Early History of U.S. Policy on IT Workers

By William Aspray

The supply of information technology workers in the United States has become a major policy issue over the past several years, involving visa legislation, tax credits for corporate training programs, regional training programs, and major investments in education and research. To provide some perspective on these recent events, this article and a follow-up that will appear in the November issue of CRN will trace the history of U.S. policy concerning information technology workers.

Generally, the federal government has not established direct labor policy for information technology workers. Instead, it has been embedded in policies for science, education, public welfare, economics, and business. In the period covered here (up through the late 1970s), IT labor policy was closely related to science and education policy-especially as it applied to the National Science Foundation (NSF)-and to one failed effort connected to social welfare policy.

In the decade following the end of the World War II, the computer was largely conceived as a scientific instrument. Only near the end of that period did people begin to recognize its potential for use in data processing. This change brought with it a demand for many more computers and IT workers, as well as the formation of a computer manufacturing industry.

The first conference on the training of workers for the computing field was held at the end of this era, in 1954, at Wayne State University in Detroit. It was clear at this time that mathematicians were in demand to fill most computer jobs, even in data-processing environments. The vast majority of practicing mathematicians had been trained before the computer was invented, and some uneasiness was expressed about a shortage of mathematicians to fill these positions. While research-related agencies, such as the Office of Naval Research and the National Science Foundation, participated in this conference, there was no government action at the time, other than to monitor numbers of practitioners and students in mathematics. No one at the time had a good sense of the number of computer workers that would be needed.

There continues to be active interest in the IT worker situation at both the regional and the national level across the United States. Here is a small sample of the recent activities, which is by no means complete: The National Research Council has recently launched a study of Workforce Needs in Information Technology and Other High Technology Fields. They are sponsoring open hearings around the country, the first on September 24 in Santa Clara, California. For more information about the project, see http://www4.nationalacademies.org/cpsma/itwpublic2.nsf. Joint Venture: Silicon Valley Network, which is a civic action nonprofit located in San Jose, has recently released Joint Venture's Workforce Study: An Analysis of the Workforce Gap in Silicon Valley. You can reach them for a copy at 408 271-7213 (or 800 573 JVSV outside 408 area code) or by email at jvsvoffice@aol.com. Other interesting information is available on their website, http://www.jointventure.org. Computing Research Association has distributed more than 4,000 copies of its report, Peter Freeman and William Aspray, The Supply of Information Technology Workers in the United States. Single copies are still available free of charge from CRA, or the report can be viewed and downloaded from the CRA website http://www.cra.org.

The Information Technology Association of America has a variety of on-going programs and projects associated with increasing the supply of information technology workers in the United States. They are undertaking a study of IT skills used in the workplace, scheduled for completion in October 1999. Other projects include engaging IT employers in school-to-work programs, helping women in transition achieve economic self-sufficiency through IT job training, and IT training and placement programs for persons with disabilities, as well as other studies, partnerships, and conferences. For more information, seehttp://www.itaa.org/workforce/.

The period from 1956 to 1962 was marked by rapid growth in computing use and the number of computing machines, and by an equally rapid expansion of the computer manufacturing industry. This growth was driven in part by Cold War needs. For example, System Development Corporation (SDC), which was formed in 1956 to carry out the programming for the computers in the SAGE missile defense system, trained 7,000 programmers in its first five years. Because of the rapid turnover of employees at SDC, many U.S. programmers received their first programming training there. Very few universities provided formal training in computer science at this time. It soon became clear that there would not be an adequate supply of mathematicians to fill computing positions, and also that it was probably not necessary for mathematicians to fill most programming jobs. SDC and other organizations thus began using standardized aptitude and personality tests to identify people who were analytical and meticulous to train for programming careers, rather than looking for mathematicians.

This era was also marked by the responses to the launching of the Soviet artificial satellite, Sputnik. The National Defense Education Act was passed, providing fellowship support for many people who became computer scientists. NSF's budget was expanded, and increasing numbers of grants were made for teacher training and curriculum development across the sciences, including computer science. Most important, however, was the computer facilities program established by NSF, which ran from 1957 through 1973 and helped some 500 U.S. colleges and universities acquire their first computers.

Throughout the 1950s, computer science education was provided on campus mainly through computer centers, and at a few schools through programs in established departments-most often electrical engineering or mathematics. Computer science programs were developed very rapidly throughout the 1960s, and half of the doctoral programs that exist in the United States today were formed between 1962 and 1972. Training began at the doctoral level, moved to the master's level and finally reached the bachelor's level. This transition was rapid, and by 1970 more students were graduating in computer science with bachelor's degrees than with doctorates. Curriculum development grants, research grants, and fellowships from the federal government were essential to this growth.

