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

Чтение книги онлайн.

СКАЧАТЬ 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 СКАЧАТЬ