Читать книгу Computing and the National Science Foundation, 1950-2016 - William Aspray - Страница 8

Preface

The Computer and Information Science and Engineering (CISE) Directorate and its predecessors at the National Science Foundation (NSF) have played a seminal but untold role in the growth of computing¹ from the 1950s to today. Since the mid-1990s, CISE has provided a large majority of all funding for basic research in computer science and closely related disciplines in the United States, as well as substantial support for other fields that study computing or push the state-of-the-art of advanced computation. The results have formed the foundations on which modern computing is built.

Two of the authors of this book, Peter Freeman and Rick Adrion, were aware of much of this history and knew also that, to date, no comprehensive record of the influential role played by CISE and its predecessors existed. As a result, in late 2016, we undertook to remedy this situation by producing a documented history of NSF’s role in modern computing. Recognizing that we had no formal training as historians, we enlisted William Aspray, an historian who had published extensively on computing-related subjects including at NSF,² and with whom we had worked in other contexts, to join us on the project; his experience has been essential. This book, and a related publicly available collection of research materials³ deposited at the Charles Babbage Institute (CBI) of the University of Minnesota, are the principal results of our efforts.

Our project had four objectives. The first was to bring together as much information as possible that pertains to the history of computing⁴ at NSF. We have collected approximately 4,000 paper and electronic records, which were donated to the CBI.⁵ We spent considerable time talking with longtime members of the CISE staff to locate materials and develop context for later project activities. We also collected materials and consulted various archival collections.⁶

We have augmented this written material with approximately 50 oral histories,⁷ which have been transcribed and lightly edited. (Most of these will be available through the CBI as well.) They include interviews with several NSF directors and eight of the nine living Assistant Directors (ADs) of CISE.⁸ Additional oral histories were conducted with staff within CISE (program officers, division directors, or chief scientists) as well as with other members of the Washington computer science community; for example, former members of the White House Office of Science and Technology Policy (OSTP). A list appears in the back of this book.

A second objective of this project was to rigorously document major events in the history of NSF support for computing research and education. Throughout the text, we have provided citations to numerous primary sources, including NSF internal memoranda and internal plans that are no longer sensitive, published documents, and other government publications. In those cases where materials we cite in this book would be difficult for readers to obtain, we have placed them with the CBI.

A third objective was to write a set of narratives describing the history in a readable and accessible way. This has been greatly facilitated by the fact that both Adrion and Freeman served as employees or rotators⁹ at NSF on several occasions for a combined total of 18 years, and were not only active researchers and educators (professors) but also engaged members of the professional community for almost 50 years each. Additionally, Aspray had led a team in the early 1990s that produced a large body of unpublished research on pre-CISE activities based on internal NSF documents. When he became part of this project, we gained particular advantage for satisfying this objective as he accessed those writings and drew from his 40-plus years of experience as an historian of computing.

Our fourth objective was to analyze what we have learned. Conclusions are indeed drawn in Chapter 13 as well throughout the rest of the text. However, generally speaking, we have not evaluated CISE programs or the individual projects that CISE supported; where we have offered judgmental opinions, these are solely the opinions of the chapter’s author(s). Further analysis must await future authors. We mention other major government funders including DARPA, NASA, DoE, and the military research agencies; science policy in both the legislative and executive branches of the federal government and in the National Academies; and computing professional organizations including ACM, the IEEE Computer Society, SIAM, AFIPS, and the Computing Research Association. While we occasionally discuss the relations of these organizations to NSF, we have not identified and analyzed the many connections among the various players in this milieu and NSF. Nor have we tried to evaluate their relative contributions and merits.

Our primary focus has been on CISE (created in 1986) and its predecessor organizations, such as the Office of Computing Activities (OCA, created in 1967) and the Office of Science Information Services (OSIS). However, computing activities within NSF often extended beyond the boundaries of OCA and OSIS.¹⁰ We mention these, but typically do not follow them in the same detail that we give to CISE and its predecessors.

