Eric Roberts March 7, 2016

A History of Capacity Challenges in Computer Science

Those who cannot remember the past are condemned to repeat it.
—George Santayana, Reason in Common Sense, 1905

Since the 1970s, the number of students graduating with bachelor’s degrees in computer
science has fluctuated significantly. As shown in Figure 1, computer science degree
production in the United States has experienced two episodes of rapid increase followed
in each case by a precipitous collapse. The first peak occurred in 1986, the second in
2005, and we are once again on a steep upward trajectory, which began in 2009.1

It is important to keep in mind that the number of bachelor’s degrees produced in a

year inevitably lags in comparison to enrollment data. As students reach their junior and
senior years, switching fields becomes less likely. Thus, the peaks in degree production
in 1986 and 2004 reflect changes in student enrollment patterns that occurred two to three
years earlier.

If you look at the graph in Figure 1, the first conclusion that jumps to mind is that
student interest in computer science is cyclical. That interpretation, however, is
insufficient. Most importantly, it fails to recognize the fact that the downturns in the mid

1 Unfortunately, the most recent digest from the National Center for Education Statistics includes data only
through 2012. To offer a more informative picture of the current situation, Figure 1 uses growth rates
recorded in the Computing Research Association’s Taulbee surveys to estimate the broader numbers for
2013 and 2014.
 –2–

1980s and the early 2000s happened for different reasons. The more recent downturn
was clearly caused by the dot-com collapse. After the tech bubble burst in 2001, student
interest in computer science waned throughout the United States, a downturn exacerbated
by a popular mythology suggesting—entirely contrary to fact—that all jobs in technology
were about to be shipped offshore to low-wage countries like India and China.

The earlier collapse in the mid 1980s was very different in its origins. The cause of
that decline was the inability of universities to attract enough faculty to meet growing
student demand. Beginning around 1984, most computer science departments were
forced to limit course enrollments and to restrict admission to the computer science
major. These actions led in turn to a steep decline in degree production a few years later.

In order to make any useful predictions about the likely outcome of the current
expansion, it is essential to undertake a more detailed analysis of the reasons for the
variations in degree production that computer science has experienced in the past. To
understand the history from a national perspective, it makes sense to analyze the three
peak periods independently, which gives rise to the following three questions:

• What happened during the downturn in the 1980s?

• What happened during the downturn in the 2000s?
• What is the nature of the enrollment expansion today?
 –3–

What happened during the downturn in the 1980s?

Who controls the past controls the future; who controls the
present controls the past.
—George Orwell, Nineteen Eighty-Four, 1949

The first boom-and-bust cycle in academic computer science began with a steady rise in
bachelor’s degrees throughout the 1970s, which became more rapid at the end of the
decade. This dramatic rate of increase continued until sometime around 1984, when the
number of students entering the field reached its peak. The peak was followed by a
decline in degree production that eventually flattened out in 1994, when degree
production was down by 42 percent from its earlier high. These statistics are illustrated
in Figure 2, which extracts the relevant years from Figure 1 and adds labels showing the
most likely explanations for the changes in direction.

The rapid increase in student demand at the beginning of the cycle is easy to explain.
The late 1970s and early 1980s saw the introduction of the personal computer, which
brought many more people into contact with computing. With the release of the Apple II
in 1977 and the IBM PC in 1981, a large number of prospective college students had
access to computing for the first time in history. The excitement associated with the
advent of personal computers coupled with the widespread availability of well-paying
jobs in computing drew many students into the field.

The cause of the decline in student numbers that began in 1984 is more difficult to
explain. The excitement that fueled the boom was, after all, still growing. January 1984,
for example, marked the release of the Macintosh, heralded in Apple’s Super Bowl
 –4–

commercial as “the reason that 1984 won’t be like 1984.” Although the overall U.S.
economy experienced a small recession beginning in 1981, that downturn had a minor
effect on the technology sector. In an article published by the Bureau of Labor Statistics
in 1985, economist John Burgan noted that “employment declines in high-tech industries
were not as deep as those in manufacturing” and that, in particular, those companies with
the largest concentration of highly skilled technical workers were the only ones that
outperformed the rest of the economy.2 There seem to be no economic or technical
reasons to explain a collapse of student interest beginning in 1984.

If one looks closely at the downturn of the 1980s, however, it quickly becomes clear
that the reasons for the collapse in student enrollments had nothing at all to do with
student interest. Student demand for computer science courses and degrees remained
high throughout that period. Students in the mid 1980s did not decide against majoring in
computer science but were instead prohibited from doing so by departments that lacked
the resources to accommodate them.

I believe that what happened in the 1980s is best described as a capacity collapse in
which universities and colleges were simply unable to satisfy the growing level of student
demand. Departments tried a number of strategies to increase their teaching capacity,
including retraining faculty from other disciplines and hiring adjunct faculty from
industry. In the end, however, demand overwhelmed capacity, and colleges and
universities were forced to restrict admission to the computer science major, which gave
rise to the subsequent downturn.

The sections that follow examine the history of this capacity collapse in more detail.

The capacity collapse of the mid 1980s

The first capacity collapse in computer science occurred around 1984, now more than 30
years ago. The passage of time, coupled with the fact that a more recent collapse
occurred for different reasons, means that few people today understand the pressures that
departments of computer science experienced during those years. That loss of historical
understanding is particularly unfortunate because the problems we see in computer
science education today closely resemble those from the beginning of the 1980s.

The challenges facing computer science in the 1980s were widely recognized at the
time. Rising enrollments and the shortage of qualified faculty were a central focus of the
fourth Snowbird Conference in 1980.3 The discussions at Snowbird led to a report
entitled A Discipline in Crisis, which was published in the June 1981 issue of
Communications of the ACM. That report begins with the following sentences, which
offer a succinct review of the problem:

2 John Burgan, “Cyclical behavior of high-tech industries,” Monthly Labor Review, May 1985.
3 The Snowbird Conferences are a biennial gathering for the chairs of computer science departments in
research universities. These conferences are sponsored by the Computing Research Association, which was
called the Computer Science Board prior to 1990.
 –5–

There is a severe manpower crisis in Computer Science. There are acute

shortages of well-trained computer people at all levels, especially the Ph.D.
level. The Ph.D. shortage is especially serious because it threatens our ability to
conduct basic research in Computer Science and to train the next generation of
computer experts.

