computer Science Research Equipment program. The program made seven grants in FY 1977, totaling $753,200, to departments that later received CER grants: Cornell, Arizona, Illinois, Pennsylvania, Utah, Washington, and Wisconsin. It also made grants to other departments, totaling $817,700, some of which competed unsuccessfully in the CER program. In FY78 the program made eight more grants, totaling $790,249, to departments that later received CER grants: Brown, Stony Brook, Berkeley, Illinois, UMass, Utah, and Wisconsin; and additional grants totaling $755,403. In many ways, the equipment program was as important to the computer science community as CER, moving departments from “pencil and paper” to a point where they had facilities for experimental research. The program continued, under various titles, from the 1980s until 2001, when the CISE Research Instrumentation program was incorporated into a revised CISE Research Infrastructure program along with the successors to the CER program. I will return to the CER infrastructure programs below.

There were attempts to develop a national research network, or regional ones, prior to the Office of Computing activities move to the Research Directorate. Don Aufenkamp, then head of the OCA Applications in Computing Section, announced at the 1972 EDUCOM meeting that NSF was going to sponsor research that might lead to a network linking universities and other institutions.⁶¹ He and Ed Weiss delivered a paper⁶² at the International Conference on Computer Communication in October 1972 discussing further details. In Science in October 1973, Greenberger et al. noted that NSF had funded

. . . EDUCOM to bring together interested users and administrators with those possessing shareable resources and relevant experience in a series of three 2-day working seminars. . . . The seminars . . . were designed to help identify the central organizational, political, and economic issues in building and operating networks on a national basis.⁶³

What happened to this effort is not at all clear. It may have been inspired by the success of the CONDUIT regional networks described in Chapter 1, but with a broader national vision. Historian Janet Abbate⁶⁴ speculated that it was because the Office of Computing Activities (and its successor, the Division of Computing Research) had a limited budget or because the importance to researchers was not yet realized. When I arrived in 1976, NSF leadership was not interested in anything of the scale and management demands of an ARPANET-like national network and remained unconvinced that a network for sharing resources and collaboration had value, given the cost.

An opportunity for NSF to be involved in networking arose in 1977. Fredrick Weingarten inherited what was left of the DCR applications efforts in his Special Projects program. I knew that he was supporting research on the economics and social impacts of networks and computer-based collaboration, often jointly with Edward Weiss and others in the Division of Information Science and Technology. An opportunity surfaced at the 1977 Foundations of Computer Science (FOCS) conference in Providence, RI, where I met with several researchers during the conference reception at the Marriott Inn. We discussed whether NSF might entertain a proposal to support an email system for collaboration. The group included Lawrence Landweber, Richard Lipton, Richard Demillo, and Edward Robertson. After the meeting, Fredrick Weingarten and I decided to encourage Landweber to submit a proposal (NSF 7801689, An Electronic Mail-Box and Teleconferencing Network for Theoretical Computer Science). Landweber agreed to add Starr Roxanne Hiltz of Uppsala College to analyze the impact on collaboration and research output.

Thirty or so theoretical computer scientists in the United States and Australia participated in the Theorynet project by using Telemail running on a University of Wisconsin computer and accessing it over Telenet,⁶⁵ a commercial packet-switched network. Research collaboration rose steadily and the 1978 ACM SIGACT program committee communicated via Theorynet/Telemail. Although she had some difficulty monitoring usage and interviewing users, Hiltz⁶⁶ was able to show positive outcomes in terms of collaboration and jointly published papers. Theorynet’s modest success lent credibility to the CSNET and NSFNET projects.

2.7Cryptography and Interactions with the National Security Agency

The various controversies that arose around the National Bureau of Standards (NBS) Digital Encryption Standard (DES) were a prologue to issues related to cryptography. IBM submitted a cryptographic algorithm as a candidate for the DES in 1974 and NBS requested that the NSA evaluate it.⁶⁷ NBS also asked IBM to grant the U.S. government “a nonexclusive, royalty-free license to make, use, and sell apparatus that implemented the algorithm.” NBS published a notice in the Federal Register in August 1975 of the proposed standard and requested comments. Martin Hellman and Whitfield Diffie of Stanford University criticized the proposed DES standard and outlined a potential attack on the algorithm.⁶⁸ In April 1977, NBS issued the DES standard.⁶⁹

In November 1976, in the midst of the DES controversy, Diffie and Hellman published “New Directions in Cryptography,”⁷⁰ which introduced several new concepts: public key cryptosystems, one-way authentication (or functions), trap-door one-way functions, and digital signatures. At about that time, El (Elias) Schutzman, the program director in engineering systems, approached me about co-funding grants to Hellman at Stanford and I agreed. Diffie was at the time a research assistant working with Hellman. I was also funding Ronald Rivest, who was developing a number-theoretic public-key encryption algorithm⁷¹ with Leonard Adleman and Adi Shamir (all at MIT), which became the RSA algorithm.

What we did not know at the time was that James Ellis of the British Communications Electronics Security Group (CESG) had published a classified paper⁷² containing the idea of public key cryptosystems and that Clifford Cocks, also with CESG, had proposed an implementation similar to RSA.⁷³ Both of these British papers predate the Americans’ work by four to five years; however, since they were classified, the NSF-funded researchers would not have known about them before CESG de-classified the work in 1997. The National Security Agency, however, was aware of Ellis’s and Cocks’s work.

In August 1977, J. A. Meyer of Bethesda, Maryland (later identified as an employee of the National Security Agency), wrote to the IEEE suggesting that some attendees at the September 1977 IEEE Symposium on Information Theory held at Cornell might be violating provisions of the International Traffic in Arms Regulations (ITAR) Act. Hellman and Rivest turned the problem over to their universities’ lawyers and opted to wait until the lawyers finished looking into the issue. Cleared to attend, they both limited their public discussion to the mathematical and technical aspects of cryptography and did not discuss possible national security implications of their work.⁷⁴

In April 1978, the NSA placed under a secrecy order a patent application from the Wisconsin Alumni Research Foundation on behalf of George Davida of the University of Wisconsin-Milwaukee.⁷⁵ NSA invoked provisions of the Invention Secrecy Act preventing Davida from discussing any aspect of this research and severely limiting his ability to pursue research in cryptology for several months. The secrecy order was later lifted. That fall, Davida joined NSF, replacing me as program director for Theoretical Computer Science.

The American Council on Education (ACE), in response to a request by the NSA, assembled the Public Cryptography Study Group⁷⁶ in March of 1980. NSA indicated concern that information contained in some articles in professional journals and in monographs might be a risk to national security. The study group held a series of meetings through February 7, 1981, and produced a report⁷⁷ that recommended a voluntary system of review of papers in cryptology. No author or publisher would be required to participate. Davida contributed a minority report⁷⁸ that argued against restraints on non-governmental research in cryptography.

According to a report in Science in August 14, 1980,⁷⁹ Leonard Adleman was told by an NSF program officer that parts of his grant proposal would not be funded. Vice Admiral Bobby Inman, NSA Director, was quoted as saying that the reason the NSF chose not to fund parts of Adelman’s proposal is that NSA wanted to fund the research itself. Soon afterward, NSA Director Inman wrote to Science indicating that NSA, as the government’s primary user of cryptography, was increasingly interested in investing in primary research in cryptography as well as related fields, such as mathematics. He mentioned NSA’s assistance with evaluating NSF research proposals in cryptographic areas but stated, “NSA does not now have and does not intend to seek the authority to prohibit NSF funding in this area.”⁸⁰ Inman hoped that NSA would become an increasingly important sponsor of research in this area.

In November 1980, NSF Acting Director Don Langenberg clarified the respective roles of NSF and NSA in support of cryptologic research.⁸¹ Since 1977, NSF routinely referred proposals with relevance to cryptology to NSA for review. The process I used⁸² as program director for Theoretical Computer Science, following guidance from the NSF management and attorneys, was to include a designated NSA expert among the referees from whom I solicited proposal reviews.

Langenberg stated that NSF long had a policy of encouraging other agencies to support basic research and had encouraged NSA to establish an unclassified basic research program, but “if an investigator prefers to apply only to NSF, the proposal will be processed in the usual manner, without prejudice.” Langenberg added that the Foundation would ensure reporting requirements that would allow it to meet its responsibilities with respect to classification.⁸³ The Adleman proposal was approved by the NSF on December 9, 1980, and the award letter included a statement of NSF policy and elaborated reporting requirements. After negotiation between NSF and MIT, a grant was made to Rivest on September 25, 1981, with an altered policy on reporting.

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 provide up to $2,000,000 for capital investment plus up to $500,000/year for operating and maintenance costs and software development. They placed a somewhat lower priority on new investigator and career development awards, graduate fellowships, and traineeships.

In a report⁹¹ issued in December 1979, the Advisory Committee recommended that NSF invest $15 million each year in a national competition for resource grants. These grants would “total no less than $250,000 and no more than $2 million,” include maintenance and software support of 10% of the capital costs and be available to individual researchers and departments. The report expected that after five years, the program would “produce at least 25 well-equipped university laboratories among the 64 computer science Ph.D. degree granting universities.”

John Pasta and Kent Curtis responded to these recommendations by “taxing” the Standard Research Projects Support (SRPS) budget in FY 1980 by $1 million, almost the entirety of the FY 1980 budget increase, to create the Coordinated Experimental Research (CER) program. It had three main thrusts: a CER facilities program, a program to assist the research community in developing networking services in support of computer science research, and grants to attract experimentalists into a university environment. Curtis sent a “dear colleague letter” in November 1979 inviting proposals for what would become the CER facilities program. Program descriptions for a New Investigator program and a Postdoctoral program came after for FY 1981 funding. In the first year the CER program funded one facilities grant and a CSNET study grant.

Today, electronic “dear colleague letters” quickly gain broad audiences, but in 1979 Curtis mailed a letter to the computer science Ph.D.-granting departments. The timeframe was short and we received only seven proposals. Predictably, most of the proposals came from people familiar with experimental computer science within NSF and the Computer Science Section. I said in a 1990 interview, “in the first set of proposals there was one good proposal, one sort of half-good proposal, and the rest of them were . . . bad proposals [although they involved] some very good people”⁹²—bad in the sense that they seemed to be independent projects “stapled together” rather than a unified coherent proposal.

