3
Demand Crosscurrents: Emerging and Disappearing Jobs

This chapter summarizes the observations of workshop participants on demand for computing professionals. It discusses crosscutting conditions and trends that may affect the overall demand for computing professionals and highlights specific factors that affect demand for individuals in computer-related research, systems and applications development, and deployment.

INTRODUCTION

What employment opportunities exist for computing professionals? The demand for computing professionals varies by specific occupation, as discussed below, but there are general principles that apply to demand for any and all computing professionals, and it is possible to make some general observations about demand for the group as a whole, much as it is possible to discuss demand for health professionals or other multifaceted groups.

The level of demand for any occupational group comes from a combination of employee turnover, replacement hiring, and plans for growth or shrinkage in the volume of employment, among other factors (Box 3.1). Demand can be high not only when employers are expanding, but also when turnover is significant; it can be low when employers are downsizing and also when their work forces are relatively stable.1 Thus, actions taken by computing and communica-



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 29
Computing Professionals: Changing Needs for the 1990s 3 Demand Crosscurrents: Emerging and Disappearing Jobs This chapter summarizes the observations of workshop participants on demand for computing professionals. It discusses crosscutting conditions and trends that may affect the overall demand for computing professionals and highlights specific factors that affect demand for individuals in computer-related research, systems and applications development, and deployment. INTRODUCTION What employment opportunities exist for computing professionals? The demand for computing professionals varies by specific occupation, as discussed below, but there are general principles that apply to demand for any and all computing professionals, and it is possible to make some general observations about demand for the group as a whole, much as it is possible to discuss demand for health professionals or other multifaceted groups. The level of demand for any occupational group comes from a combination of employee turnover, replacement hiring, and plans for growth or shrinkage in the volume of employment, among other factors (Box 3.1). Demand can be high not only when employers are expanding, but also when turnover is significant; it can be low when employers are downsizing and also when their work forces are relatively stable.1 Thus, actions taken by computing and communica-

OCR for page 29
Computing Professionals: Changing Needs for the 1990s BOX 3.1 Factors Impinging on Demand for Computing Professionals Factors That Increase Demand Economic growth Organizational growth High turnover Government technology programs Increased exports of information technology goods and services Expansion of the user base Factors That Decrease Demand Recession Organizational shrinkage (downsizing) Low turnover Defense cutbacks (also, cutbacks in other government programs) Increased imports of information technology goods and services Factors That Are Harder to Quantify Shift to distributed computing Shift to more diverse equipment base Shift in employee mix tions firms to reduce their work forces by tens of thousands (if not more) employees, including computing professionals, and by many major users of computing and communications systems to trim their information systems organizations at the same time, are important indicators of demand, but they represent only parts of the picture. Demand has qualitative as well as quantitative dimensions. For example, whether or not the total level of demand changes, employers may decide they need more of some kinds of talent and less of others.2 Thus it is not uncommon to see a company release some employees while it is simultaneously hiring others. John McSorley, human resources manager at Apple Computer, provided a specific illustration in noting that at the time of the workshop (October 1991), Apple had just laid off 600 employees but that at the same time, it had about 720 openings. Demand can also be affected by intangible

OCR for page 29
Computing Professionals: Changing Needs for the 1990s considerations relating to the desired quality of employees; it is difficult to gauge, for example, the number and types of people who might be employed if employers could find the individuals they think they want. There is no single monolithic labor market for computing professionals. In particular, computing professionals are not bound to specific industries (unlike, say, miners or aerospace engineers).3 Producing, enhancing, advancing, and using computing and communications technologies and applications all create demand for computing professionals across an ever-broadening range of organizations. As a result, computing professional occupations are growing in relative importance in all sections of the economy. Although the relative growth has been greater in manufacturing, computing professionals are also becoming more prominent in services (Table 3.1). Moreover, within and among industries, different kinds of computing professionals are employed to do the same or similar work, even within the same organization. This variation in hiring patterns and preferences for a given kind of work is among the factors that make it difficult (and potentially misleading) to generalize about the demand for computing professionals. The rapid development and proliferation of computer-based technologies since their emergence in the late 1940s have made computing a growth field over the past 4 decades. Opportunities for computing professionals overall have grown rapidly relative to opportunities for workers in other occupations. However, because the computing professional occupations are relatively new and also require specialized skills, the absolute number of computing professionals has remained relatively small—fewer than 1 million in a U.S. work force of over 122 million.4 TABLE 3.1 Employed Computer Specialists as a Percentage of Total Employed Work Force in Private Industry, by Sector, 1980 and 1989   1980 1989 All private industry 2.6% 4.0% Manufacturing 2.6% 5.3% Nonmanufacturing 2.7% 3.6%   SOURCE: National Science Board, S&E Indicators, 1991, NSB 91-1, National Science Foundation, Washington, D.C., p. 267.