The report goes on to outline the problems faced by the 83 Ph.D.-granting institutions
included in the Taulbee surveys. All participants agreed that finding faculty to satisfy the
growing demand was a critical challenge. In 1979, for example, American and Canadian
universities produced only 248 Ph.D.s in computer science. The report then noted that
“fewer than 100 of these Ph.D.s chose academic careers, and they had over 650 academic
positions from which to choose.” In other words, there was approximately one applicant
for every seven advertised positions, at least in terms of the new-Ph.D. pipeline. Six of
those seven positions would either go unfilled or be offered to a candidate with less
educational preparation or a degree in another field.

In a later section, A Discipline in Crisis offers the following description of the

pressures on existing faculty:
Pressures on faculty are intense. In the United States, Ph.D. Computer Science
faculty have grown from 805 in 1975 to 837 in 1979—virtually no growth. The
undergraduate student demand for Computer Science has risen at 15 percent to
20 percent annually during the same period.
Thus overburdened, faculty cannot find adequate time to conduct research or
to supervise graduate students in research. This atmosphere is a strong incentive
for research-oriented faculty to seek positions in industrial research groups.
Departments must find ways to give faculty more time for exploring new ideas
with their graduate students while continuing to fulfill teaching commitments.
Limiting or cutting back enrollments would be counterproductive given the
societal need manifested in the rising enrollments. The only way in the long
term to meet this need is to train, hire, and retain new faculty.

Although the numbers today are of course much higher, reading this assessment from the
early 1980s creates a clear impression of déjà vu.

In addition to the report on A Discipline in Crisis, the June 1981 issue of

Communications of the ACM included a letter from ACM President Peter Denning
entitled “Eating our seed corn.” Although Denning did not introduce the seed-corn
metaphor—and indeed says in his President’s letter that “the phrase ‘eating our seed
corn’ appears everywhere”—he certainly helped to popularize it and bring the issue
before a larger audience. He cites in particular an article in the Business Week issue of
November 17, 1980, which charges that
Industry is eating some of its own seed corn. Not only are they taking students
who would become faculty, they are recruiting faculty.

The community’s awareness of the looming capacity crisis deepened over the next few
years. The Snowbird Conference in 1982 led to the preparation of a new report entitled
Meeting the Crisis in Computer Science, which appeared in Communications of the ACM
in December 1983. Although this follow-on report identified some encouraging signs, it
concluded that “the basic critical situation had not yet been ameliorated. Ph.D.s in
 –6–

computer science are still being produced at about 250 per year, while the demand is still
about five times that. The number of undergraduates entering computer science
departments continues to increase, and the number of unfilled computer science faculty
positions is greater than in 1980.”

Response from governmental agencies

One of the encouraging signs identified in the report from the 1982 Snowbird Conference
was increased awareness by government agencies of the problems facing academic
computer science. In October 1980, the Department of Education and the National
Science Foundation released a report entitled Science and Engineering Education for the
1980s and Beyond, which highlighted the faculty shortfall throughout engineering and
computing fields.
There are, today, severe shortages of qualified faculty members in most fields of
engineering, as well as in the computer professions. Industries have expanded
their research and development efforts and have increased the rate at which new,
sophisticated products are introduced. To effect this, they are luring faculty
members away from the universities into challenging well-paid positions. At the
same time, they are making such attractive job offers to bachelor’s degree
recipients that many who would once have gone to graduate school now opt for
positions in industry. The net effect has been a reduction in the ability of
universities to provide education in engineering and the computer professions,
although undergraduate demand for these areas is more intense than ever.
Unless the problem of faculty erosion is alleviated, it is possible that many
engineering schools and departments that educate computer professionals may
have to reduce their enrollments during this decade, thereby reducing the
numbers of trained people in these fields that the Nation’s future requires.

The last sentence of this paragraph raises the specter of precisely the sort of enrollment
caps that computer science departments were forced to institute beginning around 1984.

In addition to focusing government attention on the problem of faculty shortages in

computer science and other applied fields, one of the important contributions of the
DoE/NSF report was that it introduced economic analysis into the debate. As the report
notes, it is usually possible to allow market forces to correct labor imbalances, given that
an increase in job opportunities attracts more people to that sector. The report argues
against that course of action as a strategy for correcting the imbalances in technical fields,
saying:
While market forces may ultimately relieve current and future shortages, we
believe that the innovative capacity of American industry will be severely
hampered in the interim. We simply cannot afford to wait for the slow workings
of the marketplace.

Over the next few years, the National Science Foundation continued to assess the
problem of shortages in key technical specialties, including computer science. In 1982,
NSF staffer Kent Curtis presented a report to a meeting of the Computing Research
Board on the labor shortages facing computer science. I believe that the insights Curtis’s
report offers for the situation in the 1980s are of such direct relevance today that I have
scanned his report (previously available only in an extremely poor photocopy) and made
 –7–

it available on my web site under its original title, Computer Manpower—Is There a
Crisis? Curtis’s report argues that academic computer science faces special challenges.
Let us consider the conundrum facing the computer field in higher education
first. It is experiencing an exponentially increasing demand for its product with
an inelastic labor supply. How has it reacted? . . . 80% of the universities are
responding by increasing teaching loads, 50% by decreasing course offerings
and concentrating their available faculty on larger but fewer courses, and 66%
are using more graduate-student teaching assistants or part-time faculty. 35%
report reduced research opportunities for faculty as a result. In brief, they are
using a combination of rational management measures to adjust as well as they
can to the severe manpower constraints under which they must operate.
However, these measures make the universities’ environments less attractive for
employment and are exactly counterproductive to their need to maintain and
expand their labor supply. They are also counterproductive to producing more
new faculty since the image graduate students get of academic careers is one of
harassment, frustration, and too few rewards.