Our first challenge was to determine how to review the proposals. We decided on a multi-stage process: mail reviews, site visits, and a final panel. We decided that I with two external reviewers would personally visit the sites of all of the proposing institutions, following receipt of mail reviews. Principal Investigators would have the mail reviews to react to, along with questions by the site visitors. The final panel included no academics but instead the heads of major industry and non-profit laboratories.

The first CER award was made to the University of Washington to construct the Eden operating system with a goal to build a system coupling the performance of powerful personal machines with the resource sharing and accessing capability of a modern time-sharing system. This major research project involved a majority of the departmental faculty and produced a facility that could support a variety of research projects. The Eden Project attracted co-funding from Intel and Digital Equipment, whose technical staff collaborated with Washington on the research.

In FY 1981, CER became an official program with $3.6 million dollars of the Special Projects budget identified as “experimental computer science” with other line items for CSNET, young investigator, and postdoctoral awards. As we discuss below, a revised set of CSNET proposals were received and approved by the National Science Board. One postdoctoral award and four new investigator awards were made. The CER program received 24 proposals responsive to the program announcement, which were distributed to other NSF, Office of Naval Research (ONR), and DARPA programs. We hoped for significant DoD involvement in developing CER sites as DARPA, ONR, the Air Force Office of Scientific Research (AFOSR), and the Army Research Office were planning a “Computer Resources Initiative” in FY 1982 with $30 million among the DoD science agencies.

With cooperation from the DoD agencies, we selected 11 proposals for site visits similar to those of the prior year. From the eleven sites visited, five were discussed with DARPA and ONR. Following a budget negotiation, four proposals went to the National Science Board, which approved three immediately and a fourth later. DARPA eventually funded a version of a fifth proposal. The four new NSF CER awards went to Cornell to support investigation into the programming process, Illinois for the construction of computer aids to program and system development, the University of Wisconsin–Madison for construction of a 50-node network of powerful computing devices, and Yale for facilities to support artificial intelligence and natural language processing, numerical computing, and computer architecture.

In the succeeding years when I managed the CER program, five awards were made in FY 1982 to Rice, Brown, Utah, UCLA, and Texas. Four awards were made in FY 1983 to North Carolina Chapel Hill, Pennsylvania, Maryland College Park, and Duke. SUNY Stony Brook, Rochester, Arizona, and New York University received grants in FY 1984. After I left for Berkeley, Harry Hedges joined NSF from Michigan State to run the CER program. In FY 1985, Hedges and Bruce Barnes made awards to Princeton, UMass Amherst, Colorado Boulder, and Minnesota.

As the CER program began, we did not completely agree on its goals. Some supported the concept of “Centers of Excellence”; some supported funding large, multi-investigator “experimental” research projects; and some promoted large-scale facilities grants, which would include equipment, maintenance, supplies, and technical staff. Clearly the Eden Project fell into the large, multi-investigator “experimental” research project category. While I personally favored a focus on large, collaborative research projects, there were few awards in this category. Reviewers prioritized facilities support and grants to institutions with an existing core of potential experimental computer scientists. Almost all of the grants had a unifying theme, but the available funding limited grants to support for equipment, maintenance, and support staff, with some support for the lead principal investigators, and provided few or no funds for graduate students, postdocs, or faculty salaries.

I attempted to create a CER community based upon the DARPA model. We held a two-day CER principal investigator (PI) meeting⁹³ in February 1984 where the PIs presented their research in a series of focused sessions. Even though large, integrated projects usually were not the primary focus, the new state-of-the-art facilities resulted in many significant research projects. Jack Schwartz’s 1983 taxonomy of parallel computers⁹⁴ included several that were developed or extended under CER grants. These included the NYU Ultracomputer,⁹⁵ the Illinois CEDAR machine,⁹⁶ the Texas Reconfigurable Array Computer (TRAC),⁹⁷ the Berkeley Hypertree (also at Wisconsin),⁹⁸ the Utah Applicative Multi-Processing System,⁹⁹ the Wisconsin GAMMA database machine,¹⁰⁰ the Maryland ZMOB,¹⁰¹ Yale’s ELI-512 computer,¹⁰² the Duke Boolean Vector Machine,¹⁰³ and the Blue CHip Project¹⁰⁴ at Washington (begun as the Purdue Configurable, Highly Parallel (CHiP) family). The Eden Project¹⁰⁵ expanded on ideas from the efforts at Xerox PARC, SRI, and other industry labs, and developed an influential operating system. In a similar direction, the Crystal Project¹⁰⁶ at Wisconsin developed a shared multicomputer. The Cornell CER started a long career by Ken Birman¹⁰⁷ in distributed operating systems, which included the Isis Toolkit, the Horus system, the Ensemble system, and currently Isis2, Gradient, and the reliable TCP solutions. At Cornell, Bob Constable worked with Birman on Horus and Ensemble and developed a program development system called PRL (“pearl”)¹⁰⁸ that provides automated assistance with explaining and proving. There are many additional examples.

Several CER grants became the basis for early Science and Technology Centers relating to computing: the $38 million Science and Technology Center for Research on Parallel Computation at Rice University with the California Institute of Technology, Syracuse University, the University of Tennessee, Argonne National Laboratory, and Los Alamos National Laboratory (NSF 9120008); the $21 million Center for Research in Cognitive Science at the University of Pennsylvania (NSF 8920230); and the $35 million Science and Technology Research Center in Computer Graphics and Scientific Visualization at the University of Utah with Cornell University, Brown University, the University of North Carolina, and the California Institute of Technology (NSF 89202191). The more recent Team for Research in Ubiquitous Secure Technology (TRUST) at Berkeley, with Carnegie Mellon, Cornell, Mills College, San Jose State, Smith College, Stanford, and Vanderbilt (NSF 0424422) can trace some of its activities back to research that came out of the Cornell CER some 20 years earlier.

There was concern early in the CER program that these investments were severely limiting the funds available for regular grants. Jim Ortega at the May 29, 1981, CS Advisory Committee requested a review of the impacts of CER and CSNET on Standard Research Project Support (SRPS). While the Computer Science Section budget had increased 24% from FY 1980 to FY 1982, SRPS support had only increased 10.4%. In comparison, the Mathematical and Physical Sciences budget increased 19%. After substantial discussion, the advisory committee concluded that “the CER and CSNET are essential to the furtherance of computer science research and that it is too early to modify the direction being taken.”¹⁰⁹

In 1982, the DoD planned to expand its agencies’ support to include as many as 10 or 15 institutions. This DoD program never materialized, but DARPA upgraded facilities for its major contractors and expanded its smaller ($250–300,000) equipment contracts. ONR was able to provide a few Special Research Opportunities contracts in computer research with some facilities support. Without the planned DoD programs, the CER program grew in an attempt to fill the need.¹¹⁰ Through 1985, NSF had committed $49.89 million to 22 institutions for experimental computer research. In addition, DARPA had major contracts with MIT, Stanford, Carnegie Mellon, and California-Berkeley, which had supported experimental computer research. When NSF began the CER activity, it expected to support approximately 15 institutions. With more than 70 Ph.D.-granting departments of computer science and engineering, it was estimated that 25 to 30 would require research facilities of the magnitude provided by the CER program.

A report in 1986¹¹¹ noted that “universities have been funding CS growth at rates significantly higher than in any other major discipline. But national funding policy has favored the growth of basic research in CS at a rate no greater than that of other scientific, mathematical, and engineering disciplines.” The authors warned that “the late 1980s will witness the departure of our best and most mobile computer scientists and graduate students for industrial careers. Inevitably, the universities will be unable to maintain the centers of academic excellence in CS that have been so carefully developed during the past five years.” A year later, “[t]here has been a dramatic increase in federal funding for both total and academic CS research between FY 1976 and FY 1987 . . . Funding has shifted away from basic and toward applied research, both in CS federal funding as a whole and within academic CS.”¹¹² NSF convened an Infrastructure Workshop in July 1991. The workshop report¹¹³ placed a high priority on maintaining the Institutional Infrastructure programs at $20 million per year. It also proposed developing a matching program of $8 million to support facilities for individual and small group grants.

Recognizing that no more than 30 computer science departments¹¹⁴ would have enough experimental computer scientists to require CER-scale funding, a new Institutional Infrastructure program was announced with both “Large-Scale” (II-LS) and “Small-Scale” (II-SS) grants. Given the shortcoming of DoD funding, described above, CISE invested in, expanded, and replenished the experimental facilities at around 30 institutions. The II-LS program essentially replaced CER. II-SS was aimed at units with fewer experimental computer scientists and a reduced need for facilities support. Figure 2.3 shows both “large” (CER and II-LS) and “small” (II-SS).

The CISE Institutional Infrastructure program continued until 1993, when it was replaced by the CISE Research Infrastructure (RI) program. The RI program had institutional, instrumentation, and “shared” facilities, such as the CISE Advanced Distributed Resources for Experiments (CADRE). Figure 2.3 shows that 60 institutions benefited from the CER, II-LS, II-SS, and RI—many receiving three or more awards. In 1989, CISE introduced a facilities program directed toward minority-serving institutions (see Figure 2.4). One of those awards to University of Texas at El Paso became the basis for the CISE BPC CAHSI (Computing Alliance of Hispanic-Serving Institutions) Alliance, and in turn the CAHSI INCLUDES project, one of the first five $10 million NSF INCLUDES Alliances.

When I returned to NSF in 2000 as Division Director for Experimental and Integrative Activities, we moved the Minority Institutional Infrastructure to the education and workforce block of programs and began to redesign the remaining infrastructure programs of CISE. Today, CISE supports a Community Research Infrastructure (CCRI) program to encourage “discovery and learning in the core CISE disciplines . . . by funding the creation and enhancement of world-class research infrastructure. This research infrastructure will specifically support diverse communities of CISE researchers pursuing focused research agendas in computer and information science and engineering.”¹¹⁵

Figure 2.3Infrastructure awards FY 1980–1998.

Figure 2.4Minority institutional infrastructure awards FY 1989–1998.