OCR for page 29
Computing Professionals: Changing Needs for the 1990s Moreover, although computing competence is expected to increase in the general population, the number of "true" computing professionals (i.e., individuals who are employed in professional computing fields, as opposed to other professionals who use computer-based tools), like the number of other scientific and engineering personnel, is likely to remain small relative to the labor force as a whole. Table 3.2 gives BLS estimates of civilian employment in 1990 and forecasts of employment in 2005 for nine major occupational groups. Professional specialty occupations, in the BLS taxonomy, included a total of fewer than 16 million individuals in 1990, among them systems analysts and computer scientists, mathematicians, electrical and electronics engineers, physicists and astronomers, college and university faculty, and numerous other professionals. Despite their small number, computing professionals are having a disproportionately large impact as the technologies they develop, produce, and deploy are being used by increasing numbers of people—both in the work force and in other settings. The fact that computing professionals as a group appear to be relatively well paid suggests that they are not in oversupply. Betty Vetter points out in Appendix A that both starting and continuing salaries for computer scientists are in the top ranks of professional salaries, exceeded by law and business (management) but higher than those for other science-or technology-based occupations. FACTORS IN THE MACRO ENVIRONMENT AFFECTING DEMAND While workshop participants expected that overall demand for computing professionals would continue to grow for the foreseeable future, a number of factors suggest that the growth rate is slowing because of short-run and long-run economic conditions (Box 3.2). The Recession and Slow Recovery At this writing, the most obvious influence on demand is the recent economic recession together with a subsequent slow recovery, which have depressed demand for employees across the economy. The recession and slow recovery have dampened growth in applications of computing technology and have depressed new business startups, both potential sources of demand for computing professionals. However, the computing sector has continued to grow; sales volume has increased, while prices and margins have fallen. In general, recessions tend to influence the demand for labor only

OCR for page 29
Computing Professionals: Changing Needs for the 1990s temporarily. However, what has been unusual about the recent recession is that it was the first to have had a significant negative impact on demand for computing professionals, shattering the perception that these occupations are immune to business cycles. Elizabeth Nichols, director, Systems and Software Education at IBM Corporation, explained at the workshop that the recession has motivated companies to scrutinize the value they have received from investments in information technology. As a result, they have become more conservative about making new investments in the technology and in the people associated with it. These observations are echoed by reports in trade journals and the business press about the slowdown in orders for new systems, consolidation of information systems (internal application development and support) units, and layoffs by manufacturers and other vendors. The volume of cutbacks has led some observers to speculate that a more fundamental, structural contraction has been taking place, suggesting that many of the jobs eliminated over the past few years will never be restored. Such an inference is qualified, however, by the observations that contraction in the computing sector, per se, has occurred among those segments associated with older approaches to computing and that segments associated with newer approaches, particularly with smaller, distributed systems, have been growing. Also, on the user side, the overhaul of corporate and public computing infrastructure has fed product demand. Longer-Run Factors A number of factors led workshop participants to expect slower growth in demand for computing professionals even after the economy has recovered. Some of these factors may contribute to the structural changes alluded to above. Shifting National Priorities Joe Kubat noted that the impact of the recession has been amplified by defense cutbacks5 and by structural consolidation in the financial services and other industries in the northeastern United States, traditionally a region replete with computer specialist job opportunities. Several computer companies based in New England have contracted, merged with others, or gone into bankruptcy. This combination of events may, Kubat suggested, produce significant and enduring declines in demand. Consolidation may also promote a shift in the mix of job opportunities; it is particularly likely to lower demand for

OCR for page 29
Computing Professionals: Changing Needs for the 1990s TABLE 3.2 Civilian Employment by Occupation, 1990 and Projected 2005, and Change from 1990 to 2005     Number Employed. (Percent of Total) Change, 1990 to 2005 Major Occupational Group Detailed Occupational Groupa 1990 2005 Number Percentage Executive, administrative, and managerial occupations 12,451 (10.2) 15,866 (10.8) 3,415 27.4 Professional specialty occupations 15,800 (12.9) 20,907 (14.2) 5,107 32.3   Computer, mathematical, and operations research analysts (including actuaries, statisticians, operations research analysts) 571 987 416 72.9   Systems analysts and computer scientists 463 829 366 79.0   Mathematicians and all other mathematical scientists 22 24 2 9.0   Engineers 1,519 1,919 400 26.3   Electrical and electronics engineers 426 571 145 34.0   Physical scientists (including geologists, geophysicists, oceanographers, meteorologists) 200 241 41 20.5   Chemists 83 96 13 15.6   Physicists and astronomers 20 21 1 5.0   College and university faculty 712 846 134 18.8 Technicians and related support occupations 4,204 (3.4) 5,754 (3.9) 1,550 36.9   Engineering and science technicians and technologists 1,327 1,640 313 23.5   Electrical and electronics 363 488 125 34.4   Science and mathematics 246 305 59 24.0   Computer programmersb 565 882 317 56.1 Marketing and sales occupations 14,088 (11.5) 17,489 (11.9) 3,401 24.1 Administrative support occupations, including clerical (including 665,000 computer and communication operators) 21,951 (17.9) 24,835 (16.9) 2,884 13.1 Service occupations 19,204 (15.7) 24,806 (16.9( 5,602 29.2 Agriculture, forestry, fishing, and related occupations 3,506 (2.9) 3,665 (2.5) 159 4.5 Precision production, craft, and repair (including 530,000 electrical and electronic equipment mechanics) 14,124 (11.5) 15,909 (10.8) 1,785 12.6 Operators, fabricators, and laborers 17,245 (14.1) 17,961 (12.2) 716 4.2 TOTAL 122,573 (100.0) 147,191 (100.0) 24,618 20.1 NOTE: Numbers in thousands. Projections for 2005 according to scenario for moderate economic growth developed by the Bureau of Labor Statistics. a Partial listing of selected groups. b Note that the count in this category could include some software engineers as well as lower-level programmers. SOURCE: Adapted from George Silvestri and John Lukasiewicz, ''Occupational Employment Projections," Monthly Labor Review 114 (11), November 1991, pp. 64–94, Tables 1 and 2.

