Date: Fri, 27 Aug 2004 23:23:28 -0700 From: Norm Matloff To: Norm Matloff Subject: the future of academic computer science in the U.S. To: H-1B/L-1/offshoring e-newsletter I hope readers consider my Subject: line above to be weighty. I certainly believe it is a weighty topic, one very important to the future well-being of the nation, touching on such questions as: Is a national role of _managing_ tech development (by foreign workers, both abroad and on work visas in the U.S.), instead of doing the development itself, a sustainable U.S. economic policy? Does that management role require that the manager have a strong technical background? If so, will university students still be interested in majoring in computer science (CS), if the major leads only to nontechnical work as managers? And how will they acquire the necessary technical experience on which to base that management, if development work is done by foreign workers? During the late 1990s, university students were wildly implored to study CS by all our major American institutions--business, the press, Congress, the educational establishment. Young people heeded the call, and CS enrollments soared. It was a national mobilization similar to that of the aftermath of the Russian Sputnik in the 1960s. That earlier mobilization arguably was highly successful and worthwhile, ultimately giving rise to our highly successful high-tech industry. By contrast, the mobilization of the 1990s produced graduates who are now not finding jobs. Did our institutions fail us this time? Will anyone heed the next mobilization they call for? Since I am so involved personally, I feel I must begin with an anecdote/disclaimer. (If you wish to skip this, have your e-mail utility search below for the phrase "no hidden agendas".) In 1998, during the time the industry lobbyists were stridently claiming a severe tech labor shortage and were pressuring Congress to double the yearly cap on new H-1B visas, employment writer Marty Nemko ("Cool Careers for Dummies") invited me to appear as a guest on his radio show to discuss the issue. (For a voice from the other side of the issue he invited TJ Rodgers of Cypress Semiconductor, coincidentally the second time I had debated TJ that week.) During the show, Nemko was polite and friendly (something I certainly can't say about TJ), but he definitely came down on the side of the industry on the issue. He also began his show with a paean to some immigrant fruit vendor he had patronized (a word which I suspect was true in two senses), the message basically being that a liberal immigration policy is a good thing. After the show, though, Nemko and I exchanged e-mail for a couple of days, during which time he was openly hostile. I was mystified at his attitude, until the truth came out: In my bio, I had mentioned that I had served as chair of the Academic Senate Committee on Affirmative Action at UC Davis, and it turns out that Nemko considers himself to be a terrible victim of affirmative action, having been passed over for a faculty position himself at UCD. Nemko considers affirmative action to promote hiring of the weak, and he decided that when I said employers should hire American programmers/engineers instead of H-1Bs, that again meant I favor hiring of the weak, in his mind. Since that time, Nemko has from time to time again given the pro-industry viewpoint on this issue on his show. Two listeners whom I know but who are unknown to each other and did not know about my previous interaction with Nemko, wrote to him to object to his stance. Both cited me, and in both cases Nemko responded by castigating me. To one of the listeners, Nemko cited industry reports which supposedly contradict what I say. When the listener pointed out that the industry has an obvious vested interest in producing these reports, Nemko claimed that I have vested interest too: When I criticize offshoring, Nemko charges, I am merely trying to protect my job as a computer science professor, which Nemko thinks would be at risk if most programming jobs were to go offshore. So, before going into the issue of the future of CS at U.S. universities below, I must make the following disclaimers: 1. My job is not at risk. (Not that it is much of an issue for me anyway, at age 55 with 29 years as a UCD professor.) On the contrary, my colleagues and I are glad to see class sizes finally come down to reasonable levels (a couple of the recent articles on declining enrollments of CS majors have quoted other CS faculty on this point too). All along, about half of our department's enrollment has consisted of nonmajors who want to learn more about computers. That will hold steady and in fact should grow. Consequently, as one of my colleagues has pointed out, U.S. CS departments are headed for a state similar to that of Physics departments, described in academia by the term "service departments," with enrollments consisting of a small number of majors and serving a large number of students from other departments. Needless to say, most faculty would prefer to teach mainly majors, and if the pool of majors shrinks that will adversely impact CS departments' ability to run their graduate programs, a big disappointment to most faculty. But their jobs are not at risk. 2. I don't support hiring of weak workers. On the contrary, I've repeatedly criticized employers for hiring weak workers, by having their first applicant screening be on the basis of unwarranted skills lists, rather than on overall programming talent. I've repeatedly cited research showing the paramount importance of talent over skills. And concerning H-1Bs, I've repeatedly supported rolling out the immigration red carpet for the world's "best and brightest." So, in