Creating the Needed Interface

Computer Science and Government: ARPA/IPTO (1962-1986)

This paper is a beginning effort to explore the role of the U.S. government in building the Internet. The Information Processing Techniques Office (IPTO) created within the Advanced Research Projects Agency (ARPA) in the U.S. Department of Defense (DOD) is the early and most significant institutional form of this role. Working within this institution, scientists provided leadership in creating the new field of computer science and in giving birth to the Internet. Understanding the role of government in the creation and development of the Internet involves exploring the interface between the computer scientists working as part of IPTO and the military officers in the DOD. More fundamentally, this interface is actually an interface between the computer science community and the U.S. government.

During the war there developed a partnership between military men and scientific men. It was not brought about automatically; it is not a thing that occurs readily. These men come from different backgrounds, and it is hard for each group to understand the other....I can say to you that the morale of the scientists today as I meet them is so low, so low that while they will not refuse to serve, they will serve without enthusiasm and without fruitful inspiration.

Vannevar Bush, Riehlman Comm hearing 1954, pg. 454-455

During much of its 25 year existence, from 1962-1986, the Information Processing Techniques Office funded and provided leadership, not only for the creation of the new field of computer science, but also for a large number of significant accomplishments in this field. Among these accomplishments are the creation of time-sharing and interactive computing, of packet switching networking, VLSI (Very Large Scale Integration), AI (Artificial Intelligence), the ARPANET, and perhaps most sensationally, of the Internet. Also, under its direction and support, interactive computing and the Internet have spread into many aspects of our society and lives.

And yet the Office of Information Processing Techniques was ended in 1986. This raises the question of how did it provide the leadership to make such accomplishments possible? And then, if it was successful in doing such important feats, why was it ended?

ARPA is considered throughout the field as being the main supporter and perhaps the most important force in the course of U.S. and probably world history in the computer....the country never would have grown in the computer field the way it did if it hadn't been for ARPA.

[Ibid., pg. X-22. Discussion with Dr. L. Roberts, April 23, 1974]

Before the creation of ARPA, and IPTO, there was concern within the scientific community and in the U.S. government about how to fashion an appropriate peacetime institutional form within government to support basic scientific research. ARPA/IPTO succeeded in a significant way in providing such a form, but it also encountered problems that eventually ended its existence. This paper suggests that study of IPTO's birth, development and ending will be helpful in trying to determine what institutional form within the U.S. government is necessary to continue to provide leadership for computer science research and for the continued growth and development of the Internet.

The development and problems of the National Science Foundation (NSF) are also a relevant research questions to be studied toward determining what form of institution is needed for the future. However, since such important developments in computer science were made under leadership from ARPA/IPTO, it is more important to explore how this happened. Future study is needed, however, to examine the extent to which the NSF contributed to this effort and the problems this agency encountered that prevented any greater contribution.

To state the problem more simply, I am proposing that there is a need to study ARPA/IPTO, both its achievements and the problems it encounter, as it presents important experience toward determining how to design a U.S. government institution to support the continued development of basic research in computer science. This study is also important to provide an answer to the question of how to design a government institution to provide the needed continued oversight and support for scaling and other critical functions for the child of computer science and the IPTO, i.e. for the Internet. This paper is intended as a contribution.

Creating an Institutional Form in Government for Science

Mr. McCormack The important thing about a man in science is that he must have demonstrated ability to think originally, isn't that right.
Mr. Marchetti Yes
Mr. McCormack They are discovering things and looking ahead maybe 10 and 20 years sometimes.
Mr. Marchetti That is right.

Riehlman Comm. hearing, pg. 249

During WWII, the Office of Scientific Research and Development (OSRD) was created to bring scientists into the U.S. government to do research in support of the war effort. The OSRD was an independent agency within the Executive Branch of the U.S. government. The agency was situated so that the head of this agency, Vannevar Bush, could go directly to the President on issues involving scientists in the war effort. Commenting on the importance of this direct line of authority, Don K. Price in his book "Government and Science", (New York, 1954, pg. 45) explains that the purpose of such authority was not to have conversations with the President, "It was much more important to give him (Bush-ed) the leverage he needed in dealing with the vast network of administrative relationships on which the success of a government agency depends." What is the leverage Price is referring to? This is a question that is at the crux of the interface between government and science.

It is crucial to examine the political placement of the OSRD if one is trying to understand the modern problem of placing computer scientists into the U.S. government structure. Price traces the role that scientists and scientific studies have played in the development of government institutional forms and functions from the earliest days of the U.S. Republic through the end of WWII. Such scientific contributions include the creation of the National Bureau of Standards, the Census Bureau, helping to create various kinds of legislation including anti trust, labor, and health and safety laws and the machinery within government to enforce them. Thus he is placing the issue of the creation of an appropriate institutional form within the context of how institutions connected with scientific functions have evolved within the U.S. government and how scientific regulation has been developed.

Price documents how the interface of the scientific community and government is constantly evolving throughout the history of the U.S. Therefore there is a challenge to determine what the new institutional forms are that are needed and how to fashion them appropriately. Price refers to the creation of land grant colleges to provide for the development and spread of agricultural science as an example of a successful institutional form interfacing science and government to solve a significant problem.