OCR for page 29

OCR for page 29
Computing Professionals: Changing Needs for the 1990s BOX 3.2 Observations on Demand for Computing Professionals "My conclusions for the information technology area are that the human resource is growing, but it is growing at a slower rate than it has grown in the past. The new jobs are going to be in the new companies, at the rapidly growing smaller companies, and then the larger companies are going to be focused on the systems integration and systems management problems that have evolved with the rapid growth that they have seen in the past."—Elizabeth Nichols "I think one of the things that we see here is the possibility that we are no longer, in computer science, in a position where the supply falls short of the demand."—Paul Young "[S]upply is not a problem . . . in the commercial business, where the level of intellect may not be as high [as in] research . . . and, therefore, it is possible for us to develop people . . . ."— Don McLean "If what we see in our crystal ball is correct, . . . we would see much less demand for what I will call the low end or what we have traditionally given to entry-level people."—Linda Pierce a variety of individuals involved in applications and systems development (although other factors may increase that demand). Globalization of Markets and Production The globalization of markets can affect a broad range of employment opportunities, including those for individuals involved in developing computer systems. On the one hand, growth in exports or cross-national ventures, including the opening of new markets (e.g., Eastern Europe), may create new opportunities for computing professionals. On the other hand, the development of overseas markets inevitably implies a growing role for foreign citizens, and international competition in all markets creates pressures for cost containment that could slow growth in U.S. employment. For example, interest is growing among U.S. firms—both producers and users of computer systems—in employing software developers based in other countries.6 The motivations for this trend, referred to as "outsourcing," include the lower cost of using software developers overseas and the desire to service foreign markets with local personnel. Noted Linda Pierce, manager, Staff Development

OCR for page 29
Computing Professionals: Changing Needs for the 1990s and Training at Shell Oil Company: ''If you can change the way you package work so that you concentrate on the design component . . . and outsource it overseas, you can get people in India and in Manila and other places . . . with graduate-level degrees . . . who will do this work, too . . . ." Reliance on foreign programmers will affect the mix as well as the level of job opportunities in the United States. Particularly susceptible to displacement are lower-level programming jobs; potential impacts on employment of software developers are uncertain. Jane Siegel underscored the need to monitor and better understand the offshore outsourcing of software development. She reported that her team's attempt to understand trends in offshore software development started with an examination of half a dozen countries and has grown to address activity in at least 23 countries. Her general observations were reinforced by comments from other participants about their own firms' efforts to use foreign programmers. Nichols contrasted the high levels of demand for computing professionals among developed nations (the United States, Canada, Japan, and Western European nations) with the oversupply of such professionals in certain developing nations, notably India and China: "The estimates I have seen are that there are about 100,000 computer scientists in India, and about 200,000 in China. China has jobs for about 10,000." Nichols added that international cooperation may affect the whole product cycle, in that a global division of labor may be invoked:7 If we take a look at what the global demand is, what we are going to see are products developed in more than one country. We can already find examples of that today. The research may be done in one country, the development in another. Someone integrates [a product] or puts the components together and assembles it. Then this product can be sold worldwide under multiple logos or brand names and it will have a wide market. New Economic Activity Peter Freeman, dean of the College of Computing at Georgia Institute of Technology, observed that consolidation in some sectors could be accompanied by the appearance of new jobs for computing professionals in others, commenting that "[A]ll . . . the activities that corporate America is shedding are going to show up in the economy in other ways." The most obvious shift arises from outsourcing. Increasingly, major computing-and communications-using organizations have been purchasing services such as data processing or software development in lieu of employing workers to undertake them