my analysis below of the future of academic computer science departments, I have no hidden agendas. Now, let's get to the analysis. I am enclosing two articles for you below. The first, from InformationWeek, shows you the message that university officials would like the outside world to see: "Don't worry about all that stuff you see in the press about offshoring and unemployment, university students. There are lots of exciting applications of computers coming up, and you will be needed to develop them. Don't abandon the field!" Several other articles with this theme are at http://heather.cs.ucdavis.edu/Archive, with titles starting with "CSEnrollmentDrop". But the second article enclosed below is different. This is from a magazine devoted to engineering education. Here you are seeing what engineering faculty are saying more "privately," in speaking to "their own kind." Here the message is more mixed: "University students, that stuff about offshoring of software development jobs is true. So, go ahead and have fun programming during your time in school, just like you had assumed when you signed up for the CS major, but just keep in mind that your future career won't be like this at all. Your work will be pretty much the same as if you had majored in Business Administration instead of CS. You'll never touch a computer, except to check your e-mail." Of course, my above summaries of the two articles are overdrawn. In particular, the second article doesn't quite get to such a strong stance as what I described above. There is considerable commonality between the two articles. The "mainstream" article does mention a move on the universities' part to require more business courses for CS majors, for instance. And both articles claim that America's answer to the offshoring problem is innovation, a favorite "solution" claimed by the pro-offshoring crowd. But the second, "private" article more sharply raises the key question: Should engineering education become less technical and place more emphasis on business and soft skills? In other words, instead of producing "real" engineers, should we be producing business administration experts who have some background on the technical side? The following passage is very significant: "The people skills needed to work overseas are more important than ever before," [Olin College of Engineering president Richard] Miller continues. "Our graduates have to learn to be willing to accommodate and not offend. That's hard enough to do when you're sitting across the desk from someone. It's even harder to do when your contact is primarily by phone or e-mail." The message: Engineering graduates should prepare for careers as managers, not as engineers. Note carefully that this is less true for, say, civil engineers, but highly true for the computer fields. The other quote of Miller also shows that in practice engineers tend to start moving out of the technical area within five years of graduation. Many, in fact, never work in the technical area at all. The comments of Miller and others below shows that this phenomenon will become a lot sharper in the coming years. (I remember the irony, again in 1998, in hearing a talk from a Texas Instruments rep to a bunch of engineering professors, telling them to increase engineering enrollment at universities because of a "desperate" shortage of engineers--when this TI rep herself was already working in a nontechnical capacity only a couple of years after earning her electrical engineering degree at a prestigious university.) As I've said before, when this is explained to prospective CS majors up front, most are not interested. As one student put it so well to me, "If I'm going to end up with an Econ-type job, I might as well major in Econ" (and not have to stay up to 3 a.m. debugging programs for his CS classes). As I mentioned above, both articles cite innovation as the "solution" to the offshoring problem. It has become a mantra, an incantation, an article of faith--choose your own metaphor here, but the point is that it is repeated over and over again with little or any thought. The reality is that innovation isn't a solution at all, for the following reasons: 1. Even in the most optimistic scenario, the number of programmers working on innovative projects would be only a small fraction of the number of programmers we've had "traditionally" (i.e. over the last 10-20 years). That's hardly a solution, is it? 2. The venture capitalists are pushing the firms they fund to do most of their R&D offshore. (See the files with names starting with "VCsDemandOffshoring" in http://heather.cs.ucdavis.edu/Archive) And don't forget, what isn't offshored will typically be done by H-1Bs, not by U.S. citizens and permanent residents. 3. Most innovation is seredipitous. It comes when people constantly work in the field. The premise of the innovation argument is that Americans are more innovative. Fine, but the point is that if you decimate the number of Americans working in the field, you decimate the amount of innovation. In other words, offshoring, based on the notion that India has a comparative advantage (namely, cheap labor) over the U.S. in programming, would ironically cripple America's own putative advantage, the propensity to innovate. This is even worse in light of the fact (shown in past research) that when job opportunites in a profession are radically reduced, the students most likely to bail out of majoring in that field are the most talented students. How about that, Marty Nemko? Norm http://www.informationweek.com/shared/printableArticle.jhtml?articleID=29100069 InformationWeek By The Book Declining computer-science enrollments should worry anyone interested in the future of the