2.10CSNET

In parallel with the “crisis” in experimental computer science, a number of researchers at leading universities did not have access to the ARPANET.¹¹⁶ Curtis and Pasta’s strategy for the Coordinated Experimental Research program included a computer network for research. Lawrence Landweber invited a number of researchers and government representatives to the University of Wisconsin–Madison in May 1979. His goal was to “discuss how computer network services like those of ARPANET could become available to the entire community of computer science researchers.”¹¹⁷ The attendees included Kent Curtis, Bob Kahn, and individuals who had experience with Theorynet and other similar “mailbox” systems hosted on commercial networks.¹¹⁸ The participants agreed that ARPANET’s mail, file transfer, and remote login services had “enhanced research productivity and had generated a strong community spirit among computer science and engineering departments that hosted ARPANET sites.”¹¹⁹

A consortium of universities including Georgia Tech, Minnesota, New Mexico, Oklahoma, Purdue, UC-Berkeley, Utah, Virginia, Washington, Wisconsin, and Yale submitted a proposal in November 1979 for a “CSNET” that would create a separate and independent network to provide ARPANET-like services to all U.S. computer science departments. Given the cost of duplicating the ARPANET infrastructure (estimated at $100,000 per institution), the proposed network would be built on commercial X.25 networks such as Compuserve, Tymnet, and Telenet. NSF declined the proposal in March 1980. Reviewers felt that the proposers were reinventing the ARPANET and not extending it, that they lacked a strong project management plan, and that for NSF to pay for the network it would have to reduce research support.¹²⁰ The reviewers’ skepticism was not unlike the reaction I had heard during CER site visits in 1980–1981 when asking proposing PIs about the CSNET plans. Many of those outside the CSNET proposal development did not see a real justification for an ARPA-like network, and some not even the need for email.

The NSF offered to fund a thorough study of CSNET. Landweber organized a meeting in Berkeley on June 15, 1980, at which DARPA announced its support for CSNET and assigned Vinton Cerf to help develop a plan to connect CSNET and the ARPANET. Landweber convened a 19-person CSNET planning committee, including Cerf and others who had extensive computer networking experience. The group worked throughout the summer of 1980 to devise an implementation strategy. The outcome was a plan to design CSNET as a network on multiple communication platforms interconnected via an Internet protocol. DARPA was moving from NCP to TCP/IP and the MMDF-based¹²¹ Phonenet system had been developed by David Farber and David Crocker at the University of Delaware. Phonenet was a low-cost mail relay system similar to the UUCP-based mail relay developed by Bell Laboratories to connect computer science departments that had Unix platforms. The UUCP protocol¹²² supported email and file transfer, but required explicit addressing and, unlike MMDF, was not compatible with IP networks at the time. The proposal would integrate ARPANET access, X.25 networks running TCP, and Phonenet to provide multiple tiers of services and costs for departments wishing to be connected to CSNET.

Landweber and colleagues from the University of Delaware, Purdue University, RAND Corporation, and the University of Wisconsin submitted a revised CSNET proposal to NSF in October 1980, and the National Science Board (NSB) approved the five-year proposal the following January. To address concerns about how CSNET would be managed required an unusual structure in which NSF itself, under Project Director C. William Kern, would directly manage the project for two years (through 1983) by means of contracts. NSF management would focus on setting up the organization to collect and disburse funds, and after two years the project would be sufficiently advanced that users would be willing to begin paying dues and fees.¹²³ Contracts were established with the University of Delaware, Purdue University, Rand Corporation, and the University of Wisconsin for CSNET development. Bolt, Beranek, and Newman (BBN) was contracted to run the CSNET Coordination and Information Center (CIC) for managing the network and distributing software.

Figure 2.5CSNET architecture 1981.

On March 6, 1981, NSF announced the establishment of CSNET, which would become a major step along the path to the Internet. On May 28, 1981, Bill Kern presented the status of the CSNET effort to the Computer Science Advisory Committee.¹²⁴ He discussed the two-year NSF management plan and the expectation that CSNET would become self-supporting in five years. He also told the Advisory Committee members that DARPA would develop the CSNET/ARPANET gateway and that software, systems, and services would target the Berkeley UNIX 4.3BSD operating system on VAX computers. He indicated that CSNET would initially comprise three subnets (Figure 2.5)—ARPANET, Telenet, and Phonenet—but would be designed to support expansion to other available networks. CSNET initially provided the same services as ARPANET: mail, file transfer, remote login, and an on-line name server. CSNET’s $5 million project budget, limited staffing, and the five-year timeframe for self-sufficiency put significant pressure on the CSNET team.

Figure 2.6CSNET map 1983.

In just six months, CSNET was operating,¹²⁵ including a Phonenet site at NSF in the Computer Science Section, the first NSF Internet connection. In addition to NSF, Phonenet sites included Cornell, FCC-NET, HP Labs, Purdue, Princeton, UC Irvine, and Delaware, with plans to expand to New Mexico Tech, Pennsylvania, Georgia Tech, Duke/UNC, Fairchild, and Maryland-College Park.

When Kern stepped down as CSNET Project Manager in October 1982, I assumed the role of CSNET Project Director with Landweber as Chair; Peter Denning, Richard Edmiston, David Farber, Anthony Hearn, Kern, and me as members of the Management Committee. By the time of the first CSNET Newsletter,¹²⁶ 56 Phonenet sites were operational and 27 were nearing operation (see Figure 2.6). These connected through the two CSNET relays on the ARPANET at RAND Corporation and the University of Delaware. CSNET was beginning to meet with European network leaders to investigate international connections.

After its two-year management, the NSF selected the University Consortium for Atmospheric Research (UCAR)¹²⁷ on May 3, 1983, to host and manage CSNET, with Leonard Romney (UCAR executive director) as PI and a member of the CSNET Management Committee.¹²⁸ As UCAR assumed control of CSNET, the Management Committee was replaced by a larger Executive Committee¹²⁹ with Peter Denning as chair, representing the computing research community; and the operation of the CSNET relays and technical services moved entirely to BBN. BBN housed the CSNET Coordination and Information Center (CIC) to provide operational management of CSNET. In 1983, CIC staff included: Dr. Richard Edmiston (CIC Director); Laura Breeden (CIC User Liaison); Dan Long (CIC Technical Liaison); and Beth Johnson (CIC Staff Assistant). Leonard Romney left UCAR in May 1984 and was replaced by Stanley Ruttenberg.

Figure 2.7CSNET executive committee 1983.

By October 1983, Lawrence Landweber was leading an effort to create gateways and connections among BITNET, and Canadian and European networks, including SERCNET (United Kingdom), SUNNET (Sweden), CERNEY (Switzerland), and UNINET (Norway). CSNET connected to BITNET through a University of Wisconsin gateway. At the time, BITNET was a fast-growing network connecting university computing centers via IBM store-and-forward software and leased lines. Connecting to international and other U.S.-based networks raised issues about how to manage the costs associated with traffic transiting multiple networks. In 1983, the initial agreement called for each network to bear the costs of message traffic into other networks. Security issues also arose concerning international traffic in and out of ARPANET via BITNET and CSNET.

In June 1984, I described new NSF networking plans (see NSFNET below) to the CSNET Executive Committee and asked them how CSNET might interact with this expanded vision. CSNET established new gateways with SUNET (Sweden), the Israeli Network, and DFN (Germany). NSF paid for CSNET dues for undergraduate institutions. Dennis Jennings, then chairman of the European Academic Research Network (EARN), visited the CSNET Executive Committee in September 1984. He would soon be recruited to NSF, replacing me as the program director for Networking in the Office of Advanced Scientific Computing (OASC).

The last NSF payment for CSNET operations was in mid-1985. By 1986, CSNET connected more than 165 university, industrial, and government computer research groups serving more than 50,000 researchers and students, including accounts for 1000 Internet hosts. Network services were operational and numerous networks outside the U.S. were connected.¹³⁰ CSNET was self-supporting and received significant industry funding. CSNET clearly demonstrated, for the first time, that users were willing to pay for network services.

CSNET actively collaborated with colleagues in other countries, supporting and often enabling the international expansion of the Internet. CSNET had mail connection via CSNET/Internet and USENET/EUNET/UUCPNet connections to foreign affiliates and their gateways. These included: CDNNET (Canadian Academic Network, via the University of British Columbia); SDN (System Development Network, with a gateway at the Korea Advanced Institute of Science and Technology); SUNET (Swedish University Network, via Chambers University of Technology); CHUNET (Swiss University Network, via ETH-Zentrum); INRIA (French University Network, through INRIA/Rocquencourt); DFN (Deutches Forschungsnetz); JUNET (Japanese University Network, through the University of Tokyo); Finnish University Network (via Helsinki University); AC.UK (Academic Community, United Kingdom, via University College, London); ACSNET (via a UUCP-based connection at the University of Melbourne); New Zealand Academic Network (via Waikato University, Hamilton); and the Israeli Academic Network (via Hebrew University of Jerusalem).

At its meeting in Ann Arbor in June 1988, the CSNET Executive Committee discussed a potential merger of CSNET¹³¹ and BITNET. As vice chair of the Executive Committee, I was assigned to the CSNET-BITNET merger team, planning a merged network called “ONENET.”¹³² Eventually, in 1989, CSNET and BITNET were brought under the Corporation for Research and Educational Networking (CREN), a non-profit corporation initially composed of the organizations that had participated in BITNET and CSNET. NSF funded the expansion of CSNET and BITNET, as well as the development of TCP/IP services as adjuncts to NSFNET. Because of the success of NSFNET and the regionals, CREN discontinued CSNET services in 1991. CREN ended their support for BITNET in 1996, due to the growth of TCP/IP-based networks, and by 2003, CREN dissolved itself.

2.11The Office of Advanced Scientific Computing and NSFNET

Beyond a brief overview of the high-performance computing programs, the details of which are covered in Chapter 10, this subsection examines the developments that led to the NSFNET. Chapter 9 provides details on NSF’s broader role in networking before, during, and after the NSFNET project.

The Lax Report¹³³ identified two problems: “important segments of the research and defense communities lack effective access to supercomputers and students are neither familiar with their special capabilities nor trained in their use”; and “the capacity of today’s supercomputers is several orders of magnitude too small for problems of current urgency in science, engineering and technology.” The panel recommended a program that would increase access via high bandwidth networks; increase research on computation, software, and algorithms; train personnel; and increase R&D on new supercomputer systems.