OCR for page 29
Computing Professionals: Changing Needs for the 1990s internally. Nichols pointed to systems integration as a growing arena for outsourcing because of its requirements for specialized skills: Because of . . . the heterogeneous environment of the many vendors and the communications, . . . systems integration is expected to grow at about 22 percent over the next 5 years. Systems management is expected to grow at about 17 percent. A lot of companies aren't able to do that themselves. So, they are turning to vendors, who have some expertise in how to tie these together. [Such an approach] won't always show up as additional human resources within the [current information systems] community. It will show up as new service companies . . . entering the [market]. The trend toward outsourcing has supported employment growth among a variety of service providers and consulting and contract-labor organizations; workshop participants agreed that it is likely to continue, at least in the near term. Harder to predict are entirely new forms of economic activity that would stimulate demand for computing professionals. It is relatively easy to identify where contraction is taking place, but harder to forecast new sources of growth, many of which traditionally emerge in small businesses. New businesses demanding computing professionals may arise both within the computer sector and elsewhere in the economy. CHANGES IN COMPUTER-BASED TECHNOLOGY AND APPLICATIONS AFFECTING DEMAND Changes in the nature and application of computer-based technologies are having a profound effect on the level and mix of demand for computing professionals. These changes are fundamental to declines in some elements of the computing business and increases in others. Recent Trends and New Directions Perhaps the most obvious trend is the shift away from large, centralized computing systems used largely for "crunching" numbers and generating management reports and attended to by teams who took care of the equipment and served as intermediaries between the consumers of computing (the so-called end users) and the computers themselves. The last decade has seen a strong movement to a much larger number of smaller, personalized computing systems that are increasingly interconnected and easier to use. Paralleling this trend is the increase in embedded computing capability in equipment of all

OCR for page 29
Computing Professionals: Changing Needs for the 1990s types—devices and systems for communications, entertainment, transportation, manufacturing, and so on—and the marriage of computing with telecommunications, video, and other technologies. These trends broaden and enlarge the mix of people who work with computer systems in any capacity. Employment opportunities reflect the proliferation not only of computing products, but also of applications. Applications are being introduced for virtually every activity undertaken by a business, from marketing (electronic catalogues, point-of-sale data collection) to sales (optical scanning "cash registers," electronic data interchange for order processing, portable systems for field sales representatives) to technical and professional functions (computer-aided design and engineering, knowledge and expert systems for underwriting) to strategic planning (simulation and forecasting systems) and management (labor scheduling, decision support). William Gear, vice president of Computer Science Research at NEC Research Institute Inc., pointed out that in the future there will be even more types of applications. In particular, he suggested, "The collection of data will become increasingly important . . ., as will [their] distribution . . . and publication." Paul Maritz, vice president for Advanced Operating Systems at Microsoft, explained that the proliferation of applications goes hand in hand with an evolution in software from products best suited to specialized users to products aimed at laypeople (Box 3.3). The magnitude of the potential impact was characterized by Freeman: My hypothesis is that very broad and basic application of computer technology and of computer science principles is going to drive a demand curve that is much beyond anything that we currently see. Linear projections aren't going to apply. Here in California, for 30 or 40 years now, freeways have been put out into virgin land [with] about one exit or entrance every two miles, and then houses and industries spring up around those, and then the freeways are overcrowded and there aren't enough exits, and so on. I think the same thing is going to happen in the area of computing. This view was echoed by Maritz, who referred to the emergence of an information infrastructure with substantial but unknown implications: [I]n the next 10 years, the infrastructure will be laid down to . . . bring digital information, initially in the form of movies, into your house, but the fact that you can move data at the rate of gigahertz per second around the world into your home, . . . [and] the implication of that . . . in terms of products, is a fascinating one . . . which whole new companies and industries could be built on.

OCR for page 29
Computing Professionals: Changing Needs for the 1990s shrink as more computer science and engineering Ph.D. holders move into academic positions. Senior members of the academic computer science and engineering research community have expressed concern about the recent weakness of overall demand in the academic labor market. Historically, it was very easy for a new computer science or engineering Ph.D. recipient to secure a job at a top academic department, but maturation of those departments, declining enrollments, and constraints on research funding have diminished those opportunities.)12 Anecdotal reports indicate that universities are now receiving several hundred applications for posted positions (and may receive tens of applications or more even when positions are not posted), with applicants including new Ph.D. holders, recent Ph.D. holders previously unable to obtain research positions, individuals who fail to get tenure, and individuals from sister fields, such as electrical engineering or mathematics. The result is a crisis in expectations: new Ph.D.s seeking academic employment must reorient themselves to lower-tier universities or to nonresearch institutions, the 4-year (and potentially 2-year) colleges, where openings do exist. New Ph.D.s may also have to consider options in industry or government to a greater extent than previously. Although industry's interest in higher skill levels provides encouragement, Ph.D. preparation traditionally has encouraged students to favor academic employment. A similar situation exists in physics and mathematics, which face added pressure from the new availability of talented individuals from the former Soviet Union.13 Another effect of weak demand in the academic labor market is the likely inhibition of the movement of senior computer scientists from industry to academic positions, due to crowding out by younger and less highly paid Ph.D. holders. A diminished flow of personnel from industry to academia would further increase the need for interaction and communication between the two sectors. Three factors may improve prospects in the academic labor market in the mid-1990s. First, based on demographic projections, the student population is expected to expand at that time.14 Second, although computer science and engineering faculty are younger than faculty in other sciences and engineering, there will be some retirements.15 Third, the federal High Performance Computing and Communications (HPCC) program, if fully funded, should begin to have some effect; it is expected to stimulate computer science and engineering research, and thus stimulate Ph.D. production. However, questions arise as to where such new Ph.D.s may be employed. Some senior computer scientists have expressed concern that the HPCC program might increase the supply of Ph.D.s without stimulating corresponding increases in demand.