U.S. IT industry. By Eric Chabrow, InformationWeek Aug. 16, 2004 Approaching San Angelo, Texas, on Route 87, travelers behold apanorama of a stony, flat terrain blanketed with buffalo and Indian grasses, cacti, yucca, prickly pears, and mesquite trees. The arid landscape segues into fields of maize and cotton. Nearby, sheep and cattle graze on large ranches. San Angelo, an oasis of a city of nearly 90,000, is situated near the junction of the southern stretch of the American Great Plains and the northern tier of the Great Chihuahuan Desert. It hosts a school that for three decades has been a quintessential training ground for the American IT workforce and that, like computer-science programs around the country, finds itself at a crossroads. Angelo State University has been turning out computer-science graduates since 1974, supplying American businesses with the professionals who develop, implement, and operate their IT systems. Angelo State and other schools offering undergraduate computer-science programs are facing declining enrollments as the profession loses some luster, and fields as diverse as biotechnology and criminal justice are seen as more exciting choices for talented science-minded students. Perception has become reality. Enrollment at undergraduate computer-science programs peaked at the turn of the millennium, then plummeted in the past few years by 30%. "In the '90s, people saw computer science as a quick opportunity for lucrative, high-paying jobs," says Stuart Zweben, chair of Ohio State University's computer- and information-science department. "Then companies began layoffs, and people heard about offshore outsourcing. That got them scared to go into this field. Students are becoming cautious. They have cold feet." Undergraduate enrollment at Angelo State's computer-science department is down sharply. In the late 1990s, the school, which has 6,000 total students, graduated about 40 computer-science students a year; last spring, 18 received bachelor's degrees from the program. Angelo State isn't alone. Boston University's Metropolitan College counts some 400 computer-science majors, down from more than 500 just a few years ago. Intensely competitive, elite universities such as Carnegie Mellon and Stanford have no problem filling their classrooms, but they're getting fewer applications. Carnegie Mellon's School of Computer Science's incoming freshman class has 130 students, but applications have fallen about 40% from a peak of 3,200 in 2001. Mark Stehlik, assistant dean of undergraduate education, characterizes the 2001 application figure as artificially high. "We had too many kids with parents who dreamed of six-figure initial job offers," he says. Stehlik says it would take another 40% drop-off in applicants to adversely affect the quality of the students admitted to the program. Yet declining computer-science enrollments should worry anyone interested in the future of the U.S. IT industry. While the Carnegie Mellons and Stanfords of the world won't have trouble filling their chairs, the future of IT innovation depends on them getting their fair share of the very best young science minds to come up with the truly breakthrough ideas in still-emerging fields such as robotics, artificial intelligence, and next-generation information security. On the more practical level, if companies can't get enough people in the United States trained in the IT skills they need, it provides one more reason to ship work to places such as India, which will mint more than 100,000 graduates in IT-related disciplines in the coming year, according to Nascomm, an Indian IT business association. Here's one look at where the numbers have gone: In 1995, some 10,000 undergraduate students at Ph.D.-granting schools--which represent about a third of the nation's computer-science programs--declared majors in computer science and computer engineering, according to research conducted by the Computer Research Association, a group supported by more than 200 departments of computer science, computer engineering, and related fields. That number doubled two years later. By 2000, the number of students declaring computer-science majors at these universities approached 24,000 and hovered at that level for another two years. Enrollment in computer-science programs soared in the mid- to late 1990s, as year 2000 remediation and the dot-com and telecom booms created an IT labor shortage. "Parents steered their children to computer science for dreams of instant wealth," Stehlik says. "It wasn't that the field was cool, but the dollar signs were cool." However, the bust doubled the nation's IT jobless rate to about 5.5%, and IT started looking like most any other field--you had to scrape and hustle for a job out of college, especially a high-paying one. The number of undergraduate students declaring as computer-science majors at Ph.D.