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

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

OCR for page 3
CONCLUSIONS AND RECOMID3NDATIONS CONCLUSIONS Computer science has a much more strongly developmental character than most other sciences, and the developmental consequences of the few inventions most central to it (the stored-program computer, the integrated circuit chip) have not yet been exhausted. As computer technology has matured and become more complex, leadership in carrying out large projects has shifted from universities toward industry. It is thus particularly important that those involved in university research keep in close touch with the industrial and commercial arenas out of which much essential innovation flows. Universities are good at abstracting, codifying, analyzing, disseminating, and inventing those areas where aesthetics or mathematical tradition is a sufficient guide to what needs to be done. However, an insufficiently broad familiarity with rapidly changing technology often limits the applicability of university research. In attempting to overcome this deficiency, university computer science departments have sometimes involved themselves in substantial development projects. This is not to be discouraged, but universities should remember that when they undertake development projects, their aim must be achieving new levels of understanding of principles basic to the management of important real problems; conceptual simplification and clarification of structured approaches; and transmission of organized ideas and new principles to the larger world. ~ should maintain their emphasis on long-range problems rather than shift to short-term development work. m ey must also avoid involvement in "miniproducts" that are unable to compete with more advanced industrial products. Also, their work should concentrate on domains in which they have sufficient experience to discern pragmatically vital issues correctly. Part of the panel's effort was directed at examining the university equipment situation vis-a-vis that of industry. m e panel found that the best equipped universities have computing facilities comparable to those of industry, provided that only raw cycles are measured and availability of floating point capacity and terminals is ignored. However, few universities have access to the large specialized equipment available at strong industrial research laboratories, e.g., 3 Universities

OCR for page 3
4 computers instrumented as performance test stands, robot manipulators, speech analysis hardware, and state-of-the-art computers for carrying out large numerical calculations. RECOMMENDATIONS 1. The Panel recommends further steps to encourage university- industry interaction. Possible initiatives in this direction are an follows: · increased funding for joint university-industry projects · funding for sabbatical visits to or from industry, emphasizing new courses · direct support of graduate students pursuing doctoral research in industry . and industry organization of special research grants funded jointly by NSF 2. To satisfy the important unfulfilled requirements of the scientific computing community and the growing needs of robotics research, and also to strengthen the ability of university departments to supply industrial researchers trained to a high standard, university access to computing facilities needs to be improved. The aim should be both to bring more schools up to the industrial Standard and to provide specialized equipment. 3. Tb contribute more substantially to manufacturing (as distinct from information-based) industry in general, and to the new field of robotics in particular, university computer science training should include a higher proportion of classical applied science and mathematics than is typical today.