OCR for page 29
Computing Professionals: Changing Needs for the 1990s Some researchers have advocated postdoctoral positions as a means to provide a practical transition from Ph.D. studies to research jobs and to regulate the flow of Ph.D.s into the labor market. Postdoctoral associates are the norm in some fields (e.g., biology), building on traditions in those fields for developing experience, but they have not been common in computer science and engineering.16 There was limited discussion of this concept at the workshop. Gries, for example, argued that ''a postdoctoral program with 200 more people in it would provide the field with 200 more full-time researchers, which it needs. More importantly, an extra year or 2 of full-time research after the Ph.D. would allow the Ph.D. to mature and gain experience." A postdoctoral program would also provide an opportunity for new researchers to pursue interdisciplinary work, which is of growing interest in academia and especially in industry. Others, however, argued that the direct impact of labor market conditions is needed to regulate supply and demand for Ph.D.s. According to Zelkowitz, "As academic positions dry up or good ones become harder to find, individuals will make hard choices—a $45,000 assistant professor position at a small 4-year college versus a $25,000 postdoc position at Massachusetts Institute of Technology or Carnegie Mellon University with the chance of a good position there or elsewhere in 2 or 3 years. Different individuals will make different career decisions." The literature on postdoctoral appointments suggests that they, too, are driven in part by labor market conditions. The number of Ph.D.s holding postdoctoral appointments rises when employment demand is weak and falls when demand is strong, suggesting that such appointments are used to tide new degree recipients over during periods of limited job opportunities.17 However, if demand remains weak over a long period of time, postdoctoral appointments may simply delay the inevitable, postponing the point at which individuals must make hard choices about their careers. In the biomedical research fields, in which postdoctoral experience is a recognized requirement for obtaining attractive faculty appointments, long postdoctoral apprenticeships are the norm. Moreover, many who have held such apprenticeships eventually wind up in soft-money (nontenured) university positions or in nonacademic employment, generally in industrial or federal laboratories.18 Industrial Research Industry laboratories, like universities, employ people with a mix of backgrounds to do computer science and engineering research. Unlike academic researchers, researchers in industry do not necessar-

OCR for page 29
Computing Professionals: Changing Needs for the 1990s ily have Ph.D.s. In industry, the mix seems driven both by the availability of senior computer scientists and engineers and by interest in multidisciplinary projects. Robert Kraut estimated that about 50 percent of the active professional Ph.D.-level computer science researchers at Bellcore lack computer science Ph.D.s and that they "range from mathematically oriented scientists . . . physicists who are doing computer science research now, to people who are not mathematically trained but who are doing research on one of the [Bellcore] core areas, like human-computer interaction." Unlike the more senior faculty at a typical Ph.D.-granting university, explained Kraut, the mix at 9-year-old Bellcore reflects particular decisions on the hiring of new Ph.D.s. These individuals appear to maintain links with both computer science and the other fields in which they are trained, attending two sets of professional meetings, for example. Paula Stephan, of the Policy Research Center and Department of Economics at Georgia State University, remarked that the flexibility in professional focus at Bellcore described by Kraut is consistent with observations about employed Ph.D. holders overall. Data show "an enormous number of Ph.D.s in science in the United States who say they are working in different areas in different years" when they are surveyed. For example, the Survey of Doctorate Recipients from United States Universities shows that 77 percent of people working as computing professionals (people working in computer science or computer engineering) have degrees in fields other than computer science or computer engineering, while 13 percent of degree holders in computer science or computer engineering are working in occupations other than computing professions. However, Lucy Suchman suggested that individuals with Ph.D.s in computer science might be somewhat less flexible than those whose Ph.D.s were earned in other fields, in part because some employing institutions may not encourage flexibility. Whereas industrial research in computing has been largely applied (with the exception of some activities in larger companies such as IBM or Digital Equipment Corporation), some new interest in basic research is emerging in companies that believe it is necessary to enhance technology and enlarge the market. This point was made by Paul Maritz of Microsoft, who contended that the broadening of the applications base, which is linked to improvements in the sophistication and ease of use of computer-based technologies, will motivate new research as well as new development activities: [F]or the first time in our company, we have recently opened up a pure research department . . . which is what you need . . . if you want to expand your marketplace out into a broader set of people.