The period between 1962 and 1967 was one in which the federal government began to understand the national importance of computers. In 1962 the Information Technology Processing Office of the Defense Advanced Research Agency was founded, and within one year it was providing more funds for research in computer science than all other federal agencies combined. The federal government also began to improve the data it collected about computing. For example, NSF contracted with John Hamblen of the Southern Regional Education Board for a series of surveys about academic computer expenditures, sources of funds, and utilization for research and instruction.

It was also a time for national studies of computing. With NSF funding, the National Research Council conducted its first computer study, beginning in 1962 (published in 1966). Known widely as the Rosser Report for its chairman, the mathematician J. Barkley Rosser, the study focused on computers and their use in research. The report was somewhat useful as a planning tool for the federal agencies, documenting the rapid growth of academic computing and the critical role of the federal government in financing it. However, the report was too technical and was viewed by Congress as self-serving, so it did not succeed in increasing federal support for academic computing.

A second study (1965-1967), commissioned by the Presidential Science Advisory Committee and focusing on computers and education, was more politically effective. This report-known by the name of its chairman, the distinguished electrical engineer from Bell Labs, John Pierce-convinced President Lyndon Johnson to make computers a cornerstone of his educational platform, as a delivery system to poor and rural areas. As a result of a presidential directive, NSF formed an Office of Computing Activities (OCA), reporting directly to the NSF director (previously computing had been a program under mathematics within NSF). OCA not only supported the computerized delivery of education programs, but also supported computer science research and education. These latter programs would endure, whereas expense of computerized education coupled with the technology not being advanced enough at the time led NSF within a few years to largely divest itself of the charge to promote computerized education.

Federal action during the period between 1968 and 1976 had a generally unfavorable effect on the fledgling field of CS. In the summer of 1968, domestic programs were cut in order to pay for the Vietnam War. NSF cut its support for capital items in order to continue support for individual researchers. In effect, this was the beginning of the end for NSF's computer facilities program, which for a decade had been the federal government's most effective program for helping the universities train people in computer science. NSF decided to spread its cutbacks approximately evenly across the various scientific disciplines. This hit computer science especially hard, inasmuch as it was growing rapidly and had not existed long enough to build an adequate financial base.

Other legislative and executive actions had a similarly negative effect on computer science. The 1968 NSF Act, known generally for its sponsor, Congressman Daddario of Connecticut, split the NSF funding pie into more pieces by mandating new programs in the social sciences and applications. In order to exert greater control, the Nixon Administration largely eliminated the institutional grants that were being provided by various federal agencies, including NSF, to the universities. The universities had been able to use these grants at their own discretion, and they were one of the principal ways to support academic computing centers. Computer center directors quickly found they could not survive on the computer usage fees they received from scientific research grants, which was the way in which the Office of Management and Budget (OMB) wanted the centers supported.

The amendment to the 1972 Military Procurement Authorization, sponsored by Senator Mike Mansfield, required military agencies to divest themselves of research grants that were not directly military-related. Other mission-oriented agencies, such as NASA and the Atomic Energy Commission, voluntarily followed this practice as well. The result was an overall reduction of federal support for computing research (and hence for computer education, which was closely interlinked). Although NSF had a one-time increase in its computing support as a result of the mission agencies transferring programs to NSF, the process was disruptive and the funding was not sustained in subsequent years. Finally, in this era, there were a number of hostile questions from OMB, the General Accounting Office, Congress, and even the National Science Board about the arcane nature, practical relevance, social impact, and need for computing research given the strong industrial base.

This was the federal context in which federal agencies, particularly NSF, were working to meet the national need for computer scientists. Undergraduate enrollments were growing rapidly, but not enough doctorates were being produced to staff faculty positions in computer science. There were 200 computer science faculty members in the United States in 1967. Projections were that 223,000 undergraduates and 61,000 graduate students would take computer courses in 1968, and these would require 400 faculty members to provide the instruction. But only 40 new doctorates in computer science were being produced each year, and only 80 percent of them were entering academic careers.

OCA tried to build up research and education in computing, but it was receiving resistance, especially from OMB. Noting the surplus of scientists in many fields, OMB began a campaign in 1971 to eliminate NSF support for traineeships. In the end, both fellowships and traineeships for computer science were reduced. Although the origins are unclear, in the early 1970s there are various references within the computer science community to a goal of producing 1,000 doctorates per year by 1980. Unfortunately, the community did not come close to succeeding; 107 doctorates were produced in 1970 rising to 213 in 1975, but then reaching stasis and only producing 218 doctorates in 1980.

The solution to this problem and others in academic computing resulted in a series of new programs, which would lead the computer science community into a new era. These will be explored in the next article. This article is based on a paper on the history of IT labor policy presented by the author in July 1999 at a conference on the international history of information technology policy held at the London School of Economics. Oxford University Press will publish the conference proceedings in 2000. The research was supported in part by the National Science Foundation.