In response to the Lax Report, NSF organized an internal working group¹³⁴ in April 1983 to help the Foundation meet the computing needs of academic science and engineering. NSF also held a workshop in May 1983 with 13 scientists from diverse disciplines to define an initiative in large-scale computing and networking. According to what became known as the Bardon-Curtis Report,¹³⁵ NSF should: (1) coordinate with other federal agencies; (2) increase support for local computing facilities; (3) encourage proposals to provide supercomputer services and access and be prepared to support 10 supercomputer systems within three years; (4) support networks linking laboratory researchers with each other and with supercomputer centers to provide access, file transfer, and scientific communication; (5) support academic research in advanced computer systems design and computational mathematics; and (6) establish an NSF advisory committee for supercomputing.

In November 1985, the House Committee on Science and Technology held hearings¹³⁶ on supercomputer and network resources for science research. During the hearings, NSF Director Edward Knapp cited the Bardon-Curtis Report and indicated that Edward Hayes, the NSF Controller, was chairing an NSF Task Force on Advanced Scientific Computing. Knapp also indicated that NSF was gathering information from grantees about their immediate needs for access to Class IV¹³⁷ computers and would negotiate with suppliers who could provide appropriate blocks of time. Under this plan, NSF would continue to support research in the theoretical and experimental design of computers as well as on computational mathematics, software, and algorithms. NSF indicated that its networking initiative would be part of Advanced Scientific Computing. Subsequently, the 98th Congress voted $40 million to fund the recommendations of the Bardon-Curtis Report.

NSF established an Advisory Committee on Supercomputer Access¹³⁸ chaired by Neal Lane, then Chancellor of the University of Colorado at Colorado Springs, that would become the Advisory Committee for the Office of Advanced Scientific Computing. I staffed a subcommittee¹³⁹ on networking options, chaired by Joe Wyatt, the Chancellor at Vanderbilt. At that time, NSF expected that a supercomputing network would be developed in two phases: Phase I using conventional network technology and expanding existing viable networks, and a Phase II using satellite transponder facilities and optical fiber trunks as they became available.

In December 1983, Landweber wrote to Edward Knapp encouraging him to “proceed as quickly as possible to establish a national Science Net [and] to use existing technologies . . . such as ARPANET, CSNET, and BITNET.”¹⁴⁰ A few weeks later, Jack Schwartz (NYU) wrote¹⁴¹ to Edward Hayes asking him to have the advisory committee consider other needs for high-bandwidth communication beyond access to the supercomputing centers. Even before the Office for Advanced Scientific Computing was established, the community, in particular Landweber and Kenneth Wilson, were pushing for a national network.

As the CER and CSNET programs grew, Kent Curtis and I had been discussing ARPANET opportunities with DARPA. In mid 1983, Curtis asked Frank Kuo of SRI for advice on expanding or duplicating ARPANET technology to support supercomputer access¹⁴² in a network called “USERNET” that might support 200 academic research institutions or 2000 college and university sites. Kuo pointed out that splitting off MILNET from ARPANET would leave only a 40-node network intended to be a “research and development” testbed. He estimated that developing a 200-site USERNET using ARPANET technology would require $7.5–11.5 million for IMP¹⁴³ hardware and $17 million in operational costs and 10 times that much for 2000 institutions. He also raised the issue of NSF competition with commercial packet networks such as TELENET, TYMNET, UNINET, or NET1000, and suggested that NSF look instead into using commercial networks for a backbone. My thought at the time was that no commercial network supported full network services and there might be alternative “tiered” approaches.

In April 1984, Kent Curtis and I met with DARPA’s Bob Kahn, who said that the ARPANET, as an R&D network, could only expand by an additional 20 nodes and for $1 million in capital costs and $4.8 million in annual costs. At that time, the Division of Computing Research was already supporting ARPANET sites at RAND and Delaware for CSNET and at a few CER sites.

NSF created the Office of Advanced Scientific Computing in May 1984, with John Connolly from the Division of Materials Research as Director, Larry Lee from Mathematics as Program Director for Centers, and me (on loan from the Division of Computing Research) as Program Director for Networking. The first awards for time on Class IV supercomputers totaling almost $19 million were made July 1, 1984, to Purdue University, University of Minnesota, and Boeing Computer Services. In 1985, OASC expanded access to existing supercomputer resources, adding centers at AT&T Bell Labs, Colorado State University, and Digital Productions (an early computer animation company).

An NSF OASC review panel¹⁴⁴ met in February 1985 to consider 22 applicants. On February 25, 1985, NSF announced¹⁴⁵ funding for four National Advanced Scientific Computing Centers: the John von Neumann Center (JVNC) at Princeton University, the San Diego Supercomputer Center (SDSC) at the University of California, San Diego, (managed initially by General Atomic), the National Center for Supercomputing Applications (NCSA) at the University of Illinois, and the Cornell Theory Center. Later, NSF named the Pittsburgh Supercomputing Center (PSC) as a fifth center. (The National Center for Atmospheric Research (NCAR) was sometimes considered a sixth center, but was always dedicated to climate researchers and never a part of the program.) Each of these centers was associated with academic and industry partners and had developed a tentative “customer” base of scientists needing access to high performance computing.

With responsibility for the centers and a network under OASC, the OASC Networking Advisory Committee recommended the establishment of a “Sciencenet Phase 1”¹⁴⁶ using available and proven technology to implement a network as soon as possible. The preferred strategy was to expand and interconnect existing networks such as ARPANET and BITNET with selective use of commercial network services. By 1985, the Defense Communications Agency had begun to use ARPANET as an operational DoD network following the cancellation in 1983 of the new command, control, communications, and intelligence (C3I) network, AUTODIN II. DoD was looking into splitting off the “research” sites and using the ARPANET (as MILNET) only for military purposes. This action complicated any approach to leveraging ARPANET for supercomputer access and eventually accelerated the growth of NSFNET.

The planned Sciencenet Phase 1 effort involved the development of Internet protocols, access protocols, and a management strategy for the network. David Farber and Landweber defined a Phase I strategy¹⁴⁷ to quickly enable users of existing networks (ARPANET, BITNET, CSNET, MAILNET, MFENET) to run jobs on supercomputers at the national centers. ARPANET, CSNET/X.25, and MFENET users could remotely log in to supercomputers and run interactive or batch services. BITNET, CSNET/Phonenet, and MAILNET users would have to depend on electronic mail or file transfer for batch submissions. The Landweber-Farber Report also recommended that NSF should (1) add a Sciencenet manager and management team (or contract for such services); (2) establish a working group representing the centers, networks, and NSF management; and (3) establish a permanent Technical Advisory Committee (TAC). Because someone had trademarked “Sciencenet,” the network quickly became known as NSFNET.

After I left for my IR&D assignment to Berkeley in August 1984, I was still involved remotely with both CSNET and NSFNET. NSF was looking for a replacement in OASC and concurrently UCAR was looking for a permanent CSNET Executive Director. Landweber had met Dennis Jennings, the Director of the BITNET-based European Academic Research Network (EARN) and Computing Center Director at University College Dublin. He encouraged him to apply for the CSNET directorship. Jennings visited NSF and spoke with John Connolly in August 1984. Offered both the CSNET and NSFNET positions, Jennings accepted the NSFNET directorship and began in January 1985. As he recalled, “So when I arrived at the NSF on January 2nd, 1985, the key components were in place: The demand from key researchers; a significant budget for networking—roughly 10% of the supercomputer program budget was devoted to the network; and the CSNET experience that provided the confidence in the internetworking concept and technology. What was required was a Catalyst—and that was my role.”¹⁴⁸

Jennings identified several key decisions made under his leadership.¹⁴⁹ The first was to develop a general-purpose network for all science and engineering research rather than a network only providing supercomputer access. There was considerable disagreement on this issue between the OASC networking subcommittee and John Connolly—and to some extent the centers. Connolly was reluctant to separate the network development from the centers, but Gordon Bell eventually split the networking program off as a separate division in the Computer and Information Science and Engineering (CISE) Directorate, as described below.

Another important decision was to adopt a “network of networks” approach. CSNET employed a network of networks approach by integrating Phonenet dial-up services, public network X.25 services, and ARPANET; and ARPANET, as it transitioned to an R&D network, was also integrating networks with quite different communications layers: satellite, phone lines, etc. After a visit to the Cornell Theory Center, Jennings met with Richard Mandelbaum, who was then working with Cornell and other New York state universities, corporations, and research laboratories to develop a statewide network, NYSERnet. Jennings provided some seed funding to NYSERnet, and later to other regional networks including SURAnet, BARRnet, MIDnet, Westnet, Merit, NorthWestNet, and NEARnet. The network of networks model evolved from supporting networks¹⁵⁰ with differing transport and physical layers and a common Internet layer (such as in CSNET) to also support “tiered” networks that included campus Local Area Networks (LANs), regional networks, and a national backbone. Similarly, the NSFNET program funded the center-based SDSC and the JVNC networks.

As an interim arrangement in October 1985, NSF and DARPA¹⁵¹ agreed that ARPANET could be used to access the centers hosted on ARPANET (Illinois, Cornell, Minnesota, and Purdue). This agreement opened up ARPANET hosts, typically servers in computer science and engineering departments, to a broader set of users via campus-wide networks. The NSFNET program also funded CSNET and BITNET to develop advanced TCP/IP services to provide similar access.

In September 1985, NSF announced its intention to implement a national backbone linking the five NSF supercomputer centers and NCAR, with connections to regional and campus networks. There was some initial pushback from the centers, concerned that they could lose customers in moving from “star networks” and proprietary protocols to a broadly accessible national network with common protocols. A related decision was the selection of Dave Mill’s “fuzz-ball” PDP-11-based routers due to the high cost of ARPANET Interface Message Processors and the lack of commercial alternatives.

The decision for which Jennings may be best known is the adoption of the DoD TCP/IP and related ARPANET protocols as the standard for NSFNET. NSF had originally intended to use the International Standards Organization (ISO) Open Systems Interconnection (OSI) protocols, but they were not yet widely available.¹⁵² The scientific communities planning to use the centers had developed preferences for protocols used by specific disciplines: MFENET by the magnetic fusion energy community, DECNET by the high energy physics community, etc. TCP/IP was mostly available on Unix systems and not on the Cray Time-Sharing System (CTSS) that was running at many of the centers.