Readers will find that there is some variation in the nature of the three main parts of this book, and even variation in style among its individual chapters. We wrote some chapters as participant accounts, but wrote others more objectively as historians who did not directly participate in the events described. In parts of Chapters 3, 4, and 9, for example, Peter Freeman writes from the perspective of a direct participant; in Chapter 12, he reflects on his time as the AD/CISE. In Chapters 1 and 2, Rick Adrion draws upon his early role as a program director and on his later key management experience at NSF to tell the story of critical events before and after CISE was created. In Chapter 13, Freeman and Adrion reflect on the history of NSF and computing to identify some themes that may help in future understanding. William Aspray, who has never been employed by NSF, has worked as both an historian and as the executive director of the Computing Research Association (CRA—one of the major non-profit players in Washington on computing research policy). The chapters he wrote on CISE’s role in the development of modern computing are informed by this perspective.

While there has been coordination among the authors to ensure thorough coverage of the history of computing at NSF in the period from 1950 to 2016, this book is best read as a collection of linked essays rather than as a tightly written monograph. The three authors each have their individual voices, and no effort has been made to harmonize them completely. While we have all read and critiqued each other’s chapters, we did not have a goal of forging a unified position throughout.

Our book is composed of distinct parts that present the results of our work on our project over the past two years. Part I provides a narrative of the history of NSF’s involvement in the world of digital computing, especially as it relates to the funding activities of CISE and its predecessors. (Table P.1 gives a timeline of some of the key events in this narrative, to assist in comprehending some of the milestones passed.) Part II goes into more depth on a selected set of important topics. Part III provides our conclusions, and the appendixes present NSF organizational charts over time, a list of the interviews we conducted, a non-exhaustive set of short biographies, a description of the archive we prepared, and a list of abbreviations and acronyms,

Before providing a guide to using this book, we provide very short characterizations of each of the 13 chapters.

Chapter 1 covers computing activities related to science information, facilities, education, and basic research in the period from 1950 to 1974. The most active early support entailed providing science information and support for research in information retrieval, databases, and computational linguistics. Computing facilities and education were supported more heavily than computing research, but did enable the creation of some of the earliest computer science academic departments. Creation of the Office of Computing Activities (OCA) in 1967 was a landmark development because it strengthened support for computing research and provided organizational status; that resulted in stronger ties to other NSF programs and the NSF imprimatur to fledgling academic computer science departments.

Table P.1Select events in the history of NSF and computing, 1950–2008

Year	Event
1950	NSF enabling act signed in November; operations begin in 1951
1951	Office of Science Information (OSI) created
1953	Assistance given to buy computer for research
1954	First training/education grant.
1955	von Neumann panel recommends research on design of computers; National Science Board approves facilities program
1957	First grants for computing research
1958	Office of Science Information Services (OSIS) created; NSF expands computing facilities, research and education investments
1963	Early Training grants led to the first CS curricula and departments
1966	Rosser Report
1967	Pierce Report; Office of Computing Activities (OCA) created
1974	Division of Computer Research (DCR) created, then recreated in 1984
1978	Theorynet and Debate on Public Cryptography
1980	CER (experimental research) and CSNET (networking) programs begin
1984	Supercomputer Centers created and NSFNET begins
1986	CISE created
1995	NSFNET converted to Internet
1999	ITR program started
2003	Major reorganization of CISE
2004	GENI Program started
2005	Broadening Participation Program started
2005	Office of Cyberinfrastructure (OCI) created in O/D
2006	CCC created
2008	Expeditions in Computing, Cyber-Enabled Discovery programs begin
2013	OCI moved from O/D to CISE as Division of Advanced Cyberinfrastructure (ACI)
2016	ACI made into an office (OAC) within CISE to provide better connection with rest of NSF

Chapter 2 covers the years from 1974 until the founding of CISE in 1986. In addition to organizational changes and further strengthening of computing programs, there was support for efforts to professionalize and define computer science. A series of reports (e.g., Feldman, Snowbird, Hopcroft-Kennedy, Lax, Bardon-Curtis) shaped NSF’s computing efforts. Among the results were cryptologic research, the Coordinated Experimental Research (CER) program, and the CSNET and NSFET networking initiatives; these are afforded expanded discussion. This growing importance of computer science and of computing, coupled with internal efforts by several NSF staff, led to the founding of CISE.