To win WWII, it was necessary to enlist the active participation of the scientific community in contributing technical and scientific knowledge to the war effort. The OSRD provided the institutional form to make the needed scientist- government interface. As the war was winding down, several in the scientific community and in government recognized the need to continue an appropriate scientist-government interface for peacetime purposes. They realized that once the wartime mobilization was over, the ability to involve civilian scientists in the U.S. government in general, and in working with the Department of Defense in particular, would become more difficult. With the devastation of Europe during the war, which country would assume leadership in science and technology, especially in basic research, became a burning question in the quest for postwar world leadership. The contest was between the U.S. and the Soviet Union. Science and technology were thus to become much more important in post war America, and several in the scientific community realized the need to examine the wartime experience to see how the lessons from it could be applied to peacetime conditions.

Writing during the early 1950s, Price proposes that the OSRD was fashioned in a particularly skillful way. Its structure provided the means for scientists to remain autonomous with respect to the military but to still be able to contribute to the war effort and cooperate with the Department of Defense. Similarly, in a hearing of the Subcommittee on Military Operations of the Committee on Government Operations of the U.S. House of Representatives, convened by Congressman R. Walter Riehlman from New York [hereafter Riehlman hearings], in 1954, John von Neumann, the computer pioneer, described the problem that civilian scientists working with the military presented. He explains ("Organization and Administration of the Military and Development Programs: Hearings before a Subcommittee of the Committee on Government Operations," House of Representatives, 83rd Congress, pg. 373-374) :

There are all sorts of difficulties in marrying the civilian function of research and development with the military and those things deserve careful study in each case.

Among the issues that Von Neumann was referring to were the differences in function and nature of the scientist and the military officer. While the scientist needed to maintain his independence and his ability to search for the new and underlying laws of nature, the military officer was trained to have an unusual respect for authority and tradition.(Ibid, pg. 255, pg 32).

The Riehlman hearings were held to explore this clash of temperaments and to make recommendations for accommodating it in the institutional forms that were being proposed to bring scientists into postwar research work in government.

Explaining the problem that had been previously encountered in trying to create a partnership between scientists and the military, Assistant Secretary of Defense Donald Quarles explains there were, "A lot of Military people...(who) have a very adequate understanding of technology but not in the sense of being scientists." (Ibid., pg 25) While the scientists, according to Quarles, resented that the military officer was "trying to regiment them too much in an area that he is not competent to dictate in." (Ibid.)

Quarles explains that in doing basic research in science there is a need for "latitude," not "regimentation." But once particular applications are being developed, those in the military can participate to specify the requirements they have.

Proposing his view of the problem, Congressman Frank Ikard observed (Ibid., pg 129-130) :

[T]he very nature of the animal involved -- one is a man that is a nonconformist and a man who rebels against the accepted, or he would not be a scientist; he is searching new fields and explaining new avenues, and the military man is pretty much on the opposite side. It seems to me there could conceivably be just through the human element, considerable conflict between the two.

Responding to him, Quarles agreed (Ibid., pg 30):

If there were not breadth of vision in their problems and a very deliberate effort to get both sides to understand what it is, and to work it out, surely we would have the kind of conflict that you speak of.

The hearings raise the institutional question (Ibid. pg 32):

Does the military organization have a research and development program to create conditions which retard the work of scientists or make it difficult to retain the services of highly qualified civilian scientific and technical personnel?

Also testifying was J.W. Marchetti, who had been employed as a civilian scientist administering a laboratory that was under military control. He describes how in fact the conditions he had worked in within the U.S. government did retard scientific work and deter qualified civilian scientists from working with government. He explains (Ibid.):

If the scientific laboratory is healthy it is ever reaching for something new and to officers trained in military organization this seems chaotic. They feel it is their duty to put the house in order. When they do, they stifle the enterprise and effectively kill the goose that lays the golden eggs.

Price also considers how the diversity in the roles and temperaments of these two different kinds of professionals could be accommodated. He describes how the design of the OSRD was based on the model of the National Research Council (NRC) of the National Academy of Science (NAS). This design provided scientists with support for the independence they need to pursue their scientific activity. Price proposes that with scientists working as part of government for the U.S. military, there must be recognition of the pressures on these scientists. Along with the difference between role and function of the scientist and the military officer, there are also institutional pressures that must be taken into account when fashioning an institutional form within government. In particular, the pressure that the OSRD had to recognize was the pressure from the different branches of the military services (hereafter Services) to maintain and extend their jurisdictional boundaries and powers. Those scientists who were to work in such a situation needed a means of protection from becoming the victim of the fierce competition among the Services.

Believing that this problem had been solved in the creation of the OSRD, Price describes the mechanism of this solution. Giving the OSRD a direct line of authority to the President, it had the Presidential support necessary for Bush to protect the independence of the scientific community working with government. Furthermore, the political protection provided by the OSRD made it possible for the decisions within the OSRD to be made on the basis of scientific rather than political criteria.

In his book, "Science and the Navy", Harvey Sapolsky describes a confrontation between Bush and Rear Admiral Harold G. Bowen of the Navy. Bowen had rejected a proposal for the NRC to work with the Navy. Bush believed that this action had to be treated seriously. Bowen was penalized and a notation put into his file. Also Bush penalized the Navy in an upcoming contract award, and assigned the role of building the atom bomb to the Army rather than the Navy. Bush later explained that the reason was that the Navy had not demonstrated "sufficient respect for and an ability to work with civilian scientists." (Sapolsky, pgs 15-18)