Jennings left NSF at the end of March 1986, having developed a model for NSFNET and moving it forward. He had established a Networking Technology Advisory Group (NTAG) and put a staff in place. Following a brief stint as acting president at the John von Neumann Center, he returned to Ireland and was replaced at NSF by Steven Wolff. Wolff had met Dave Farber when DARPA was arranging an ARPANET connection for the Delaware CSNET Relay. Wolff was working at the Army Ballistic Research Labs (BRL) located in the Aberdeen Proving Ground in Maryland, and BRL had provided the ARPANET line connecting the University of Delaware for the CSNET relay. Farber convinced Wolff to join NSF on a detail from BRL as NSFNET program director. Wolff brought substantial experience to the position, having served on the faculty of Johns Hopkins after receiving a Ph.D. from Princeton. His experience at BRL had included work on TCP/IP. He became Division Director for the Networking and Communications Research and Infrastructure when Gordon Bell split it off from the Division of Advanced Scientific Computing in April 1985—April Fool’s Day as Wolff¹⁵³ remembers it. He was responsible for much of the development, expansion, and eventual privatization of NSFNET. Details are in Chapter 9.

2.12The Beginning of CISE

In 1986, the NSF programs and offices supporting computing and information research and applications were brought together for the first time since NSF was founded. The new Directorate for Computer and Information Science and Engineering (CISE) would become the organizational core NSF used to exert federal leadership in computing.

At the request of Richard Nicolson, AD/MPS, in April 1985 Kent Curtis carried out an analysis of the options for organizing computing programs within NSF. Curtis considered the NSF Office of Information Services (OIS), the IT support organization led by Connie McLindon. He concluded that, while OIS was funding some projects such as EXPRES and working with the networking program, it “had no primary research role” and should remain an administrative unit. In his analysis, he looked at programs funding “informatics” viewed broadly. These included the Computer Engineering program, the Division of Computer Research, the Division of Information Science and Technology, the Office of Advanced Scientific Computing, and various elements of Materials Research, Mathematical Sciences, Electrical and Computer Engineering and Behavioral and Neural Sciences. Curtis considered various combinations of these programs and even a new directorate encompassing Mathematical Sciences, Cognitive Science, Linguistics, Systems Engineering, and Management Science. There were “substantial benefits and faults to be expected from any decision” he noted, and added that “the Director should feel free to follow his instincts because there is no obvious wisdom to suggest a particular course.”¹⁵⁴

In the late fall of 1985 after I returned to Washington, I began to work with Chuck Brownstein on Bloch’s plans to consolidate NSF computing activities. Bloch officially announced his intentions to create a new directorate and hire Gordon Bell to run it on March 3, 1986. Bell had already begun consulting with Bloch and Mary Clutter. In February, Bell requested that Brownstein take on the role of Executive Officer of the new directorate, Jerry Daen be added as Planning and Administrative Officer, and I join half-time on loan from DCR.

Albert Bridgewater, the Deputy Assistant Director for Astronomical, Atmospheric, Earth, and Ocean Sciences (AAEO), wrote a memorandum¹⁵⁵ to Bell in February 1986 suggesting a process and schedule for a new Computer and Information Science and Engineering (CISE) Directorate. This process included meetings with the National Science Board (NSB), developing long-range plans, and presentations to Bloch and the NSB in the spring of 1986. Bell, Brownstein, Daen, and I began to develop a strategy to address the deliverables outlined in the Bridgewater memorandum.

In a second memorandum,¹⁵⁶ Bridgewater encouraged Bell to be a proactive Assistant Director Designate by: Identifying areas needing greater emphasis or support; organizing community support; organizing National Academy studies; encouraging links and cooperation with other agencies and directorates; and keeping the National Science Board, the Office of Management and Budget, the Office of Science and Technology Policy, and NSF management informed. Bridgewater was really telling Bell that, with Bloch’s backing and his national credentials, he had an opportunity that had not been given to the NSF computing program leaders in the past.

In Bell’s typed and hand-annotated notes of February 26, 1986, he began to sketch out ideas for CISE:

CISE encompasses fields that are predominately concerned (measured either in design effort or system cost) with the understanding (computer science) and design of computers (computer engineering). These SYSTEMS include: traditional and specialized computers, all forms of computer and communications networks, various transducer interfaces for computers and robots, specialized signal processors, and VLSI circuits and their design systems to implement the particular information processing system.

CISE is not concerned with the phenomena or processes necessary to implement the above systems . . . although it is concerned with the design and implementation of the large systems that integrate and carry out complex, manufacturing processes.

CISE supplies supercomputer resources and network access to programs in all directorates. CISE will initiate programs to facilitate more effective use of supercomputers, including: understanding vector multiprocessors, improved algorithms, software development faster networks and high speed graphics workstations for more effective and enhanced use.¹⁵⁷

Bell decided to argue that CISE encompasses all fields in which the major fraction of the intellectual discipline is computer science or engineering (e.g., robotics, VLSI, signal processing), and that other disciplines would have a “non-trivial” portion of their budget devoted to computing as an “experimental apparatus” and would be responsible for their own applications and for utilizing the computer as a simple component. CISE would “provide the scientific and engineering knowledge for these fields.” Bell wondered if educational activities, including supercomputing training and multidisciplinary projects, should be included.

In a memorandum dated February 27, 1986, to Bloch, Clutter, and Engineering AD Nam Suh, Bell proposed that CISE should include the Computer Research Division (from Mathematical and Physical Sciences), the Information Sciences and Technology Division (from Biological and Behavioral Sciences), the Office of Advanced Scientific Computing, including NSFNET, programs in real time computing applied to signals and communications systems, image understanding, and systems theory (from the Engineering Science in Electrical Communications and Systems Engineering Division), Computer Engineering and Manufacturing Engineering for Computers and Semiconductors (from the Design, Manufacturing and Computer Engineering Division), the Columbia University Engineering Research Center, the Advanced Technology program (from Science and Engineering Education), and the EXPRES Project.¹⁵⁸

Alarmed by Bell’s wide-ranging vision, Nam Suh responded in a memorandum to Bell (copied to Bloch and Clutter) dated February 28, that “the only thing that really deals with the essence of Computer and Information Sciences and Engineering that you ought to take into your new directorate is Computer Engineering. In the rest of the programs, the computer is a peripheral tool, but not the intellectual driving force behind them. You will find that this view is widely supported in engineering schools throughout the country.” Suh added, “Sometimes we have the feeling that this world evolves around computers [but the] role of computers in our society has got to be looked at in the proper context.”¹⁵⁹

Bloch issued a memorandum,¹⁶⁰ dated March 3, 1986, to all NSF staff indicating that he was officially appointing Bell as a consultant to assist him in reorganizing the NSF computing activities with the intention of naming him AD/CISE. Bloch intended to “consolidate into a new directorate several computer-related divisions and programs [including] the Division of Computer Research (MPS); the Division of Information Science and Technology (BBS); the Office of Advanced Scientific Computing (O/D); and certain engineering programs from ENG.” In an attachment, Bloch stated the following rationale—that creating CISE:

(1) Brings together ongoing activities now spread among several NSF units; (2) Simplifies formulation and coordination of new policy directions; (3) Makes it possible to deal easily with full span of functions, from basic research through systems engineering in an area critical to national well-being; (4) Facilitates internal management; takes program activities out of Director’s office and puts them into a technical area; and (5) Will [create a] small disciplinary research directorate—in the range of $110–130 million, [with an] approximately 50 person staff drawn largely from other parts of NSF.

The allocations of budget and personnel attached to Bloch’s memorandum are shown in Table 2.1. The budget increase from the FY 1986 current plan to the FY 1987 Estimate is $21.89 million (19.5%), but with almost half of the increase ($10.35 million) going to OASC.

Table 2.1Bloch’s initial allocation to CISE (in $ millions)

a. Staffing did not include the approximate 6 for the AD Office.

b. Portions of DMCE and ECSE yet to be determined, with the transfer amounts estimated in the $10–30 million range. Amounts shown are mid-range estimates.

During March 1986, many people in the Engineering Directorate became alarmed about the potential scope and definition of CISE. Nam Suh was not happy with Bloch’s initial decision and mobilized¹⁶¹ members of his Advisory Committee (Frederick Garry, Sheila Widnall, Lester A. Gerhardt, Paul C. Jennings, and Herbert H. Richardson) at NSF on March 18, 1986. Meeting attendees also included Suh’s Task Group on Computing (Herbert Voelcker chair, with program directors Alan de Pennington, John Mayer, Howard Moraff, Michael Gaus, Michael Polis, Elias Schutzman, and Donald Silversmith) and Frank C. Huband, Division Director of Electrical, Communications, and Systems Engineering (ECSE). Voelcker was also the Deputy Division Director for Design, Manufacturing, and Computer Engineering (DMCE). ECSE and DMCE were the Engineering divisions most likely to be impacted.

The Engineering Advisory Committee had access to Voelcker’s Task Group report,¹⁶² which considered a “broad” and a “narrow” construct for CISE; but the draft report failed to get the full support of the Task Group. The Task Group also analyzed the DMCE Computer Engineering program in a report¹⁶³ concluding that many elements of the Computer Engineering program had stronger connections to the Engineering Directorate programs than to the CISE programs. The committee members were particularly concerned about a home for the joint DARPA-NSF MOSIS VLSI fabrication facility, which became a key activity of the MIPS Division in CISE. Norman Caplan, the Deputy Division Director of ECSE, forwarded to the attendees his memorandum¹⁶⁴ to Nam Suh concerning robotics, which raised concerns about the definition of the field of robotics being assigned to CISE.