Chapter 3 covers the years from the founding of CISE in 1986 through 1998. There was a succession of short-term ADs: Gordon Bell, William Wulf, Nico Habermann, Paul Young, and Juris Hartmanis (all served approximately two years each). In spite of some internal pushback, the new Directorate quickly established its structure and importance within both NSF and the federal government. During the 1990s the first easily usable browser (Mosaic), conversion of NSFNET into the Internet, and the emergence of Google were all enabled in some way by CISE support.¹¹ By 1999 CISE started to receive greater funding from Congress, increased respect within NSF, and sustained leadership from its scientific community.

Chapter 4 covers 1999–2006 when Ruzena Bajcsy and Peter Freeman served as CISE ADs. Major initiatives increased support for cyberinfrastructure, greatly expanded the field with the Information Technology Research (ITR) program, reorganized CISE, started new funding programs in networking research (GENI—the Global Environment for Network Innovations), cybersecurity research (a centers program), and the Broadening Participation in Computing (BPC) program. Direct actions by CISE made significant management changes in the supercomputer centers and strengthened the cyberinfrastructure and basic research programs. Initial plans were laid for later initiatives including the Expeditions in Computing program and Cyber-enabled Discovery and Innovation (CDI).

Chapter 5 covers 2007–2016. Three individuals served as AD/CISE: Jeannette Wing, Farnam Jahanian, and James Kurose. Budgets were tweaked to ensure that basic computer science research was protected and that CISE received fewer but better proposals. A major one-time appropriation was received and successfully managed as part of President Obama’s stimulus package. Several major programs, such as GENI, Expeditions in Computing, and Cyber-enabled Discovery and Innovation were furthered during this time. The Computing Community Consortium was continued and there was increased partnering with other directorates and industry.

Chapter 6 provides a detailed analysis of the NSF programs in computer facilities and computer education prior to the founding of CISE in 1986.

Chapter 7 provides case studies of early NSF support for research in circuits, computer architecture, software, numerical analysis, computer engineering theory, artificial intelligence, and computer graphics.

Chapter 8 covers the Information Technology Research Program from its beginning in FY (fiscal year) 2000 through to 2005, when it became part of base CISE research funding.

Chapter 9 provides a case study of NSF’s support of research on concepts and mechanisms of networking, and deployment of operational networks.

Chapter 10 covers High Performance Computing, an activity NSF has supported even as the power of such machines has grown exponentially.

Chapter 11 covers CISE’s programs to broaden participation in computing to women, underrepresented minorities, and the disabled.

Chapter 12 provides a personal view of what a CISE AD does.

Chapter 13 recaps the narratives in Chapters 1 to 5 and provides a set of high-level conclusions about the history of computing and NSF funding.

Readers seeking an overview of NSF activities in computing research and education, as well as related activities, are encouraged first to read Chapters 1–5, and then follow up by reading any deeper studies that are of particular interest. The organizational charts in the appendixes may also be useful in understanding one aspect of the changing relationship between NSF and computing.

Readers with limited time and/or scope of interest may want to read only the chapter(s) in Part II that speak to their interests. A quick scan of the chapters’ beginnings may help to determine whether one of them addresses the reader’s interest.

Table P.2 may be of use in connecting Part I chapters with Part II chapters. It illustrates the major connections between a given chronological chapter and one or more subject study chapters.