-granting schools plunged by some 30% to about 17,700 last year. Non-Ph.D. programs reported similar declines. Computer science often loses out to other fields of study, many of which depend on high-end computing. The type of student who once expressed interest in computer science now is lured by life sciences such as biology and chemistry, or even criminal justice, attracted to those fields by the popularity of criminal forensic shows such as CSI and Crossing Jordan. "Things on TV guide their interests," says Charles McCamant, head of Angelo State's computer-science department. Leaders of computer-science programs, having ridden a rising tide of employment and prominence for decades, concede they need to do a better job promoting their discipline and highlighting the great challenges ahead. Stehlik notes that in real life, criminologists rely heavily on computers to solve crimes, something represented on TV shows by images of fingerprints quickly flashing by on a PC monitor. "What's really happening here is pattern matching. That's computer science," Stehlik says. "On these shows, we see the test-tube side; there's a computer-science side, too, that's not played up. ... As a field, computer science has done a lot less PR than it needs to do." While waiting for a reversal in computer-science enrollment, schools aren't necessarily panicking. Though criminal justice and other disciplines gain at the expense of computer science, universities and colleges haven't made knee-jerk changes. At Angelo State, for instance, university president James Hindman maintains the status quo. "We don't respond to fads," he says. "If a program becomes entrenched after four or five years and passes from the fad into the mainstream, we'll adapt. The marketplace is a greater arbiter." Most troubling for the future of computer science is the idea that students wouldn't pursue the field because they believe there are no jobs, that all the work is going to India, or that all the cool stuff has been done. "It's natural for people to look at a narrow point of time and conclude that businesses aren't well capitalized and jobs aren't plentiful," Ohio State's Zweben says. "They're wrong on both counts. You've got to believe IT is an important factor in our future." Indeed, over the next decade, the Bureau of Labor Statistics sees the need for an additional 307,000 computer-software engineers, 184,000 systems analysts, 106,000 network-systems and data-communications analysts, and 103,000 managers. For the next 20 years, "every advance we can anticipate is going to require software that has not yet been written," predicts Grady Booch, an IBM fellow and author of several books on software programming. The drop in computer-science enrollment concerns business executives who'll need that talent so their companies can properly function. Bernie Francis, owner of IT services firm Business Control Systems LP, worries that lower computer-science enrollment will results in a dearth of qualified IT workers to hire in the not-too-distant future. "The development cycle time is getting shorter and shorter, and we need more and more innovators," says Francis, a member of the Texas State University System board of regents. Without an increase in the IT workforce, Francis sees employers' costs rising. "We'll go back to the situation of tremendous salary increases for those who care to choose IT for a career," he says. Beyond whether U.S. computer-science programs will turn out enough graduates lies the question of whether they'll turn them out with the right skills. Too often, they're not, IBM's Booch says. They're good at what he calls out-of-the-box thinking but weaker on fundamental business skills such as teamwork and project design. "Many universities teach people how to program, but they don't teach them how to work in projects," Booch says. "They don't teach them how to design." Computer-science programs can't function in a vacuum. As technologies quickly mature, academia must work closer with industry to identify the skills needed to be taught, says Tanya Zlateva, chair of the computer-science department at Boston University's Metropolitan College. Ohio State's Zweben, a former president of computer society Association for Computing Machinery and the Computing Sciences Accreditation Board, says he believes U.S. colleges and universities generally do a good job of providing technical skills and an improving job of teaching students how to communicate and collaborate. But teaching students how computers help businesses be more effective is something most computer-science programs must improve. "I don't think most schools do that great of a job," Zweben says. "We don't do it here." Zweben laments that the demands of the computer-science program, especially one tied to engineering, leave precious little room for courses outside the technical area. Still, at Ohio State, computer students can minor in business if they find the time. There's growing pressure on schools to provide computer-science majors with an understanding of how information systems have an impact on an organization. It's not just business but how computers help researchers find new drugs, designers make sleeker cars, or police solve a crime. "The one thing that's more important now than before is having an understanding of the application's domain," says Gerald Engel, a University of Connecticut computer-science professor and president-elect of the IEEE Computer Society, an association of computer academics and professionals. As at Ohio State, the University of Connecticut is finding that an interdisciplinary approach to computer-science education isn't easy to pull off. At UConn, for instance, there's pressure on the faculty to reduce the number of hours for a student to earn a bachelor's degree to 120 hours from 128, the equivalent of two or three courses. The larger course commitment "makes it difficult for a student to pick up some type of minor in some application area," Engel says. "How do you decide amongst this whole array of choices out there? How do you cram more and more stuff into four years?" Some schools are trying. Carnegie Mellon requires computer-science majors to take a minor. But for many students, that means extending their stay in college for a year. Nationwide, fewer than 40% of undergraduates complete their degrees in four years. At Angelo State, the search for business experience (and spending money) leads about a third of the department's undergraduates to work part time in the university's IT department (see story, "No