Following the March 18th meeting of the Engineering Advisory Committee, Frank Garry, its chair, wrote to Bloch that the Committee concurred with “the consolidation of the following into the new CISE Directorate: DCR/MPS, IST/BBS, OASC/OD and the Program in Computer Engineering from the Engineering Directorate.” Garry went on to say that “the remaining Engineering programs outlined in Gordon Bell’s memorandum of February 27 have their intellectual base in the Engineering Directorate” and we “fear that their transfer to CISE would narrow their focus and eventually erode their disciplinary strength.” The recent reorganization of Engineering had been carefully constructed and based on broad input from the Advisory Committee, the National Academy of Engineering, and other members of the engineering community. It “would be precipitous to alter the [d]irectorate’s programs in a major way without a similar review.”¹⁶⁵

Other Engineering Directorate managers also pushed back against Bell. Frank Huband stated, “The creation of a computer-related directorate is an exciting event, and has the potential to create new opportunities for development in this important discipline” but research funded in CISE “must pass muster as computer-related” . . . the practitioners in non-computer disciplines “want—and I believe deserve—an independent home for their research proposals.”¹⁶⁶ “Assignment of program elements to ENG or CISE should be governed by considerations of their fit to the respective missions of the directorates, and of their relative contributions to strengthening U.S. research in computing or in engineering,”¹⁶⁷ suggested program director Howard Moraff. He also raised concerns about possible disruptions to the new programs created in Nam Suh’s recent reorganization of Engineering and urged collaboration between CISE and Engineering.

Bell wrote to Bloch defending his approach to organizing the directorate, suggesting:

The rationale for Engineering disciplines in CISE is that contemporary engineering naturally divides into two parts: those research areas based on the physical and biological sciences and those which deal with information. This provides a working boundary: on one side are research areas governed by the physical transformations, and on the other side those research areas concerned with the transformation of information from one form to another.¹⁶⁸

In Bell’s personal notes dated March 20, 1986, he characterized the Nam Suh position as being that CISE should be “pure computer,” while Bell himself thought it should be “pure information processing (and storage, transmission, switching, transduction), robotics, and some cross-disciplinary programs that make significant use of the computer.” In these notes, Bell discussed other organizations of traditional EE and computing, and the National Academy of Engineering taxonomies.

On March 24, Frank Huband responded to Nam Suh¹⁶⁹ concerning the probable decision by Bloch to transfer Communication Systems and Signal Processing programs (CSSP) to CISE. Huband was concerned that “important parts of the current CSSP program will not be relevant to the purposes of CISE and may thus not be eligible for future funding.” This memorandum was forwarded to Bloch and Bell, and Bell forwarded it to Chuck Brownstein, Bernie Chern (who had been reassigned from the Computer Engineering program to CISE), and me. Bell commented that “we will work it out over the next couple of months along with MOSIS funding, etc.”

Bell sent his initial plan¹⁷⁰ for CISE to Bloch for approval via the Assistant Director for Administration on April 17, 1986. This memorandum proposed transferring the Division of Computer Research (DCR) from Mathematical and Physical Sciences, the Division of Information Science and Technology (DIST) from Biological and Behavioral Sciences, and the Office of Advanced Scientific Computing (OASC) from the Director’s Office intact. It proposed creating a new Division of Computer and Information Engineering (CIE) to temporarily house the computer engineering programs (Software Systems Design; Computer Systems Architecture; Vision, Robotic, and Knowledge-based Systems) and the Communications and Signal Processing programs, each from the Engineering Directorate. The proposal added a Division Director (Bernard Chern) for the new CIE Division, a CISE Planning Officer (Jerry Daen), an acting CISE Executive Officer (Charles Brownstein, who remained Director/DIST), and an acting senior scientist for planning and program development (W. Richards Adrion, who remained Deputy Director/DCR). The new directorate would have 54 positions, 49 permanent and 5 IPAs. Bloch approved, and the directorate was officially launched on May 1, 1986. At the time it was officially created, CISE had two advisory committees: Computer Research (mainly associated with DCR) and Advanced Scientific Computing (mainly associated with OASC).

In Bell’s presentation¹⁷¹ to the National Science Board in May, he described using the CISE research budget as a “balance wheel” to DARPA and industry. He defined five research areas: parallelism as applied to parallel processing; automation, robotics, and intelligent systems; ultra-large-scale integrated systems; advanced scientific and engineering computing; and networks and distributed computing. Bell focused on these areas because they had relatively clear, long-term goals; measurable output; an emphasis on maintaining U.S. leadership in computing; significant economic and competitive impact; and a demand for undergraduate and graduate training. Across these five initiatives, CISE would support basic, front-end research throughout the entire computer research community at a time when DARPA was becoming more “mission oriented.”

Between April and August, a number of organizational structures were considered. Bell’s five initiatives helped structure the divisional organization of CISE. While the initiatives were cross-cutting, divisions were thought of as the leads: DCR for parallelism as applied to parallel processing; DIST for automation, robotics, and intelligent systems; CIE for ultra-large-scale integrated systems; and OASC for advanced scientific and engineering computing. Bell did not see OASC leading networks and distributed computing, and that eventually led to a fifth division in CISE. DIST leadership in automation, robotics, and intelligent systems led to a proposed ARIS division that eventually became Information, Robotics, and Intelligent Systems (IRIS) due to a continuing commitment to information sciences and systems. Initially CIE was to become the Ultra-Large-Scale Integration (ULSIS) Division; but with a responsibility for computing design and architecture, it was renamed the Microelectronic Information Processing Systems (MIPS) Division. The need to locate the communication and signal processing programs led to DCR taking on that responsibility as the Computer and Communications Research (CCR) Division. One other proposed restructuring would have divided OASC¹⁷² into three sections: Centers with Larry Lee as Head; Networking and Distributed Computing with Steve Wolff as Head: and a New Technologies Section with Al Harvey as Head. The plan included research and EXPRES program directors. The tension over NSFNET as a national vs. supercomputer network continued and John Connolly strongly objected.

On August 26, 1986, Bell proposed¹⁷³ restructuring CISE by reconfiguring the DCR, moving the Intelligent Systems program to DIST, renaming DCR as the Division of Computer and Computation Research, restructuring DIST and renaming it the Division of Information, Robotics, and Intelligent Systems, and restructuring CIE and naming it the Division of Microelectronic Information Processing Systems. The Office of Advanced Scientific Computing was renamed the Division of Advanced Scientific Computing. In the divisions, programs were restructured to reflect a new divisional mission. This plan was approved and became official¹⁷⁴ on October 17, 1986.

This reorganization resulted in some personnel changes. Earlier in July, Chuck Brownstein officially was named CISE Executive Officer. (He had served briefly as Acting AD/CISE until Bell was sworn in on June 17, 1986.) While I would resign officially on September 14,¹⁷⁵ remaining on a part-time basis through the fall, Gordon Bell already knew this when he wrote his August 26 memorandum. He officially transferred me into the position from DCR and argued to keep the position after my departure. Kent Curtis filled this position a year later. EXPRES was moved from DASC into the AD’s office. The last organizational change occurred on December 10, 1986, when the Networking and Communications programs in DASC became a fifth division in CISE: the Networking and Communications Research and Infrastructure Division (NCRI), with Steve Wolff as DD.

While the organizational debates were going on, the staff that were clearly moving to CISE began to address the planning process outlined by Bridgewater. I wrote¹⁷⁶ in March 1986 to Chuck Brownstein, Bernie Chern, John Connolly, and Kent Curtis about long range planning for CISE, asking them to produce planning documents covering both existing programs and potential new initiatives. This was a three-part request: an exercise to “define the base,” long-range strategic planning, and issue papers on new initiatives. All of these were due March 17.

I wrote¹⁷⁷ again to the (unofficial) CISE division directors about the need for them to develop plans to address the five Bell initiatives, asking for two-page position papers. Each acting division director needed to answer four questions Bloch had asked each directorate to address: What difference has NSF support made? What is the NSF role in [discipline] research? What are the programmatic gaps? What are your priorities in the event of reductions? Table 2.2 includes excerpts from the Long-Range Planning Material submitted to the Office of Budget, Audit, and Control¹⁷⁸ in April 1986.

Each of five research initiatives included research opportunities/breakthroughs needed, current efforts, and plans and initiatives. Two-page position papers were written on (1) parallelism (lead: Rick Adrion); (2) advanced scientific computing (lead: John Connolly): (3) networking (lead: John Connolly); (4) fabrication facilities expanding the MOSIS concept (lead: Bernie Chern); (5) robotics (lead: Chuck Brownstein); and (6) experimental systems (lead; Robert Minnick).

In July, Chuck Brownstein asked¹⁷⁹ the division directors to prepare backup materials for the FY 1988 budget request. CISE was requesting a $69.02 million increase to $192.00 million, a 56% increase over the FY 1987 current plan. Proposed initiatives included project and instrumentation support for research on parallel techniques of computing and information processing to be expanded throughout CISE; several large group or “mini-center” awards to be made to promote experimentation with large-scale systems; additional infrastructure for use throughout U.S. academic institutions, including upgrading instrumentation and improving laboratories; a major effort to be undertaken to expand university research and teaching in the design, fabrication, and use of integrated microelectronics; and a commitment to advancing and accelerating the state-of-the-art of advanced scientific computing.

More on the development of CISE is covered in the following chapters.

Table 2.2Answers to the “Four Questions” in the 1986 long-range planning exercise

What difference has NSF support made?	CISE programs have improved the knowledge base for research and commerce and have developed the national scientific and engineering personnel and facility infrastructure required for the maintenance of U.S. leadership. CISE also has a unique NSF role in the improvement of scientific and engineering computing and communications through shared use facilities, training, and network links among researchers.
What is the NSF role in Computer and Information Processing research?	NSF has unique responsibility for long-term and theoretical research and for the broad base of academic research. NSF is critical for the improvement of the national academic infrastructure in the CISE areas. Advanced Scientific Computing is a unique first step toward conditioning the general scientific research community to use computing as a new mode of research.
What are the Programmatic Gaps?	The academic base of computing and information processing research and the support base from both federal and private sources are too small. Too many research universities lack the necessary manpower and infrastructure in computing and information processing research to achieve the critical mass needed for quality research in this important technology.
What are your priorities in the event of reductions?	None of the major areas of CISE would be dropped; fewer awards would be made. Graduate student and young faculty support will have a high priority. Grant sizes will be maintained. Network access to the supercomputer centers will be given higher priority than maintaining all the centers. If necessary, we would reduce the CER program in favor of individual faculty research project support.