Public perceptions of the challenges

The problem of faculty shortages in computer science also received coverage in the
media. In February 1981, The Chronicle of Higher Education ran an article entitled “As
students flock to computer science courses, colleges scramble to find professors.”4 The
article quotes Joe Turner (still a leader in the computer science education community) as
follows:
“It’s an impossibility to fill faculty positions,” says A. Joseph Turner, head of
the computer science department at Clemson University. He says he has to
compete with other universities, as well as industry—but the industrial
competition is by far the toughest.

A month later, The Chronicle of Higher Education followed up the earlier article with
an essay by Stanley Pogrow at the University of Arizona in which he points out that the
situation facing several applied disciplines is new in the history of academia.5
In previous times, fields that were experiencing rapid expansion of knowledge
generally found it easy to attract new faculty members, and fields where jobs
were plentiful found it easy to attract graduate students. This is no longer true.
A number of fields in applied science, such as computer science, physics, and
electrical engineering, where knowledge frontiers are being rapidly extended,
are experiencing increasing numbers of unfilled faculty positions, a reduced
aging faculty, and declining graduate enrollment.

The response from academia

Faced with the extraordinary challenge of finding faculty in a labor market in which the
number of positions exceeded the number of applicants by as much as a factor of seven,

4 Jack Margarrell, “As students flock to computer science courses, colleges scramble to find professors,”
The Chronicle of Higher Education, February 9, 1981, page 3.
5 Stanley Pogrow, “In an information economy, universities and business compete for workers,” The
Chronicle of Higher Education, March 16, 1981, page 64.
 –8–

universities and colleges were forced to adopt other strategies to build their teaching
capacity. These strategies included

• Increasing teaching loads and class sizes

• Hiring more part-time and adjunct faculty
• Retraining faculty from other disciplines
• Limiting enrollments and access to the major

All these strategies are described in papers presented at the leading conferences in
computer science education at the time. The strategy of increasing teaching loads is
self-defeating, as indicated in Kent Curtis’s admonition that such measures “make the
universities’ environments less attractive for employment and are exactly
counterproductive to their need to maintain and expand their labor supply.” Hiring
part-time and adjunct faculty was at best a short-term solution that proved difficult to
implement. Given the shortage of computing talent in the industry, adjunct faculty were
also in short supply.

The strategy that had the most significant long-term effect on computer science
education was faculty retraining. Starting in the early 1980s, a number of universities
including the University of Massachusetts, the University of South Carolina, Ohio State
University, Kent State University, the University of Evansville, Brooklyn College,
Clarkson University, California State University at Fresno, Central State University in
Oklahoma, Memphis State University, and James Madison University began to offer
programs to retrain faculty from other disciplines to teach computer science, at least at
the introductory level. These programs are described in an article that appeared in The
Chronicle of Higher Education in July 1984, which begins as follows:6
Nearly 400 faculty members in mathematics, physics, chemistry, and a host of
other disciplines—scientific and nonscientific—are going to colleges and
universities this summer to learn to teach computing.
Some of them see retraining in computer science as an opportunity to move
into an exciting, growing field. Others are getting formal training in courses
they already teach. Still others are going because they recognize—or have been
told—that the future in their present fields is bleak.

The Chronicle article found that the length of the faculty training programs varied
from a high of three part-time years to a low of two weeks. The latter figure “raised
eyebrows” and prompted the author of the article to ask the question, “Can two weeks be
as effective as three years in training a faculty member to teach a computer course?”
William Weber, chairman of the computer science department at Southeast Missouri
State University, admitted that such short programs could of course not be as complete
but added that universities faced no other choice. As the article describes,
None of the 13 faculty members in his department have a doctorate in computer
science. “We couldn’t afford them if they did,” he says. Instead, the university
made a commitment to retraining.

6 Judith Axel Turner, “Growing demand for computer courses spurs retraining of college teachers,” The
Chronicle of Higher Education, July 11, 1984, page 23.
 –9–

Although faculty with formal training in computer science remained dominant at the
research universities, faculty members from outside the field, often with minimal training
in computer science, soon filled most of the positions in less prestigious universities and
liberal arts colleges. As I wrote in an article that appeared during the enrollment boom of
the late 1990s (and which reviews the strategies used during the 1980s in more detail than
I do in this section), I am convinced that “academic computer science could not have
survived were it not for the willingness of some faculty to move to a new field. For the
most part, those who migrated to computer science were extremely conscientious about
acquiring the expertise they needed to teach in their adopted discipline. Their efforts
sustained computer science education at many institutions and helped reduce the impact
of the earlier crisis.”7 Even so, the fact that so many computer science faculty came from
outside the field had profound implications for academic computer science that continued
through the next enrollment cycle.

The collapse and its aftermath

Even though institutions tried many strategies to expand their teaching capacity, they
were eventually overrun by the relentless increase in student demand. Although the
report from Snowbird 1980 had warned that “limiting or cutting back enrollments would
be counterproductive given the societal need manifested in the rising enrollments,”
universities and colleges were forced to do just that. Most of those limitations were
based on academic performance and were extremely restrictive. At Berkeley in the mid
1980s, for example, only students with a 4.0 GPA were admitted to the major in
Electrical Engineering and Computer Science.

During these years, I was chairing the newly formed Department of Computer Science
at Wellesley College. Although we were more fortunate than many colleges in that we
were able to attract a few applicants in response to our searches, making actual
appointments remained a near impossibility. In 1982-83, Wellesley made six offers
before finding someone who would take the position. Most of our candidates accepted
competing offers elsewhere at higher salaries, both from industry and academia.
Unfortunately, hiring one person in that year was insufficient to keep up with the
increasing student interest in the computer science major. In 1983, Wellesley decided to
restrict access to the major, accepting only students who met a minimum GPA threshold.