Games, Campus Work Builds Business And Life Skills"). Most schools that mix computer science with other disciplines, such as bioinformatics, do so at the graduate level, though upper-level undergraduates can sample interdisciplinary subjects. Boston University's Metropolitan College, for instance, lets senior computing-science majors take courses in cell biology and medical imaging. A decade ago, computer-science programs would have offered courses in specific languages. That's less common now. Today, courses tend to be more process-oriented, such as software development or enterprise computing, with a language such as C# or Java merely used as a tool to teach the subject matter. Topics such as information assurance and security are becoming a core part of the computer-science curriculum. At Boston University, the subject isn't just a single course or two but is embedded into the coursework of all basic computer-science classes. "That's what everyone must understand to function in today's world," Zlateva says. "Terrorism or no terrorism, information security must be in place." Teaching concepts broader than specific skills comes naturally to college educators. Because of their research, academics can identify trends five to 10 years before they're widely adopted by business, though they tend to have a harder time predicting the tools needed to implement those trends. Academics saw the coming of enterprise computing years before SAP and the Web became part of the lingua franca of the business world. Still, they didn't know until recently what specific tools the marketplace would adopt to help make that happen. "We can better tell you what things will happen in 10 years than what specific technology will win this year or next," Zlateva says. "We knew something like the Java enterprise platform or .Net platform would be important. The ideas behind them have been known for years. We just didn't know it would be Sun or Microsoft behind them." That's why computer-science programs shy away from specific languages and platforms and emphasize broader topics. "We don't train, we educate," Engel says. "The biggest thing you buy a university graduate with a computer-science degree is the ability to adapt." Angelo State president Hindman recently got into a lively discussion on that topic with Business Control Systems' Francis, the IT services executive who's also a Texas State University System regent. With only about half of the students who declare computer-science majors completing the program, Francis proposed that the school provide certificate training on specific technologies. "We need to do something with the spoilage rate," he says, referring to the students who dropped out of the program. "We need to get people prepared for life." But such training, Hindman responds, belongs at community colleges and other technical training schools, not university computer-science programs. He sees a university computer-science education not as vocational training but as a foundation for a professional career, one that could even lead to the CEO suite. "Knowledge of computer science is a stairway to reach that point," Hindman says. "That's the difference between education and training." Booch of IBM agrees with the distinction but says there's another way to fill the need that universities aren't meeting. He's on the board of a new for-profit Utah school, Northface University, that's essentially designed to provide ready-to-work software developers. He describes it as a true university where students develop into "whole people," but greater emphasis is on technical skills as well as work skills such as the social dynamics of teams, project-based work, and real-world problems such as designing for the complexities of Java 2 Enterprise Edition and .Net. "There's a fine line here," says Booch, who also sits on the board of a school of theology. "The last thing you want is to corrupt the educational process and become a trade school." Julie Horsman, human-resources manager for the IT organization at truck maker Paccar Inc., says university computer-science programs need to educate people who are adaptable and able to handle a wide range of business-technology challenges. "We want employees with broader set of skills, as opposed to a large number of people doing .Net or Java development," she says. "How many people do we need to crank out C++ code if we use rewritten components?" Not all companies are like Paccar. It's the marketplace that makes schools such as UConn teach students the language de jour. As part of a course that teaches the concepts behind programming, Connecticut requires students to learn C++. This fall, it will offer Java to half its freshman computer-science class. Engel isn't comfortable with that approach, but he understands why it's being done. "The trouble with C++--or Java, for that matter--is that they're horribly complex languages. Instead of concentrating on how computers solve problems, we can get mired in the details of C++ or Java, and that's regrettable," Engel says. "But it's a reflection that people want to have skills so they can walk out and get a job. This is the kind of thing personnel departments look for. In the '70s, you put Cobol on your resumé, and you got a job." Soft skills that help facilitate teamwork and collaboration are being integrated into many computer-science programs. Boston University requires students to work together on projects; one class has four or more students developing an application for which each person is assigned a specific responsibility that mirrors the workaday world. Sometimes the students collaborate in person, other times over the Web, simulating the experience IT managers face when