2.13Summary and Conclusions

In a mere 12 years, computing programs at NSF transitioned from two weakened offices, OCA and OSIS, to a directorate that had positioned itself to lead the major national initiatives described in later chapters. Along the way, a number of significant initiatives and activities fundamentally changed the perception of computing as a discipline. Not only did dozens of Turing Award winners begin their careers with NSF funding, but so did hundreds of ACM, IEEE, AAAI, and AAAS fellows. Theorynet led to CSNET and then to NSFNET. The Computer Science Research Equipment program laid the ground work for the Coordinated Experimental Research program. CER fundamentally altered the capacity for experimental research in colleges and universities.

Table 2.3NSF CISE FY 1988 budget request

Source: NSF OBAC 1986

I am fortunate to have had a career that spanned those 12 years and the opportunity to observe how far the field came during those years and to contribute to its growth. The narrative above names a number of important people, but it omits a great number of administrators, program managers, program assistants, and other staff who made the successes of the period possible.

Notes

1.Engineering became a separate directorate in 1978 and the Computer Science Section remained in the Mathematical and Physical Sciences (MPS) Directorate.

2.Theoretical Computer Science was a program in the Computer Science Section (CSS) of the Mathematical and Computer Science (MCS) Division within the Mathematical and Physical Sciences (MPS) Directorate.

3.National Research Council. 1982. Ad Hoc Panel to Study the Conduct of Basic Research in Computer Science and Its Interaction with Applied Research and Development. National Academies.

4.L. Fein. 1959. The role of the university in computers, data processing, and related fields. Communications of the ACM, 2(9): 7–14. DOI: 10.1145/368424.368427.

5.S. Gorn. 1963. The computer and information sciences: A new basic discipline. SIAM Review, 5(2): 150–155. https://www.jstor.org/stable/2027479.

6.G. E. Forsythe. 1967. A university’s educational program in computer science. Communications of the ACM, 10(1): 3–11. DOI: 10.1145/363018.363038.

7.A. Newell, A. J. Perlis, and H. A. Simon. 1967. Computer science. Science, 157(3795): 1373– 1374. DOI: 10.1126/science.157.3795.1373-b.

8.D. E. Knuth. 1968. The Art of Computer Programming. Volume 1: Fundamental Algorithms (1st ed.). Reading, MA: Addison-Wesley Publishing.

9.P. Wegner. 1968. Programming Languages, Information Structures, and Machine Organization, McGraw Hill.

10.T. J. Misa, ed. 2016. Communities of Computing: Computer Science and Society in the ACM, Morgan & Claypool.

11.J. Abbate. October 2013. Is computer science “Science”? A half-century debate. Keynote talk, 2nd International Conference on History and Philosophy of Computing, Paris. https://hapoc2013.sciencesconf.org/27047/document.

12.S. D. Conte, J. W. Hamblen, W. B. Kehl, S. O. Navarro, W. C. Rheinboldt, D. M. Young, Jr., and W. F. Atchinson. 1965. An undergraduate program in computer science—Preliminary recommendations. Communications of the ACM, 8(9): 543–552. DOI: 10.1145/365559.366069.

13.G. E. Forsythe. 1965. President’s letter to the ACM membership: Why ACM? Communications of the ACM, 8(3): 143–144. DOI: 10.1145/363791.363792.

14.H. A. Simon. 1969. The Sciences of the Artificial. Cambridge, MA: MIT Press, p. 21.

15.J. Abbate. 2016. From handmaiden to “proper intellectual discipline”: Creating a scientific identity for computer science in 1960s America. In Communities of Computing: Computer Science and Society in the ACM (edited by T. Misa). Morgan & Claypool, p. 28.

16.Gordon Bell. January 23, 1974. Letter to Edward C. Creutz (AD/R), internal document.

17.This idea would return in the Feldman, Snowbird, and other reports.

18.National Science Foundation. 1974. Twenty-Fourth Annual Report for Fiscal Year 1974. Washington, DC: U.S. Government Printing Office.

19.B. W. Arden. 1976. The computer science and engineering research study (COSERS). Communications of the ACM, 19(12): 670–673. DOI: 10.1145/360373.360376.

20.B. W. Arden. 1983. What Can Be Automated?: Computer Science and Engineering Research Study. MIT Press.

21.J. A. Feldman and W. R. Sutherland. 1979. Rejuvenating experimental computer science: A report to the National Science Foundation and others. Communications of the ACM, 22(9): 497–502. DOI: 10.1145/359146.359147.

22.D. D. McCracken, P. J. Denning, and D. H. Brandin. 1979. An ACM executive committee position on the crisis in experimental computer science. Communications of the ACM, 22(9): 503–504. DOI: 10.1145/359146.362786.

23.P. J. Denning, E. A. Feigenbaum, P. Gilmore, A. C. Hearn, R. W. Ritchie, and J. F. Traub.1981. A discipline in crisis. Communications of the ACM, 24(6): 370–374. DOI: 10.1145/ 358669.358682.

24.National Science Foundation. May 21–23, 1979. Summary Minutes of the Advisory Subcommittee for Computer Science. Washington, DC.

25.P. J. Denning. 1981. ACM President’s letter: Eating our seed corn. Communications of the ACM, 24(6): 341–343. DOI: 10.1145/358669.358672.

26.J. F. Traub. 1981. Quo vadimus: Computer science in a decade. Communications of the ACM, 24(6): 351–369. DOI: 10.1145/358669.358677.

27.J. Hopcroft, B. Lampson, J. McCarthy, A. Penzias, M. Rabin, J. Schwartz, D. Scott, and A. Van Dam. December 1986. Preliminary Draft Report on the Scientific Contributions of Computer Science. National Science Foundation.

28.J. E. Hopcroft, K. W. Kennedy, and K. K. Curtis. 1989. Computer Science: Achievements and Opportunities. Philadelphia: Society for Industrial and Applied Mathematics.

29.J. McCarthy. April 13, 1996. A Petition for the Withdrawal of “Computing the Future”. http://jmc.stanford.edu/commentary/petition.html; last accessed 31 December 2018.

30.National Research Council. 1997. Defining a Decade: Envisioning CSTB’s Second 10 Years. Washington, DC: The National Academies Press.

31.L. Kleinrock. 1998. Toward a National Research Network. National Academies.

32.J. F. Traub and National Research Council. 1988. The National Challenge in Computer Science and Technology. National Academies.

33.Links to CSTB publications can be found here: http://sites.nationalacademies.org/CSTB/ CSTB_042201.

34.National Research Council, 1997.

35.CRA provided three earlier reports: (1) P. Freeman and W. Aspray. 1999. The Supply of Information Technology Workers in the United States. https://eric.ed.gov/?id=ED459346. (2) J. Cuny and W. Aspray. 2001. Recruitment and Retention of Women Graduate Students in Computer Science and Engineering. Computing Research Association. https://www.researchgate.net/publication/239547208_Recruitment_and_retention_of_ women_graduate_students_in_computer_science_and_engineering. (3) J. Stankovic and W. Aspray. 2003. Recruitment and Retention of Faculty in Computer Science and Engineering. Computing Research Association. http://archive.cra.org/reports/r&rfaculty.pdf.

36.Y. Oyanagi. 1999. Development of supercomputers in Japan: Hardware and software. Parallel Computing, 25: 1545–1567. DOI: 10.1016/S0167-8191(99)00084-8.

37.National Research Council. 2005. Getting up to Speed: The Future of Supercomputing. National Academies Press.

38.Panel on Large Scale Computing in Science and Engineering, and P. D. Lax. 1983. Report of the Panel on Large Scale Computing in Science and Engineering. National Science Foundation. https://www.pnnl.gov/scales/docs/lax_report1982.pdf.

39.K. G. Wilson. 1988. On Supercomputing. In Panel on Large Scale Computing in Science and Engineering and P. D. Lax. 1983. National Science Foundation. Supplement to the “Report of the Panel on Large Scale Computing in Science and Engineering,” pp. 24–33.

40.J. F. Decker. 1983. “Report to the Federal Coordinating Council on Science, Engineering and Technology Supercomputer Panel on Recommended Government Actions to Provide Access to Supercomputers.” In Supercomputers, Hearings before the Committee on Science and Technology, U.S. House of Representatives, Ninety-eighth Congress, First Session, November, vol. 15, p. 16.

41.M. Bardon. 1983. A National Computing Environment for Academic Research. National Science Foundation: Report of an NSF Working Group on Computers for Research.

42.Committee on Science, Engineering, and Public Policy. September 7, 1984. Report of the Research Briefing Panel on Computer Architecture, National Academies of Sciences and Engineering and the Institute of Medicine. DOI: 10.17226/19410.

43.Erich Bloch. September 11, 1984. NSF internal memorandum to Carl Hall, AAD/ENG, Marcel Bardon, AAD/MPS, and John Connolly, OD/OASC. Charles Babbage Institute.

44.Kent Curtis, DD/DCR, Blake Cherrington, DD/ECSE, and John Connolly, Head/OASC. September 24, 1984. NSF internal memorandum to Erich Bloch, Director. Charles Babbage Institute.

45.R. Adrion, D. Farber, F. F. Kuo, L. H. Landweber, D. Angel, and J. B. Wyatt. December 1984. SCIENCENET: Report on the Evolution of a National Supercomputer Access Network. National Science Foundation.

46.D. M. Jennings, L. H. Landweber, I. H. Fuchs, D. J. Farber, and W. R. Adrion. 1986. Computer networking for scientists. Science, 231(4741): 943–950. DOI: 10.1126/science.231.4741.943.

47.This quote is from the introduction to the program activities of the Division of Computing Research (DCR) in the Mathematical and Physical Sciences and Engineering (MPE) Directorate. NSF Annual Report, 1975.

48.I applied and was recruited to NSF by Kent Curtis and joined in a “rotator” position. NSF brings academics into serve as “rotating” program officers using several employment strategies including Visiting Scientist, Engineer, and Educator (VSEE); Intergovernmental Personnel Act (IPA); and temporary excepted service assignments.

49.Congress changed the fiscal year start from July 1 to October 1 beginning in 1977, hoping the later start would find all federal budgets passed before the fiscal year began. The result was a need for a 3-month “Transition Quarter.” See J. J. Hogan. 1985. Ten years after: The US Congressional Budget and Impoundment Control Act of 1974. Public Administration, 63(2): 133–149. DOI: 10.1111/j.1467-9299.1985.tb00896.x.

