industry. The majority of undergraduate computer science majors find employment in industry; about half of the new Ph.D.s in CS&E do so as well.19 Without exposure to ECSE in their formal education, students would be even less prepared than they are now to engage in meaningful careers in industry. As an associate professor at a large private university noted in response to the CRA-CSTB survey (see Appendix A):

The practical aspects of compiler optimization are passed on verbally or in the occasional ''engineering oriented" compiler text, while texts that focus on different grammars and parsing algorithms are considered to be more "pure." Likewise, practical software testing methodologies are often given far less discussion than impractical but theoretically elegant methods. In architecture texts, issues such as signal delays, noise, loading, power, and cost usually take a back seat to the intellectually important issues of organization, instruction sets, and theoretical or simulated performance analysis. This bias is a contributing factor to the common complaint among industrial employers that graduates have to be retrained because they have no practical experience. Obviously, academe cannot and should not recast itself into a training ground for industrial employers, but an increased amount of practical and genuine experimentation would be a benefit to the discipline.

Finally, a strong experimental component to the research and teaching programs of CS&E departments is a necessary aspect of reaching out to other academic disciplines. As articulated in Computing the Future , the future of the discipline demands in part an attention to problems with relevance to society or to other intellectual domains.

None of this argues that universities and industrial research laboratories are equivalent research environments. Industrial laboratories such as those at AT&T Bell Laboratories generally have better resources and fewer distractions, and their efforts are funded by their parent corporations specifically in the hope and expectation that they will lead to competitive advantages in the marketplace. Sometimes they offer unique advantages, such as proprietary technology or access to profiles of (the parent company's) customers' work loads. Universities offer different advantages, including the enthusiasm and imagination of graduate students and a wide freedom to select topics for study. Despite these differences, both settings have produced important experimental research ideas in recent years.

19  

Gries, David, and Dorothy Marsh. 1992. "The 1990–1991 Taulbee Survey," Computing Research News 4 (January):8.



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