Although limiting access to the major did reduce class sizes, it was not effective in
meeting the more general goal of improving working conditions for the faculty.
Enrollment limitations are, naturally enough, unpopular with students—and with their
parents. Imposing such restrictions makes the relationship between faculty and students
adversarial, causing students to become more competitive and, in many cases, angry.
Teaching became considerably less enjoyable, and I ended up leaving Wellesley for a
research lab.

7 Eric Roberts, “Conserving the seed corn: reflections on the academic hiring crisis,” SIGCSE Bulletin,
December 1999, page 4.
 – 10 –

The imposition of GPA thresholds and other strategies to reduce enrollment led
naturally to a change in how students perceived computer science. In the 1970s, students
were welcomed eagerly into this new and exciting field. Around 1984, everything
changed. Instead of welcoming students, departments began trying to push them away.
Students got that message and concluded that they weren’t wanted. Over the next few
years, the idea that computer science was competitive and unwelcoming became
widespread and started to have an impact even at institutions that had not imposed
limitations on the major.
 – 11 –

What happened during the downturn in the 2000s?

Please God, just one more bubble.
—Popular Silicon Valley bumper sticker, 2003

In many ways, the second installment in the boom-and-bust history of computer science
is easier to explain than the first. The growth in interest that began in the mid 1990s
coincided with the advent of the web and the elimination of commercial restrictions from
the Internet. These developments ushered in a period of frenetic growth in the computing
industry generally referred to as the dot-com boom or, when it is important to emphasize
its ephemeral nature, the dot-com bubble. The excitement generated by both the new
technologies and the opportunities provided by the startup culture attracted many students
back to the field. In the years of flat enrollments between 1991 and 1996, departments
had been able to rebuild their faculties, which meant that there was capacity—at least at
the beginning—to accommodate a rise in student numbers. When the tech bubble burst
in 2001, students began to move away from computer science, which led in turn to a
multiyear decline. Figure 3 shows the rise and fall during this cycle of history.

Even though the causes of both phases of this cycle are easy to identify, it is still worth
considering this period of history in more detail. From 1997 to 2003, the number of
computer science graduates rose by an average of 15 percent per year, with many
institutions seeing considerably larger increases. That rapid rate of growth raised echoes
of the expansion of the early 1980s. Several committees were formed to study the
problem in the hope that academic computer science could avoid the meltdown it had
suffered a decade and a half before.
 – 12 –

In 2000, academic computer science did avoid a meltdown, but not for any reason one
would like to repeat. In a perverse sense, academia got lucky. The industry collapsed
first. Had it not done so, it seems likely that capacity limitations would have forced
universities to restrict enrollment, with all the negative consequences that the field
endured in the late 1980s.

The aspect of the collapse in the early 2000s that seems hardest to explain is why
things took so long for student interest to recover. The industry bounced back very
quickly and was hiring at the pre-crash rate by 2004. The student numbers, however, did
not start to rise substantially until after the subprime mortgage crisis in 2007. Somehow,
a meme arose in the public consciousness in the form of a widespread belief that
computing jobs were in danger of imminent collapse, either because they would be
automated out of existence or because all software development jobs would be shipped
offshore. Although there was no evidence to justify those fears—and ample data to
refute them—that mythology kept students out of computer science until disaster struck
in a different sector of the economy.

The sections that follow look more closely at the effect of the dot-com expansion on
academic capacity and the failure of student interest to recover even after the dot-com
collapse had passed.

The effect of the dot-com boom on enrollments

The frenzy of excitement around the dot-com explosion in the mid 1990s generated
enormous student interest in computer science. As Figure 3 illustrates, the number of
bachelor’s degrees in computer science rose steadily from 1997 to 2003. Those numbers
therefore reflect the decisions that students made about their major field approximately
two years earlier, which aligns with the years of the dot-com boom.

During those years, the situation facing computer science departments corresponded
closely to the rapid expansion of the early 1980s and generated a similar set of pressures.
Along with the double-digit annual growth in student numbers, departments faced a
shortage of available faculty. In the December 1998 issue of the SIGCSE Bulletin, Paul
Myers and Henry Walker published a review of academic hiring, which concluded that
there was “a very serious shortage of new Ph.D.s in computer science,” to the point that
in 1997-98 “only about half of the open tenure-track positions were filled.”8 While that
level of undersupply falls far short of the seven-positions-for-every-applicant crisis of the
1980s, it nonetheless generated considerable concern, not only in university departments,
but also in the media, industry, and government.

In 1999, the Computing Research Association published a report entitled The Supply of
Information Technology Workers in the United States, detailing the shortage of workers
in both industry and academia. The report observed that academic institutions faced a

8 Paul Myers and Henry Walker, “The state of academic hiring in computer science: an interim review,”
SIGCSE Bulletin, December 1998, page 32.
 – 13 –

special problem, invoking the “seed-corn problem” popularized by Peter Denning a

decade earlier.9
Many educators, industrial laboratory leaders, and government science officials
are concerned that the high industrial demand for information technology (IT)
workers will siphon out of the educational systems many students who would
otherwise pursue an advanced degree. This diminishes [the] pool of people who
will join the university faculties that perform basic research and teach the next
generation of students. This problem is compounded when industry also
successfully recruits current faculty members, including junior faculty who
would become the academic leaders of the profession in the coming decades.
This is known as the “seed-corn” problem—an analogy to those who consume
too much of this year’s crop, reserving too little for next year’s planting.