50.ESL, Inc. was a defense contractor specializing in satellite and other photographic image processing. ESL was led by William Perry, later Secretary of Defense under Clinton and earlier Under Secretary of Defense for Research and Engineering from 1977 to 1981, and Robert Fossum, who was DARPA Director from 1976 to 1981.

51.Sponsored Projects Office, University of California, Berkeley. Quick Guide to Working with NSF FastLane and Research.gov. https://spo.berkeley.edu/guide/fastlanequick.html; last accessed 19 March 2019.

52.Theoretical Computer Science was the largest program and received only about 100 proposals per year.

53.Oral history, Rick Weingarten, interviewed by Peter Freeman, July 11, 2017. Charles Babbage Institute.

54.R. Adrion and S. Mahaney. 2003. Shooting inward. Computing Research News, 15(1). archive.cra.org/CRN/articles/jan03/adrion.mahaney.html; last accessed 15 January 2019.

55.Nomination information and winners can be found at https://amturing.acm.org/.

56.The Turing Award is presented each June at the ACM Awards Banquet and is accompanied by a prize of $1,000,000 plus travel expenses to the banquet. Financial support for the award [currently] is provided by Google Inc.

57.D. E. Knuth. 1969. The Art of Computer Programming, Vol. 1. Reading, MA, Addison-Wesley Publising Company.

58.D. Knuth. 1986. Computers & Typesetting. Reading, MA: Addison-Wesley.

59.Ted Codd, Dennis Ritchie, Ken Thompson, John Cocke, Ivan Sutherland, Butler Lampson, Dick Sterns, Doug Engelbart, Jim Gray, Alan Kay, Vint Cerf, Bob Kahn, Fran Allen, and Chuck Thacker were primarily in industry or government, while Tony Hoare, Steve Cook, Niklaus Wirth, Robin Milner, Ole Dahl, Kristen Nygaard, Peter Naur, Joseph Sifakis, Silvio Micali, and Tim Berners-Lee carried out their Turing-related research in Europe, Canada, or Israel.

60.R. K. Yin. 2011. Applications of Case Study Research. Sage Publishers.

61.D. Leavitt. May 3, 1972. NSF-sponsored research could lead to national science computer network. Computerworld, p. 10.

62.D. D. Aufenkamp and E. C. Weiss. 1972. NSF activities related to a national science computer network. Computer Communications: Impacts and Implications, p. 226.

63.M. Greenberger, J. Aronofsky, J. L. McKenney, and W. F. Massy. 1973. Computer and information networks: The movement in research and education toward national resource sharing via networks is accelerating. Science, 182(4107): 29–35. DOI: 10.1126/science.182 .4107.29.

64.J. Abbate. 2000. Inventing the Internet. MIT Press.

65.Telenet was developed by Larry Roberts, the original ARPANET program manager.

66.S. R. Hiltz. June 1981. The Impact of a Computerized Conferencing System on Scientific Research Communities. FINAL REPORT NSF-MCS-77-27813.

67.M. E. Smid and D. K. Branstad. 1988. Data encryption standard: Past and future. Proceedings of the IEEE, 76(5): 550–559. DOI: 10.1109/5.4441.

68.W. Diffie and M. E. Hellman. 1977. Special feature exhaustive cryptanalysis of the NBS data encryption standard. Computer, 10(6): 74–84. DOI: 10.1109/C-M.1977.217750.

69.FIPS Publication 46-3. 1999. Data encryption standard (DES). National Institute of Standards and Technology 25(10): 1–22. https://csrc.nist.gov/csrc/media/publications/fips/46/3/ archive/1999-10-25/documents/fips46-3.pdf.

70.W. Diffie and M. Hellman. 1976. New directions in cryptography. IEEE Transactions on Information Theory, 22(6): 644–654. DOI: 10.1109/TIT.1976.1055638.

71.R. L. Rivest, A. Shamir, and L. Adleman. 1978. A method for obtaining digital signatures and public-key cryptosystems. Communications of the ACM, 21(2): 120–126. DOI: 10.1145/ 359340.359342.

72.J. H. Ellis. 1970. The possibility of secure non-secret digital encryption. UK Communications Electronics Security Group, vol. 6. http://cryptocellar.org/cesg/possnse.pdf.

73.C. C. Cocks. 1973. “A note on non-secret encryption.” CESG Memo. https://www.semanticscholar.org/paper/A-note-on-non-secret-encryption-Cocks/ fab40f7645d931a6094e1d8c0113143091942ad3.

74.D. Shapley and G. B. Kolata. 1977. Cryptology: Scientists puzzle over threat to open research. Science, 197(4311): 1345–1349. http://science.sciencemag.org/; last accessed 2 March 2018. Also: D. Shapley. 1977. Cryptography meeting goes smoothly. Science, 198(4316): 476. http://science.sciencemag.org/; last accessed 2 March 2018.

75.J. C. Cherniavsky. 1985. Case study: Openness and secrecy in computer research. Science, Technology, & Human Values, 10(2): 99–103. DOI: 10.1177/016224398501000214.

76.R. C. Buck. 1982. The public cryptography study group. Computers & Security, 1(3): 249–254. DOI: 10.1016/0167-4048(82)90043-8.

77.P. J. Denning, D. H. Brandin, D. C. Schwartz, and G. I. Davida. 1981. Report of the public cryptography study group. Communications of the ACM, 24(7): 434–450. DOI: 10.1145/358699 .358710.

78.Denning et al., 1981, op. cit.

79.G. B. Kolata. 1980. Cryptography: A new clash between academic freedom and national security. Science, 209(4460): 995–996. DOI: 10.1126/science.209.4460.995.

80.B. R. Inman. 1980. Cryptography research funding. Science, 210(4466): 134. DOI: 10.1126/ science.210.4466.134.

81.D. A. Langenberg. 1980. Cryptography, NSF, and NSA. Science Magazine, 210(4473). DOI: 10.1126/science.210.4473.960.

82.I left NSF in the fall of 1978 and returned to manage a different set of programs in January 1980, so I am not familiar with the process used by the program directors who followed me. I am quite confident that something similar to this process continues today.

83.Institute of Medicine, National Academy of Sciences, and National Academy of Engineering. 1982. Scientific Communication and National Security. Washington, DC: The National Academies Press. DOI: 10.17226/253.

84.Minutes of the May 28–29, 1981, meeting of the NSF Advisory Subcommittee on Computer Science. Jack Minker, Chair. Charles Babbage Institute.

85.J. Guttag et al. May 1981. The Role of the NSF in Supporting Cryptological Research: A Report to the National Science Foundation by Its Mathematical and Computer Sciences Advisory Committee. National Science Foundation. Charles Babbage Institute.

86.Thelma Estrin was married to Gerald Estrin and both were UCLA computer science professors. They had three daughters. Margo Estrin is a medical doctor. Deborah Estrin is a leading computer scientist, who moved from USC to UCLA after her parents retired. She led the NSF Science and Technology Center for Embedded Networked Sensing (CENS). Recently, she joined Cornell Tech. Judith Estrin is a corporate executive and entrepreneur, having founded Zilog, Bridge Communications, Network Computing Devices, Precept Software, Packet Design, and JLABS. Judy served as chief technology officer and senior vice president of Cisco Systems until 2000 and CEO of Eventlive in 2013.

87.Ettore F. “Jim” Infante had served as program director for Applied Mathematics in MCS a few years earlier.

88.K. Curtis. May 21–23, 1979. Summary Minutes, Advisory Subcommittee for Computer Science. National Science Foundation. Charles Babbage Institute.

89.Curtis, 1979, op. cit. , p. 2.

90.Curtis, 1979, op. cit. , p. 3.

91.P. Young. December 9, 1979. “ReDRAFT: NSF support for experimental computer science,” Internal document. Charles Babbage Institute.

92.Oral history, Rick Adrion, interviewed by William Aspray, October 29, 1990, in Amherst, Massachusetts. Charles Babbage Institute. https://conservancy.umn.edu/handle/11299/ 104300; last accessed 15 January 2019.

93.R. Riesenfeld and L. Hollaar. 1984: NSF Grant 8406682-1984 NSF CER Conference to Be Held February 23–24, 1984, at the University of Utah, Salt Lake City, Utah (Computer Research). https://www.nsf.gov/awardsearch/showAward?AWD_ID=8406682&HistoricalAwards=false; last accessed 14 February 2019.

94.J. Schwartz. November 1983. A Taxonomic Table of Parallel Computers, Based on 55 Designs. Ultracomputer Note #69. Courant Institute, N.Y.U.

95.Ultracomputer Project, https://cs.nyu.edu/cs/projects/ultra/home-page.html; last accessed 15 February 2019.

96.D. Gajski, D. J. Kuck, D. H. Lawrie, and A. S. Sameh. 1983. CEDAR—A large scale multiprocessor. Proceedings of the 1983 International Conference on Parallel Processing, Belaire, MI. DOI: 10.1145/859526.859527.

97.M. Sejnowski, E. T. Upchurch, R. N. Kapur, D. P. S. Charlu, and G. J. Lipovski. 1980. An overview of the Texas reconfigurable array computer. AFIPS, Proceedings of the May 19–22, 1980, National Computer Conference, p. 631. DOI: 10.1145/1500518.1500623.

98.J. R. Goodman and C. H. Sequin. 1981. Hypertree: A multiprocessor interconnection topology. IEEE Transactions on Computers, 20(12): 923–933. DOI: 10.1109/TC.1981.1675731.

99.R. M. Keller, G. Lindstrom, and S. Patil. 1979. A loosely-coupled applicative multi-processing system. AFIPS Conference Proceedings, 48: 613–622. DOI: 10.1109/AFIPS.1979.8.

100.D. J. DeWitt, R. Gerber, G. Graefe, M. Heytens, K. Kumar, and M. Muralikrishna. 1986. “Gamma—A high performance dataflow database machine.” University of Wisconsin– Madison Department of Computer Sciences.

101.C. Rieger, R. H. Trigg, and B. Bane. 1981. ZMOB: A new computing engine for AI. IJCAI, pp. 955–960. https://www.ijcai.org/Proceedings/81-2/Papers/071.pdf.