OCR for page 29
Computing Professionals: Changing Needs for the 1990s . . . [W]e are looking not only for people who have inventive skills—who know how to take established techniques and think up novel applications of those techniques—but also, in very small numbers, for pure research people . . . . We are looking at which computer science programs are preparing theoretical people, whose ideas won't necessarily go into a product immediately. William Gear and Elizabeth Nichols expressed similar views, also suggesting that there will be a greater demand for higher-level skills. Expansion into basic research by companies like Microsoft will offset, in part, the diminution of such research efforts at larger companies, such as American Telephone & Telegraph or IBM, which appear to be emphasizing more applied and more focused research in response to market pressures. However, the net effect on demand for researchers is unknown, because of the difficulty in obtaining reliable data on industrial hiring of researchers. Based on their own and their students' experiences, academic researchers in computer science and engineering have expressed concern that industrial research is contracting and that additional small endeavors cannot compensate for contraction in the larger laboratories. Applications and Systems Development The development of computing applications and systems dominates the employment of computing professionals in industry, government, and other organizations. It is a heterogeneous activity and, therefore, is difficult to describe and label. Systems development increasingly revolves around software and its many uses—in individual pieces of equipment, in networks and for integrating diverse components and subsystems into larger systems, for implementing and accessing databases, and for making a variety of tasks easier to do. The number and variety of software types are growing. In particular, development increasingly involves communications technology or applications, due to the spread of distributed systems, the development and broadening application of telecommunications infrastructure, and the blending of computing and communications technologies and systems. Corresponding to the growth in applications, there appears to be a shift among companies from investing in hardware-oriented talent to investing in software-oriented talent. Professional activities include designing, developing, and maintaining (i.e., enhancing and modifying) software; so-called coding, which involves less skill, is done by programmers that are increasingly considered technicians. There is a growing array of computer-based tools to help automate software development processes.

OCR for page 29
Computing Professionals: Changing Needs for the 1990s BOX 3.6 Skill Requirements for Development "We are not in the business of. . . research. We are in the business of building things. As long as we build things with some good methodologies, we don't need a tremendously high level of education and training. So, I don't see a problem even down the road with supply."—Don McLean "[W]e may be less sensitive to the skill levels, as far as computer science goes, because we don't do research and rarely break new ground in computer science. We often break new ground in applying old technology to new problems."—Gordon Eubanks "In developing a spreadsheet or a word processor, you can't really say there is any computer science done there . . . . [I]t . . . was appreciation of what could be done using these techniques, applying them to some problem outside of the traditional computer arena . . . . [I]nventiveness in that sense is different from research."— Paul Maritz Despite its apparent importance, the work of software engineers and what it implies for demand are poorly understood, according to Jane Siegel: "There are very few places where one can get any consensus about what the specific education, background, and/or job performance skills are for someone whom you would call a software engineer. In fact, three states in the country now have legislation on this subject,19 and it is a growing issue." Workshop participants from industry emphasized that applications and systems development does not necessarily require advanced training in computer science and engineering (Box 3.6). Nevertheless, an increase in skill requirements for these functions was reported. Gear argued that at least a bachelor's-level education appeared necessary: "I see very little hope for people . . . at the associate level in this area, and even the B.S. is probably minimal. These are the people who are developing the operating systems for the future, the packages, the spreadsheets, the compilers and so on . . . and, of course, the machine designers." According to Linda Pierce, a primary motivation for using more highly skilled people is the concern about increasing productivity. She noted that computer-aided software engineering (CASE) technology will eliminate "drudgery work" in code generation, reducing demand at the low end, the work traditionally given to entry-level

OCR for page 29
Computing Professionals: Changing Needs for the 1990s people. With the help of a ''technology infrastructure," she posited, "we would see a shift toward the higher end of the technical skills of the people that we bring on board . . . ." Applications and systems development is driven by the needs and wants of users, and both users and producers of computer systems employ development specialists. This point was made by Chris Caren, vice president of science and technology at Lockheed, who described the fruitful interaction of a small number of trained computing professionals with a larger number of people from other fields who needed computing tools (Box 3.7). The emphasis on use and users encourages the assembly of multidisciplinary development teams. BOX 3.7 Applications as Drivers of Development "[A] lot of the real developments [are] user driven . . . . [O]ur company developed a program . . . called CADAM, which is a mechanical design tool. That was done something like 15 years ago and it was done because of the L-1011 [aircraft] program . . . . That [effort] did not involve computer science people. Those were applications-type people, who learned the tricks of the trade. "In the case of the Stealth aircraft . . . most of [the] programming was done by teams of aerodynamics and electromagnetic people, . . . using new massively parallel machines . . . . [T]hat allows us to . . . not only do better designs, but also [to go] through design optimizations . . . . [A] lot of the design optimization [came] over from the computer science domain. ". . . [W]e seeded [our Al efforts] with a few computer science people, but those have been mainly self-taught, again with applications people coming in and being taught the selective methods required . . . to handle their type of problem. In fact, the place where the computer science people have probably made the largest impacts . . . has not really been . . . in computer science. [It] has been in . . . the work that we are doing in advanced software methods . . . . [W]hat we have been pushing there has been not only software reuse, but also automated programming. And, again, that has come from the computer science. "So, . . . if I were to look at the . . . major developments we have had over the last decade or so . . . it has been highly program-need driven, and that program side of things has been handled more by domain specialists . . . . [O]n the other hand, for the advanced tool [project], as I indicated with reuse, automated programs, and some of the Al stuff, that has come from the computer science [specialists]."— Chris Caren