working over the Internet with business partners in India. "Computer science requires teamwork," Zlateva says. "There's no such thing as one person developing one kind of application. On top of that, it does require very good understanding of the business environment to make a product successful. ... What's not going away, and where we need to educate our students, is on the high level of coordination and project management." Angelo State requires computer-science students to take courses in public speaking and either scientific or business writing because communication is an important skill for most IT professionals. "We have to document everything," says recent Angelo State computer-science grad Ryan Cleere, who took a technical writing class, of his work as a systems administrator for Northrop Grumman in San Angelo. Carnegie Mellon computer-science students learn collaborative skills by taking a project-based course in which students from computer science and other disciplines such as art work together to create a software-based virtual world viewed through a head-mounted display. "We're teaching our students the ability to work in teams, helping them develop communications and interpersonal skills, so they can communicate with those who are not necessarily as tech-savvy," Stehlik says. This interdisciplinary approach might be the salvation for computer science and could eventually attract a different breed of student than from an earlier generation. "The students who come in want to do more than just hack," Stehlik says. "Some students have political designs; they're interested in greater issues that confront society: security, privacy. We're seeing students who are extending the notion of computer science." Copyright © 2003 CMP Media LLC http://www.prism-magazine.org/jan04/global.cfm/ ASEE Prism Magazine PUTTING IT INTO PERSPECTIVE By Dan McGraw Two years ago, the state of New Mexico was looking to overhaul its system of supplying unemployment insurance to 72,000 claimants each year. The system was paper-based; information was entered into a 30-year-old mainframe, and field officers had to travel long distances to collect and verify information. New Mexico's Department of Labor did not have the IT expertise to create a new automated system for processing claims. The work had to be contracted out. The state took bids from companies that would overhaul the workers compensation system. TRW said they could do the work for $18 million; IBM came in at $12 million. Both companies said the work would take about six years. A third company, Tata Consultancy Services (TCS), also bid. The Indian company said it could do the job for about $6 million and could complete the work in 15 months. TCS used 110 employees to do the job: 30 in New Mexico and the rest in India. With a 24/7 operation and using a Web-based system of collecting data, TCS was able to save the taxpayers of New Mexico a significant amount of money and made the system of filing unemployment claims far more efficient. The irony of the New Mexico case is that some of the people using the new system are, undoubtedly, out-of-work IT workers whose jobs have been lost to this type of outsourcing. The underlying question in the outsourcing debate was made very clear in the choice New Mexico made in picking a company to handle its unemployment insurance: Namely, are the savings and efficiencies that TCS brings to the table worth the loss of some American jobs? More importantly, is this an inevitable trend? Raman Unnikrishnan, dean of engineering and computer science at California State University-Fullerton, says the trend is not only inevitable, it's not necessarily a bad thing. For Unnikrishnan, the nature of work in the new knowledge economy has changed in a fundamental way. "Instead of qualified people seeking work wherever work is available," he says, "work is seeking qualified people wherever they are." The reason, Unnikrishnan notes, is that the products of information technology--whether that be computer code or reading X-rays or developing engineering plans for architecture--can now be transported around the globe with few costs. There are no transportation charges, no shipping delays, no tariffs. Traditional boundaries in the workplace have vanished. If a computer engineer can do useful work at $25,000 a year, and the same work costs a company $60,000 in California, the market will place that work in India, Unnikrishnan says. "Outsourcing is a natural outcome of the information technology field," Unnikrishnan says. "It is not going to end with IT or customer service. But that doesn't mean that globalization is a bad thing. It is going to force the American science and engineering community to do what we have always done best. And that is innovation." The issue of outsourcing white-collar jobs to emerging markets like India, China, and Russia, among others, has become a hot-button issue of late. Gartner Inc., a high-tech forecasting firm, estimates that one in every 10 software jobs will be moved overseas by the end of 2004. Forrester Research, a marketing research firm, predicts that 3.3 million high-tech and service-industry jobs will move overseas by 2015, jobs that will provide $136 billion in wages. A Change in Course? While politicians and economists wrangle over solutions to the perceived problem, engineering educators wrangle over how to respond. The debate over this complex issue boils down to some simple choices. Namely, should U.S. engineering education continue to provide the basic science foundation for its students that encourages innovation--whatever that might be--or should engineering schools tweak their curriculum to provide their students the necessary