As was true in the mid 1980s, the problems of faculty recruitment were noted by the
media. In September 1999, The Chronicle of Higher Education included a news story
entitled “Computer scientists flee academe for industry’s greener pastures” that begins
with the following evocative paragraphs:10
Just as he prepared to leave Cornell University last spring to help start a new
high-technology company, Thorsten von Eicken got word that the computer-
science department at Cornell had voted to grant him tenure.
He left anyway.
Mr. von Eicken is part of a stampede of bright, young Ph.D.s in computer
science who are abandoning academe for the corporate world.
High-paying, fast-paced jobs in the computer industry are attracting both
seasoned academics and newly minted Ph.D.s who, in the past, would have
opted for careers in higher education. The upshot: Computer-science and
computer-engineering departments are suffering a serious shortage of professors
at a time when undergraduate enrollments are booming.
Many departments are losing professors faster than they can hire them. The
University of Illinois at Urbana-Champaign recruited five new professors in
electrical and computer engineering to start this fall, but lost five others who
were already on its faculty. The University of Washington recruited four
scholars to its department of computer science and engineering but lost five.
Cornell hired three but lost six.

The difficulty of faculty recruitment was also picked up by The New York Times,
which ran an article entitled “Computer science departments are depleted as more
professors test entrepreneurial waters” on August 9, 2000.11 The article quotes Ed
Lazowska, then chair of the Computer Science and Engineering Department at the
University of Washington, as follows:
It is difficult to hold a computer science department together these days. You’d
like to keep a lot of that entrepreneurial energy here. Faculty recruiting and
retention are difficult. Ten years ago, industrial research labs were the enemy;
now it’s the lure of startups.

9 Peter Denning, “Eating our seed corn,” Communications of the ACM, June 1981, page 341.
10 Robin Wilson, “Computer scientists flee academe for industry’s greener pastures,” Chronicle of Higher
Education, September 24, 1999, page A16.
11 Rebecca S. Weiner, “Computer science departments are depleted as more professors test entrepreneurial
waters,” The New York Times, August 9, 2000.
 – 14 –

In 2001, the National Academies released a major study entitled Building a Workforce
for the Information Economy that looked broadly at the questions of the computing and
information-technology workforce, including several issues concerning education. At the
request of the study panel for the National Academies and with the endorsement of the
ACM Education Board, I submitted a white paper that focused on how the shortage of
faculty candidates was making it impossible for universities and colleges to meet the
demand from employers for graduates with the necessary level of expertise.12 That white
paper was cited several times in the final report, which issued the following conclusion
with respect to higher education:
The academic research enterprise in IT continues to be strong, but industry and
academia are competing for the same small pool of highly productive, creative
individuals. Ph.D. production and faculty recruitment and retention are both
threatened by the lure of the commercial sector. Some faculty and graduate
students are leaving academia for better-compensated positions in industry;
others leave because only industry (especially start-ups supported by venture
capital) offers them the opportunity to pursue their intellectual and research
interests. . . . Compared to the benefits to be found in industry and start-ups,
academic life—with the attendant burdens of low salaries, teaching, and the
need to obtain grant support—is increasingly seen as unattractive to many
graduate students. The long-term significance of these perceptions is at present
unclear, but they do not bode well for the long-term health of the IT field.

Although many of the discussions that led to the National Academies report took place
in 1999 and 2000, the final version was not released until 2001. By that time, the
situation in the computing industry had changed entirely. The speculative bubble that
had fueled the growth of a vast array of dot-com companies collapsed, and the industry
went into a tailspin. The NASDAQ composite index—which had risen from 740.47 at
the beginning of 1995 to a high of 5,132.52 on March 10, 2000—collapsed to 1,108.49
by October 10, 2002. With that collapse, investors lost trillions of dollars, the wide-open
job market of the late 1990s disappeared (if only for a couple of years), and students who
had been lining up to major in computer science, like many faculty members and
graduate students before them, started to look for greener pastures.

The slowness of the recovery after the dot-com collapse

On one level, the decline in computer science enrollments after the dot-com collapse is
easy to understand. At the time, the news was full of stories of the demise of Internet
startups, so recently the darlings of Wall Street, many of which lost their entire value
overnight. As startups collapsed and large companies started downsizing, the high-tech
industry did not seem like a good bet for stable employment over the long term. Students
headed off in other directions.

What is paradoxical about the downturn in student interest is that it persisted for many
years after the industry had fully recovered. While there was a small dip in overall
employment in the information-technology sector between 2000 and 2002, employment

12 Eric Roberts, “Computing education and the information technology workforce,” SIGCSE Bulletin, June
2000, page 83.
 – 15 –

numbers quickly surpassed their pre-crash levels. By 2005, employment prospects for
students completing a bachelor’s degree in computer science and other specialties in
information technology were considerably better than they were for students in any other
discipline. That conclusion is underscored in the following excerpt from a December
2005 publication from the National Science Foundation:13
Continuing a pattern that has been evident for decades, recent bachelor’s and
master’s engineering graduates and computer science graduates at the bachelor’s
level are more likely than graduates in other fields to be employed full time after
graduation, and upon entering the workforce, they are rewarded with higher
salaries.

These conclusions from the Bureau of Labor Statistics were echoed in the popular
press. In May 2006, Money magazine rated “software engineer” as the best job in
America on the basis of a combination of factors including salary, job availability,
potential for growth, flexibility, and creativity.14

One of the best analyses about the shortage of software professionals appears in a talk
by John Sargent, Senior Policy Analyst at the Department of Commerce, which he
presented at the CRA Computing Research Summit in February 2004.15 Although the
entire presentation is worth viewing—in part because it is striking how little things have
changed over the past decade—the talk is particularly memorable for the slide that
appears in Figure 4, which combines data on degree production from the Department of
Education with job projections from the Bureau of Labor Statistics.