OCR for page 29
Computing Professionals: Changing Needs for the 1990s And it has promoted the development of computing applications by users and others not originally trained to be computing professionals. New categories of users are emerging in parallel with more powerful and easier to use systems, generating new kinds of applications and new needs for development. Joe Kubat, for example, described a new role in the finance industry for so-called "rocket scientists" with "an R&D-type of . . . technical understanding and expertise," who are highly numerate and may develop and use algorithms for supercomputers and other sophisticated tools to develop new financial products. Maritz generalized by saying, "[T]he fact that computers are much more affordable now makes it more cost-effective to apply analytical techniques to more and more problems." Applications and Systems Deployment The proliferation of computer systems and applications has been accompanied by a need for more people to facilitate their deployment. These individuals range from the "applications engineers" and customer support personnel employed by vendors to a range of individuals within system-using organizations, including but not limited to people in information systems departments. Some of the work involves integration of hardware and applications to meet user needs. This category of computing professional is perhaps the least well understood; it is at least as diverse as the applications and systems development category, and it potentially spans a mix of jobs that may be performed by professionals and/or technicians. In this category, too, skill requirements appear to be increasing. For example, according to an article on vendor support hot lines (which constitute an application of computing and communications technologies to support applications of computing), "Entry requirements are changing to reflect a greater emphasis on service and the increased technical complexity of products . . . . "20 As a definable group, however, this set of occupations is beginning to coalesce because the need is growing for the tasks involved in applications and systems deployment and because those tasks and the jobs that comprise them are becoming more professionalized. Thus, this is an area of work that bears watching, in the interests of better defining and measuring employment opportunities and trends. William Gear characterized a range of support functions according to their requirements for education and training: In the support area, network administrators, system administrators, people who can provide routine maintenance, the sort of people that a modest-size [company]—by "modest," I mean a company with 10

OCR for page 29
Computing Professionals: Changing Needs for the 1990s to 50 people—might need . . . to keep a system running at some level, who know when to take [a piece of equipment] down to the computer repair shop and when to just switch a board, on up through a somewhat higher level [—e.g. a lead systems administrator—] requiring a degree in . . . system administration . . . probably will go through a more traditional educational program, such as computer science, computer engineering, management of information systems . . . on the way to that path. The need for people skilled in deployment of computer-based systems is expected to grow with the dispersion of intelligent systems and applications across organizations and out of conventional information systems units. As Elizabeth Nichols explained, the dispersion phenomenon may have been launched by people who were technically literate (including scientists and engineers), but increasingly, computer systems are being used by everybody, and notably by people who, unlike pioneering PC users, may not be in a position to do their own troubleshooting. The sheer magnitude of the distribution of equipment to end users will increase the level of system support activity (including training and maintenance), while the increasing sophistication of applications and associated technology may increase both the amount and the quality of system support required as part of deployment (especially training and facilitation). The need for deployment work is not always recognized and staffed properly within organizations. In particular, the requirements for professional-level skills, training, and judgment may not be recognized up front by employers. In Appendix C Nichols cites the example of the proliferation of LANs and LAN-based applications: One estimate is that one person is needed to install, upgrade, move equipment, and resolve problems for each 25 workstations or personal computers on a LAN. Very often an undertrained individual in the end-user department is designated to do the work; in this case it is likely that the resource is never counted in the information technology totals. In addition, undertrained personnel make mistakes that often create large problems for information technology personnel to resolve. Further complicating the assessment of demand for personnel skilled in deployment and other areas is the tendency to outsource some of this work to external service firms. Thus companies may track deployment costs but not actual human resource needs. However, job seekers may find these third-party service providers a significant source of deployment-oriented jobs.

OCR for page 29
Computing Professionals: Changing Needs for the 1990s ISSUES AND CONCLUSIONS Crosscurrents are shaping employment opportunities for computing professionals in firms that supply and organizations that use computing technology. The industrial sector supplying computing technology is metamorphosing: opportunities are declining in the segments based on the centralized computing paradigm but increasing in segments oriented to distributed computing—in firms that develop smaller systems and associated software, networking, and so on. On the user side, the consolidation of the larger firms to achieve economies of scale is offset by the diffusion of computing and communications technology across a widening set of organizations. The general consensus among workshop attendees was that there may be net, moderate growth in employment. But focusing on the averages can obscure the magnitude of the shifts taking place in the mix and the nature of job opportunities for computing professionals. Budget tightening that constrains academic research and the decline of large, central industrial laboratories do not bode well for jobs in computer science and engineering research. These developments are part of a larger decline in the conventional research environment, and it is not apparent that we have an attractive model for continuing to meet national needs for research relating to computing and communications. Absent other changes, the research component of the job market may decline. Because occupations and functions are not necessarily identical, it is possible that new job opportunities may arise outside of research that can tap the higher skill levels of computer science and engineering Ph.D.s, thus conveying broad economic benefits. But broader use of these individuals is not the same as expansion of the research capability. Required levels of skill appear to be increasing. Employers represented among workshop participants reported that their skill requirements are increasing in all professional domains—research, applications and systems development, and applications and systems deployment. This circumstance suggests that requirements for formal education and training may increase, although to date nonacademic employers have not consistently demanded formal computer science education. The shift in the mix of skills needed by computing professionals implies a need for education that provides for both flexibility and continuous training. This circumstance must be recognized by educational institutions, employers, and individuals.