Work on this project was supported in part by NSF Grant #1743282, EAGER: Exploring the History and Impact of the Computing and Information Science and Engineering (CISE) Directorate of the National Science Foundation, a grant made to the Massachusetts Green High Performance Computing Center (MGHPCC). We worked independently of the NSF. Any views expressed in this book are solely due to us or named third-party sources, not the NSF nor the MGHPCC. Any errors of fact are our responsiblity.

Table P.2Relation between Part I chapters and Part II chapters

Work on this project would not have been possible without NSF support and the help of many people. Erwin Gianchandani, currently Deputy AD/CISE, guided us on the usage of NSF materials and other issues. NSF Historian Leo Slater answered questions and Assistant NSF Historian Emily Gibson provided access to some NSF records. Janet Abbate (Virginia Tech), Thomas Haigh (University of WisconsinMilwaukee), and Jeffrey Yost (Charles Babbage Institute) served as our historical advisory committee. We have worked closely with Amanda Wick, the Acting CBI Archivist, on the deposit of project materials at CBI. Several former and current CISE staff have donated material to the project. Over 50 individuals have agreed to sit for oral history interviews. A succession of four people provided diligent support to the project: Jana Vetter, Julia Fan, Jessica Ewen, and Kayla Heslin. In particular, we want to thank Julia and Jessica for their work on the oral histories and Kayla for the work at the end of the project as we compiled this book and readied materials to be sent to the CBI. Finally, we wish to thank our families, who have sometimes missed us and been neglected as we worked on this project. Many thanks to all!

Notes

1.Before the early 1960s there was no computer science, but by the 1970s the term was widely known and departments of computer science rapidly became a dominant academic unit and scientific discipline. We will use “computer science” primarily to refer to the research discipline and “computing” to refer to the broader activity of using computers and studying that usage.

2.William Aspray, Bernard O. Williams, and Andrew Goldstein, “Computing as Servant and Science: Impact of the National Science Foundation” (unpublished, 510 pages, 1992).

3.The CISE History Archive (CHA) is described in Part III of this book.

4.We use the ambiguous term “computing” to denote computer science and closely related disciplines, but not all uses of computing by other fields; however, especially in the early days, the distinction was not yet clear.

5.One motivation for this collection effort was the physical move of NSF headquarters in September 2017 from Arlington to Alexandria, Virginia, and the knowledge that valuable documents might be discarded. Another motivation was that early NSF employees are starting to pass on—and their memories and their documentation with them. Two interviewees passed away during the project and several potential interviewees were incapacitated.

6.These included documents from Gordon Bell, Mel Ciment, Mike Foster, John King, Irene Lombardo, Jack Minker, Rick Adrion, and Peter Freeman. Archival collections consulted included those of Ed Feigenbaum and John McCarthy at the Stanford archives.

7.The oral history record is strong but not complete. Many of the principal people involved with CISE and its predecessors have been interviewed, but a few are deceased, a few we could not reach or they did not agree to be interviewed, and due to oversight or lack of time, no doubt a few were missed. While there were perhaps 10 or 20 oral histories concerning the NSF computing story existing at the time we began this project (mostly at the Charles Babbage Institute Archives, the IEEE History Center, and the Computer History Museum), the new interviews we have added represent a major increase in coverage of this topic.

8.The AD is the head of the directorate; “Assistant” indicates they also have NSF-wide responsibilities, reporting to the NSF Director.

9.A “rotator” at NSF is a person on leave from their home institution to work at NSF under the Intergovernmental Personnel Act (IPA) or as temporary employees.

10.This is true of early work on science information and information science. Highperformance computing and cyberinfrastructure have sometimes been housed within CISE, but at other times either in the Office of the Director or in their own freestanding office reporting to the Director. At times, computing activities have existed in other NSF directorates: especially Engineering, Mathematics and Physical Sciences, Biology, and Education and Human Resources.

11.See the list of acronyms and abbreviations that appears in Appendix E of this book.