13 National Science Foundation, “Recent engineering and computer science graduates continue to earn the
highest salaries,” InfoBrief, December 2005.
14 Money magazine, “Best jobs in America,” May 2006.
15 John Sargent, “The adequacy of the U. S. science and engineering workforce,” CRA Computing
Research Summit, February 23, 2004.
 – 16 –

Because of its unusual effectiveness, several people have updated this slide as new
releases of the relevant data become available. The most recent version I’m aware of was
created by my colleague Phil Levis and is available from
http://csl.stanford.edu/~pal/ed/. The take-home message of the graph, however,
has remained constant over the past decade: universities are producing far too few
graduates in computer science to meet industry demand.

The bar that towers over the other data points in Figure 4 makes it immediately clear
that the number of jobs in computing-related disciplines far exceeds the number of
students trained in those fields. If students were responding to market forces, the
imbalance between degree production and job growth would have sparked a stampede
toward computer science. That stampede did not happen. Despite the many economic
advantages available to those with computer science degrees, students shied away from
the field until after the subprime mortgage crash in 2007.

The reasons behind the lingering unpopularity of computer science during the 2000s
are complex. The factors certainly included memories of the pain associated with the
collapse of the dot-com bubble and a widespread fear that computing jobs would soon be
shipped offshore to low-wage countries. The fear of offshoring was particularly intense,
even though a 2006 ACM report entitled Globalization and Offshoring of Software found
no evidence that software jobs were disappearing in developed countries. In fact, the
report found that “despite a significant increase in offshoring over the past five years,
more IT jobs are available today in the U.S. than at the height of the dot-com boom” and,
moreover, that “IT jobs are predicted to be among the fastest-growing occupations over
the next decade.”16

An interesting illustration of the disconnect between the available economic data and
popular perception appears in the online response to a keynote address at the CIO
Leadership Conference by Maria Klawe, then Dean of Engineering at Princeton
University, with the title “Blue skies ahead for IT jobs.”17 The abstract for Klawe’s talk
reads as follows:
Contrary to popular belief, career opportunities in computer science are at an
all-time high. We’ve got to spread that message among students from a rainbow
of backgrounds, or risk becoming a technological backwater.

The comments that Klawe’s talk elicited—which have, unfortunately, vanished along
with the original website at CIO Magazine—ran at least ten-to-one against her
assessment of the sunny outlook for the field. Here are a few typical reactions that I
downloaded at the time:

16 William Aspray, Frank Mayadas, and Moshe Y. Vardi (editors), Globalization and Offshoring of
Software: A Report of the ACM Job Migration Task Force, Association for Computing Machinery,
February 2006.
17 Maria Klawe, “Blue skies ahead for IT jobs,” CIO Leadership Conference, Boston, MA, May 8–10,
2006.
 – 17 –

• All this talk about “Blue Skies” ahead just can’t hide the stark fact that Americans who
don’t wish to migrate to India and/or some other offshore haven are going to have a
difficult career.
• Why would any smart American undergrad go into IT when companies like IBM and
HP are talking of stepping up their offshoring efforts in the coming years? They want
cheap labor, no matter the real cost.
• I think the latest figures from the U.S. Department of Labor are not correct.

The last of these comments, which is quoted in its entirety, seems particularly telling.
The reader offers no alternative data, just a deeply seated belief that the optimistic
forecasts of the Labor Department must be wrong. Evidence counted for little in this
debate.

Ironically, popular fears about the tenuous future of the discipline were in some cases
encouraged by comments from within the academic community. In July 2008,
Communications of the ACM published a debate about future directions for the
technology curriculum.18 Professor Stephen Andriole at Villanova University predicted
that the need for programmers would soon diminish:
Of course there will be programming jobs for our students. But the number of
those jobs will decline, become more specialized, and distributed across the
globe. . . . Today, Fortune 1000 companies have far fewer programmers than
they did because of the rise of packaged applications and the labor-rate-driven
sourcing options they now have. This trend will accelerate resulting in fewer
programming jobs for our students. Should we continue to produce more
programmers?

In my response, I argued that Andriole was looking only at one sector of the
technology industry and that the number of jobs across the industry as a whole would
continue to rise, in line with the predictions from the Bureau of Labor Statistics. In
recent years, the sustained increase in the number of software jobs makes it clear that my
analysis was closer to the mark.

Although I never published an analysis of the reasons why students continued to stay
away from computer science long after the industry recovered, I did prepare a report for
the ACM Education Board, which considers this paradox in more detail.19

18 Stephen J. Andriole and Eric Roberts, “Point/counterpoint: Technology curriculum for the early 21st
century,” Communications of the ACM, July 2008, page 27.
19 Eric Roberts, “Understanding the paradox: strategies to rebuild student interest in computing,”
presentation to the ACM Education Board, August 22, 2008.
 – 18 –

What is the nature of the enrollment expansion today?

It’s like déjà vu all over again.
—attributed to Yogi Berra

Since 2007, the number of students taking computer science courses and declaring
computer science majors has been rising rapidly. To a large extent, the expansion mirrors
the earlier expansionary periods that occurred from 1979 to 1984 and from 1994 to 2001,
in the years just before the preceding crashes. In particular, the factors fueling the
increase in student interest are similar. The last seven years have been an extremely
exciting time in computer science, with the proliferation of computing into ever more
facets of everyday life. The ubiquity of smart phones and the applications that run on
them, the enormous successes in machine learning, the rise of big data, and so many other
advances all make computer science extremely attractive. Moreover, the seemingly
boundless opportunities for employment that computing careers offer—particularly when
coupled with the enormous uncertainty facing other aspects of the economy—are certain
to draw more students into computer science, just as those same factors have in the past.

The sections that follow trace the dimensions of the current expansion and its likely
consequences, exploring both the similarities and the differences from the earlier cycles.