nonscience skills to compete in the global economy? "We have a tendency to overreact to the immediate crisis," says Nino Masnari, dean of the College of Engineering at North Carolina State University, and a professor of electrical and computer engineering. "We have to continue to give our students the best scientific education. But we must always re-evaluate what an engineering education is all about. It boils down to the question of whether we are adding enough value to our students so the American companies will see that value and hire our students. Adding that value is key to re-evaluating our programs." The situation facing engineering educators is how to view the outsourcing problem. On the one hand, a variety of factors--the economic slowdown, vastly improved communication, routine IT work becoming commoditized -have led to the outsourcing boom. On the other hand, many engineering educators feel universities need to better emphasize the skills used in the new global economy--teamwork, systems over specific knowledge, and marketing. "When I talk to CEOs from industry, they say they will observe young engineers that reach a career plateau relatively early, usually within about five years," says Richard Miller, president of the Olin College of Engineering in Needham, Massachusetts. "It is not because they are deficient in some technical way. Instead, it is because they have problems in relationships with people. They may be working on a team where they have to deal with marketing and manufacturing people. They will be dealing more and more with the business office or a client. "The people skills needed to work overseas are more important than ever before," Miller continues. "Our graduates have to learn to be willing to accommodate and not offend. That's hard enough to do when you're sitting across the desk from someone. It's even harder to do when your contact is primarily by phone or e-mail." Miller suggests that American engineering schools must play to the strengths of our system in the new global economy. American engineers, he says, lead the world in two fundamental ways: innovation and the ability to recognize and improve systems. While India may be good at writing specific computer code, Germany excels at precision, and Japan at continuous improvement, American engineers excel at creativity, Miller says. "About the time we begin to lose jobs overseas, we change the game, and it makes the argument irrelevant," Miller says. "The business of being creative is fundamental to our long-term economic health. This creativity needs to be nurtured, needs to be emphasized, needs to be measured. "The cultural and ethnic diversity foster this creativity," Miller continues. "This diversity is not replicated anywhere else around the globe. A diverse group of people has a better chance of recognizing opportunities. We need to encourage diversity, from within our own country, to having students from other countries study here. That flow of incredible talent really enhances the rate at which we innovate." It is clear the globalization trend is affecting the job market for engineers in several ways. The first is that graduates with specific, individual skills will more than likely find that their jobs can be done as well and cheaper in emerging labor markets. For example, a student whose expertise is to provide improved ways to apply paint coats to an automobile may find that an engineer in Russia can provide the same service at one-third the cost. The second trend is that the traditional job promotion track, where an engineering graduate may spend his entire career with one company, is a thing of the past. John Anderson, dean of engineering at Carnegie-Mellon University in Pittsburgh, says these trends make it necessary to change some of the ways that engineering is taught. In September 2003, Carnegie-Mellon convened a panel discussion of industry leaders and educators to discuss what specific recommendations might help students in the global economy. The conclusions were that students must be more multidisciplinary in their skills, and that working in partnerships within teams was a skill that most American companies prized. "We're still ahead of the world in innovation," Anderson says. "And U.S. students still integrate science and engineering into systems better than anyone else. But we can do a better job in bringing a global awareness through business and humanities courses. The constraint is that we still have to provide a good, solid technical education." Anderson suggests the general education requirements be changed to reflect the global marketplace. For example, some schools might require a course in economics but do not allow a business or marketing course to fulfill the requirement. Anderson also thinks studying abroad for a semester or two should be encouraged. In addition, partnerships with foreign universities, where students collaborate in teams via the Internet (Carnegie-Mellon has such a program with Technical University at Delft in The Netherlands), should be implemented. "Students need to have an appreciation of markets, what customers of technology really need," Anderson says. "Increasingly, for U.S. companies, those customers and markets are in foreign countries. When we are talking about creating new technologies, we cannot only think in terms of the U.S. market." North Carolina State's Masnari agrees that changes should be made in general education requirements to better meet the needs of students. "In the past, accreditation of an engineering program has been a bean-counting exercise," he says. "With the new global marketplace, we need to have the flexibility to use the general education requirement to better serve the