OCR for page 29
Computing Professionals: Changing Needs for the 1990s NOTES 1.   Measured demand is also sensitive to the level of aggregation of the analysis. Replacement rates arising from turnover will be higher when individual firms are the unit of analysis; they will be lower when industries are examined, since people can move among firms without increasing industry turnover overall. 2.   Employer surveys are conducted to assess demand, but they do not always produce reliable information. Generally such information is credible for current employment levels and turnover rates. It is less credible for future employment and qualitative dimensions of demand. Such information has become difficult (and/or expensive) to collect. 3.   Even specialists in computer hardware may not remain tied to a narrow set of industries producing computer-based hardware, since such hardware is becoming embedded in a growing range of products. 4.   The number of computing professionals may not seem small compared to the number of scientists, engineers, or other technical personnel. For example, unpublished BLS tables for 1991 indicate that there were 95,000 employed mathematicians and mathematical science college and university teachers, 42,000 employed physicists and physics teachers, and 1.8 million employed engineers and engineering teachers (combined). 5.   Defense cutbacks, however, do not have a clear-cut effect. For example, explained Jane Siegel, many defense systems have a 20-to 30-year lifetime, creating a demand for individuals to maintain and enhance if not create them. Also, defense systems may be associated with a requirement for hiring only U.S. citizens. For an interesting analysis of the employment implications of the recent dramatic events in Eastern Europe, see Andrew Pollock's "Technology Without Borders Raises Big Questions for U.S.," New York Times, January 1, 1992. 6.   In the information technology industry the number of new transnational corporate technology alliances has increased from 348 in 1980–1984 to 445 in 1985–1989. The Minnesota Mining and Manufacturing Company now has 2,500 technical employees stationed abroad, triple the number the company had in 1980. See Pollock, "Technology Without Borders Raises Big Questions for U.S.," 1992. 7.   For an interesting overview of the issue, see Pollock "Technology Without Borders Raises Big Questions for U.S.," 1992. 8.   One example cited by workshop participants was the drop in demand for certain workers associated with mainframe systems, such as "tape hangers" and computer operators; an earlier example might be keypunch operators. 9.   Observed Tora Bikson, "I think that what we are seeing is . . . every level shifting up, and there is not a large bunch of low-skill jobs at the bottom level." 10.   Paul Maritz explained, "[This approach] really is very effective because we get to see these people in more than just a one hour or a half hour interview and we really find those people who are not only good at interviewing, but really are good at working as well." 11.   Note also that some computing-related research can be done in other departments (e.g., interface design in psychology), involving Ph.D.s in both computer science and other fields. 12.   See Computer Science and Telecommunications Board, Computing the Future: A Broader Agenda for Computer Science and Engineering , National Academy Press, Washington, D.C., 1992. 13.   See, for example, American Mathematical Society, Employment and the U.S. Mathematics Doctorate: Report of the AMS Task Force on Employment , American Mathematics Society, July, 1992; Jean Kamagai and William Sweet, "Signs of Tighter Job Market

OCR for page 29
Computing Professionals: Changing Needs for the 1990s     Grow; More than Recession at Work," and "Ph.D.s Fared Worse in 1990 Job Market, Survey Shows," Physics Today, March, 1992, pp. 55–58. 14.   William G. Bowen and Julie Ann Sosa, Prospects for Faculty in the Arts and Sciences, Princeton University Press, Princeton, New Jersey, 1988. 15.   Howard R. Bauer and Jack H. Schuster, American Professors: A National Resource Implied, Oxford University Press, New York, 1986; and Bowen and Sosa, Prospects, for Faculty in the Arts and Sciences , 1988. 16.   According to unpublished tabulations from the Survey of Doctorate Recipients, Office of Scientific Engineering Personnel, the number of Ph.D.s who go on to hold postdoctoral appointments has been increasing in all fields, however. In the physical sciences, for example, the number rose from 2,500 in 1977 to 3,000 in 1989. The comparable numbers for computer science are 40 and 70. 17.   See National Research Council, Postdoctoral Appointments and Disappointments, National Academy Press, Washington, D.C., 1990. 18.   See National Research Council, Biomedical and Behavioral Research Scientists: Their Training and Supply, Committee on Biomedical and Behavioral Research Personnel, Office of Scientific and Engineering Personnel, National Academy Press, Washington, D.C., 1989. 19.   The legislation referred to concerns a current debate about whether software engineers should be certified or licensed, especially when they are producing safety-critical software. 20.   Alan Radding, "Support: A Hot Line to a Computer Career," Computerworld , October 21, 1991, p. 132.