The dimensions of the current expansion

Over the last four years, several members of the field who remember the history of the
two previous cycles, including myself, have sought to raise awareness of the danger of
rapid expansion in enrollment with no commensurate increase in teaching capacity. In
2011, I published an editorial entitled “Meeting the challenges of rising enrollments” in
ACM’s Inroads magazine.20 In that piece, I review the history of the earlier crises and
end with the following warning:
In the 1980s, the inability to hire new faculty made it impossible for most
departments to satisfy the increased student demand. As a result, institutions
were forced to discourage student interest by adopting such strategies as limiting
the size of the computer science major or staffing courses with inadequately
trained outsiders. . . .
As a nation, we cannot afford to repeat the failures of the early 1980s. As we
emerge from a decade in which far too few students chose to major in computer
science, it makes no sense to frustrate the renewed student enthusiasm by
turning yet another generation away because of a lack of resources. The
economy needs more people with computer science training, and we have a
collective responsibility to prepare students for those positions.

More recently, Ed Lazowska from the University of Washington, Jim Kurose at the
National Science Foundation (on leave from the University of Massachusetts), and I have

20 Eric Roberts, “Meeting the challenges of rising enrollments,” ACM Inroads magazine, September 2011,
page 4.
 – 19 –

given joint talks in various venues about the growing capacity problem, including the
NSF Future Directions in Computer Science Education summit in January 2014, the
National Center for Women in Information Technology summit in May 2014, and the
Computing Research Association’s biennial conference at Snowbird in July 2014. The
slides from our presentation, entitled “Tsunami or sea change?: Responding to the
explosion of student interest in computer science,” offer an overview of the capacity
crisis and are available from http://lazowska.cs.washington.edu/NCWIT.pdf. The
slides document the rapid growth in the number of computer science majors at several
leading research institutions, both public and private, as shown in Figure 5.

The expansion seen at these institutions is even more rapid than aggregate statistics
show because many universities insulate themselves from changes in demand by
controlling admissions to the major. For example, institutions like Carnegie Mellon and
the University of Washington experience lower variability because those institutions
admit students directly into the computer science program. The number of applicants
changes along with the national pattern of student interest, but the number of students
actually admitted remains much more stable.

The slides also document the continued strength of the job market for graduates with
strong computational backgrounds. Figure 6, for example, shows the projections from
the Bureau of Labor Statistics for job growth and job openings (job growth plus
replacement) for the STEM (science, technology, engineering, and mathematics) sectors
of the economy from 2012-22. By both measures, most of the employment growth over
the decade is in the computing disciplines, which account for 71 percent of job creation
and 57 percent of job openings in STEM.
 – 20 –

The “Tsunami or sea change” presentation ends with the warning that “we have seen
this movie before, and it wasn’t pretty.” In previous cycles, “universities did not respond
adequately,” increasing the importance of doing a better job this time around.

Today—as in past years in which student enrollments in computer science increased

rapidly—faculty openings exceed the number of applicants for those positions. From
July 2014 through June 2015, more than 700 distinct advertisements of computer science
faculty positions at North American universities and colleges were posted on the ACM
employment web site. Many of these ads listed multiple searches (sometimes as many as
five) at the same institution, so the total number of open positions is larger than the
number of advertisements. Although the precise number is impossible to determine
because many of the listings use imprecise phrases like “several positions” or “multiple
positions,” it appears that the number of open computer science faculty positions in
2014-15 was around 1000.

According to the Computing Research Association’s most recent Taulbee survey,

North American institutions produced 1,651 computer science Ph.D.s in 2014.21 Of this
number, 244 (15 percent) accepted faculty positions at North American institutions. By
this calculation, the current rate of Ph.D. production is sufficient to fill about one of every
four open positions.

Although the ratio of applicants to open positions is less than the one-in-seven shortfall
of the early 1980s, the number of unfilled positions is significantly larger in absolute
terms. If the number of Ph.D.s is sufficient to fill only a quarter of the open positions,
then the number of positions that cannot be filled from this pool is around 750. Unlike
other fields, computer science has no reserve labor force in the form of Ph.D.s who
received their degrees in prior years but who have been unable to find positions. Some

21 Stuart Zweben and Betsy Bizot, “2014 Taulbee survey,” Computing Research News, May 2015.
 – 21 –

positions, of course, will attract faculty members from lower-ranked institutions who
“trade up” to more prestigious employment. That flow of existing faculty up the ladder
of institutional prestige, which is usually referred to as churn in discussions of the
academic labor market, means that some of the 750 open positions will indeed be filled,
but only in a way that leaves vacancies in other institutions that will have to be filled in
future years. The only way to ensure stability is for the number of new faculty entering
the workforce to keep pace with the rate of departures and the growth of the field.

Where are we heading?

Figuring out how to respond to the current expansion is complicated by the fact that
history does not provide us with a clear sense of how the situation will evolve from here.
The two expansionary periods we have seen before—which were quite similar in form—
were followed by collapses that were qualitatively different. Are we heading for another
capacity collapse similar to the one the field experienced in 1984, or will we be saved by
a downturn in the high-tech industry that sends students scurrying away to other fields?
And, perhaps more importantly, is there any way to predict the actual outcome?

Given the uncertainties of economics, it is clear that the answer to the second question
is no: there is no way to predict with confidence exactly how economic forces will play
out in the high-tech industry and how those forces will affect enrollment patterns. Many
analysts believe that the situation in the technology sector is substantially different from
the “irrational exuberance” of the dot-com bubble. As The Economist reported on
July 25,22
Greed, profligacy, tiny companies with outlandish valuations: it is not hard to
detect echoes of the turn of the century, when the dotcom bubble burst
spectacularly and America’s economy stumbled as a result. But to see history as
about to repeat itself is to miss how deeply things have changed. Today’s
technology businesses are selling services and products from which they already
generate income, rather than just saying that one day they might. And the group
of people doing the investing is much smaller now than it was then. The risks
are on fewer shoulders.

This assessment, of course, is by no means authoritative. The important point is that

there is certainly no evidence to suggest that there will be a downturn in the high-tech
industry that doesn’t affect the economy as a whole. More importantly, it is foolhardy to
assume that there will be such a downturn and that academic computer science will be
saved thereby. We don’t know what is going to happen, and it is therefore important to
prepare for what might well be a capacity collapse similar to that of the 1980s.