students. We should be able to better define what our programs offer. The onus should be on the institution to do this." Masnari also believes that teamwork is essential in the global economy, and that engineering schools can do a better job of teaching that skill. "We used to be very compartmentalized, everyone had their specific niche," he says. "You basically worked within your own discipline. Today, the trend is toward larger and more complex projects. It is critical that students learn the skills of working within teams." The U.S. Department of Commerce, in a report released in June 2003, suggests that many companies want computer science and IT workers to have a better understanding of the business side of the business. Employers, according to the report, are stepping up their recruitment of people with M.B.A.s or master's degrees who also have technical skills. Overall, 12.3 percent of IT workers hold a business degree. The ability to understand the business side, according to the report, provides "a deeper `foundational' knowledge'" that "is likely to prepare them for technological change and learning new technical skills when needed, rather than just knowing the `skill of the day.'" The challenge for engineering educators is to provide some of the "softer" skills required for the new global economy, without sacrificing the necessary "hard" science that drives innovation. At Olin College of Engineering, the curriculum is being "bundled" in ways that combine different courses within a team project. Students participate in a team-based project every semester for eight semesters. In one "bundle," students study biology and business while completing an engineering project. Another bundle unites history, materials science, and engineering. "This teaches students how to work and use the resources of the team, across disciplinary boundaries," says Olin's Miller. "We have quite an interest in entrepreneurial thinking and business," Miller continues. "It is a practicum that is overlayed in much of what we do. Starting with freshman, we emphasize the ability to recognize opportunities. We want them to think about business opportunities and the relationship with technology. We're trying to weave that into the engineering curriculum. "The strategy might be better not what to teach, but how they learn," Miller says. "You cannot teach every chapter in every book. You cannot cram sufficient knowledge covering everything into four years. It's certainly a matter of balance. We need to look at how students learn, instead of perhaps what they learn. How to answer questions, how to integrate within systems, how to work in teams--those skills are important but aren't taught in books." Labor's "Manifest Destiny" None of the engineering educators interviewed for this article believe that radical changes are needed for the new global landscape. Cal State-Fullerton's Unni-krishnan says that a historical view needs to be taken into account. "If we look at the 1980s and the technological scenery from that period, we were extremely despondent about the economy," he says. "Japan was supposedly taking over the world, U.S. productivity was low, and trade was out of tilt with the rest of the world. "But everything changed in a relatively short period," Unnikrishnan continues. "Innovation led us to the boom years of the 1990s and into 2000, and Japan is nowhere on the scene, certainly not as an invincible economic superpower. The lesson we need to take is that innovation has made us great, and we are still the best at fostering creativity and innovation. That is our strength and will continue to be so." And that is precisely what makes the issue of globalization so difficult to get one's arms around. The solutions being bandied about--protectionism, job quotas, trade restrictions--are precisely the policies that tend to discourage innovation. However, no one knows where and when innovation will rise up. In 1990, very few could have predicted the Internet would have such a vast impact on the world economy. And it is difficult to predict what the next big thing will be. Nanotechnology? Biomedical advances? Wireless systems? Unnikrishnan points out that the outsourcing of lower-end jobs within science and engineering is a natural occurrence. "There was a fear that computers would lead to automation and cause unemployment," he says. "That never happened. What happened is that low-end jobs were taken over by computers, freeing people to do high-end jobs." Outsourcing, Unnikrishnan contends, raises companies' profits, providing more money for research and development and ultimately raising our standard of living. "As long as the United States continues to remain ahead in leading-edge technologies, new jobs will be created naturally," he says. That, in essence, is the challenge for engineering educators within the global economy. Technology has changed to allow work to be done without the traditional boundaries of the workplace, in countries with lower wages. Some critics of the trend have called outsourcing "the manifest destiny of labor." But as long as innovation and creativity are fostered within the nation's engineering schools, newer technologies will be created, jobs will follow, and the economy will hum along. Certainly, some changes would be welcome to better prepare students within the global economy. But the real mission of science and engineering education--creating an environment for students to dream and innovate--is still very much in the forefront of what engineering schools should do. Outsourcing of jobs in the new global economy will not change that mission. Dan McGraw is a freelance writer based in Austin, Texas. He can be reached at dmcgraw@asee.org.