There is, however, a reasonable interpretation of history that makes the differences in
the mid 1980s and the early 2000s less important. Rather than looking at the character of
the downturns, I believe it is more productive to focus on the pace of the enrollment
increases that preceded those periods of collapse. In the early 1980s, the late 1990s, and
again today, computer science departments face a rate of expansion that is much faster

22 “To fly, to fall, to fly again,” The Economist, July 25, 2015.
 – 22 –

than university departments ordinarily sustain. Those periods of expansion, moreover,

coincide with extremely tight labor markets for faculty, which makes it difficult to
respond to the increase in student load, even if the institutional will to do so is there. In
the absence of extraordinary measures that most universities have been unwilling to
undertake in the past, the rates of increase during the boom years are simply
unsustainable. And, in a marvelously succinct principle generally attributed to Herbert
Stein, economics tells us that
If something is unsustainable, it will stop.

Stein’s principle does not tell us how an unsustainable phenomenon will stop, only that it
will. The unsustainable buildup of the early 1980s ended with a capacity collapse. The
unsustainable expansion of the dot-com era ended with the collapse of the dot-com
bubble, and with it, the enrollment crisis that had threatened to overwhelm academic
departments.

The current rise looks very much like the previous ones and exists against a backdrop
of faculty shortages that bears all the hallmarks of past expansions. Unless new strategies
are implemented at a scale that has not been attempted in the past, this expansion too will
stop. Our foresight may not permit us to understand the precise mechanism of the
collapse, but those details will matter very little to those who suffer from its effects.

What strategies are currently in progress?

As the effects of increasing enrollments become more evident, professional societies have
undertaken several initiatives to address the capacity problems. These initiatives include
the following:

• The Computing Research Association has launched a “Booming Enrollments”

committee chaired by Tracy Camp at the Colorado School of Mines. As of late 2015,
that committee is beginning its data-gathering phase.
• The ACM Education Council has created a working group on the capacity crisis that I
am chairing. This document is intended primarily as input to that working group.
• The National Science Foundation and the National Academies are considering the
formation of a study panel to address broader issues with respect to the
information-technology workforce. I have spoken with staff members who are
working on the project, but do not know its current status and schedule.

Why has a concerted response from the community been so slow in coming?
Despite being several years into the latest period of skyrocketing enrollments, efforts to
address the problems are just now getting off the ground. For those of us who have lived
through past crises, this delay reflects an unfortunate change in the community’s
understanding of the problems. In the early 1980s, academic computer scientists had a
solid appreciation of the dangers they faced from massive increases in enrollment. The
same was true in the late 1990s—an understanding all the more vivid because memories
were still fresh from the capacity collapse of the mid 1980s. This time around, however,
 – 23 –

it has been much harder to get universities, departments, and individual faculty to
recognize the risks, despite the accumulation of additional historical experience.

In my view, the failure in this cycle to benefit from the lessons of history arises from a
confluence of several factors:

• Historical clarity is obscured by memories of the dot-com collapse. For computer

science departments today, and particularly for those staffed by faculty who are too
young to remember the earlier peaks and valleys of computer science enrollment, the
dominant memory is of the enrollment downturn that followed the collapse of the
dot-com bubble. For a little over a decade, computer science educators have been
working to increase student interest in the field. Once a movement like that gets
going, it is difficult to change gears, even when the number of students we serve
clearly threatens to overwhelm existing resources. Given the demand for people with
those skills, it is clear that we still need to attract more students to computing. To
work on building demand without also working to increase capacity strikes me as
suicidal.
• History has convinced many people that student demand is cyclical. When you look at
the variation in degree production over the last several decades, it is easy to conclude
that student interest varies in a cyclical fashion. As Jim Kurose said when he spoke at
Stanford’s computer science faculty lunch earlier in the year, he often meets
administrators who are convinced that they can predict the cyclical period and know
when students will next lose interest. That attitude, however, is inconsistent with
history. Students did not turn away from computer science in the 1980s by choice but
instead from a lack of teaching resources within the institutions. What’s worse is that
the assumption of cyclicity is self-fulfilling. If administrators decide not to support
computer science departments because of a belief that those enrollments will
inevitably fall, that very decision ensures a capacity collapse.
• There was a widespread conviction that technology would solve the problems. The
early years of the current enrollment increase coincided with the birth of MOOCs
(massive open online courses) and the excitement they generated in their early years.
To many people in academia, MOOCs held forth the promise of allowing universities
to reach larger communities of students at lower cost, thereby enabling an increase in
productivity. Although that vision failed to materialize, the widespread enthusiasm for
using technology to solve the capacity problem in computer science made it difficult to
consider other options.
• Teaching large computer science courses has become a more specialized endeavor.
Computer science enrollments and major counts are much larger today than they were
at the height of previous enrollment expansions. As a result, many departments have,
entirely appropriately, hired people specifically tasked with teaching the
lower-division courses that have seen the greatest increases in enrollment. The
downside of this approach is that much of the faculty is insulated from the pressures of
increased student interest until the enrollment bubble reaches the upper-division
courses.
 – 24 –

• Much of the early history lies beyond the Google “event horizon.” In putting together
this history, I was interested to discover that several relevant articles I remembered
from the early 1980s were invisible online because they predate digital archiving for
the journals in which they appear. Looking for evidence about faculty shortages in the
1980s becomes much harder when none of the references from, for example, The
Chronicle of Higher Education, show up in Google searches.

The fact that people who are responsible for making decisions that affect the future of
computer science education are less aware of the problems of the past makes it harder for
the field to act with a common purpose. That historical myopia, however, in no way
reduces the importance of finding a way to forestall a repeat of the capacity collapse of
the 1980s that cut the number of qualified computer scientists nearly in half. Our society
cannot afford to repeat that mistake.