Michael A. Harrison, University of California, Berkeley (Panel Chair)
Edward F. Hayes, Rice University
James D. Meindl, Rensselaer Polytechnic Institute
James H. Morris, Carnegie Mellon University
Daniel P. Siewiorek, Carnegie Mellon University
Robert M. White, Department of Commerce
To assess Japanese technology in advanced computing, the panel divided the subject into electronic components, data storage, computer architecture, software, computer/human interface and multimedia, and supercomputers. The panel obtained a baseline of U.S. accomplishments in these areas by reviewing literature, attending conferences, visiting laboratories, and discussing the subject with specialists. The panel then spent a week in Japan visiting five university sites, sixteen industrial sites, one consulting company, and nine government laboratories.
summarizes the positions of the U.S. and Japan in advanced computing.
Japan has made a significant long-term commitment to information technology, from research through commercialization. Policymakers, aware that Japan would have difficulty being self-sufficient in food and in energy, decided as early as 1955 to meet international competition and make international contributions by cultivating information as a resource. Japan would draw on a highly educated and motivated labor force to promote information-related, knowledge-intensive industries. Japan has implemented this goal through national programs. Industrial strategies have been coordinated, and MITI introduced a series of multi-year plans devoted to achieving excellence in information technology.
Japan's success in information technology is due in large part to its support of industries in the allied technologies -- advanced semiconductors, chip-making technology, data storage devices, and so forth.
Japan's Position vs. U.S.: Advanced Computing
Japan's success in the computer industries has led to significant market share; the profits have been reinvested in R&D, and Japanese capital expenditures have remained high. Thus the panel expected that the Japanese competitive position would remain strong for at least the next five years. Whether the U.S. could maintain its competitive position would depend on whether the U.S. was willing to match Japan's rate of investment.
The panel found Japan relatively weak in software but effective in software engineering. There was a serious shortage of talented software people who could be hired to work in the large, high-technology Japanese companies, partly because many young people chose to work for higher salaries in the financial community. Japan had nothing yet to compare with the strong community of creative and talented software people in the United States.
Japanese universities remained substantially weaker than their U.S. counterparts because they have had no large projects of the type supported by DARPA in the U.S. Japanese students graduated from universities with a good conceptual education. The companies then provided continuing education to train them in design, production, and so forth. Employer-sponsored continuing education in the U.S. was much less intensive and effective because of employee mobility.
A key theme in Japan was internationalization. Japanese companies were using the profits from their success in consumer electronics and other information industries to establish themselves in the U.S. and elsewhere. Individual companies were establishing R&D laboratories, product development laboratories, manufacturing facilities, and sales and distribution centers in the U.S.
Table 16 shows Japan's position in electronics components by indicating the number of years Japan is ahead of (behind) the U.S. in various areas.
Japan's Position vs. U.S.: Electronic Components
The panel qualified the findings in Table 16 by noting that the interval between an R&D announcement and commercial production was typically smaller for U.S. companies than for their Japanese counterparts. This tended to exaggerate the gap between the countries' positions.
compares the two countries in data storage. Most Japanese industrial research focused on near- to medium-term issues. The panel found an enormous amount of exploratory work being done on alloys for thin film media, tribology, magnetoresistive sensors, and so forth. By comparison, efforts in the U.S. appeared more fragmented but more adventurous -- for example, the holographic storage at MCC and attempts to exploit high-resolution scanning microscopy.
Japan's Position vs. U.S.: Data Storage
The Japanese were experimenting with a vast number of computer architectures. Although their projects were based on American architectures, the gap between the American roots and the first Japanese project had narrowed from over a decade (i.e., from the American Illiac IV in the mid-1960s to the Japanese PAX in 1977) to less than a year (i.e., hardware simulation engines). Furthermore, although the number of advanced architectural projects was roughly equivalent in the U.S. and Japan, the sheer volume of Japanese projects initiated since 1980 was very impressive.
The U.S. was ahead of the Japanese in computer architecture. However, the Japanese were strong and growing stronger in hardware, prototyping, vector processing and pipeline design, dedicated hardware simulation architectures, multimedia workstations, and technology transfer between research and products.
Except in software engineering, Japan has traditionally been weak in software, as is shown in Table 18. Although Japan has improved significantly in graphics, logic programming, and artificial intelligence applications, so has the rest of the international community. Ironically, the panel found that Japan had the lead in software engineering. U.S. researchers were conducting better software engineering research, but the Japanese were applying U.S. methods in a more disciplined fashion and achieving impressive results.
The panel found the U.S. to be significantly ahead in computer/human interfaces, although the Japanese were beginning to concentrate in that area. In multimedia systems, the Japanese were ahead in hardware technology because of their significant consumer electronics industry; the U.S. was far ahead in software applications. Table 19 shows the panel's rankings in multimedia systems.
Japan's Position vs. U.S.: Software
Japan's Position vs. U.S.: Multimedia and Computer/Human Interfaces
records the panel's impressions of Japanese research in supercomputers. In most areas of computational science and engineering, the number of researchers in Japan was smaller than that in the U.S. by a considerable margin. However, the numbers were growing in each of the fields surveyed.
The panel predicted that for the next five years the U.S. would continue to have more researchers working in supercomputers and scientific calculations. If U.S. researchers continued to have access to enough state-of-the-art supercomputers, the U.S. would continue to provide leadership in developing new approaches, algorithms, and software.
Japan's Position vs. U.S.: Supercomputers
In the field of advanced computing in general, the panel found Japan to be ahead of the U.S. in basic building blocks such as chips and components. The U.S. predominated in software. However, revenues for software development could not be compared to those for the manufacture of electronics, and so forth. Therefore, the panel predicted that Japan would continue to have both market share and profits, which would fund R&D.
The panel judged the United States' investment in advanced computing R&D unimpressive. Because future government funding was uncertain, industry has been left with an increasing responsibility for funding computer-related R&D. IBM has taken a leadership position in forming cooperative ventures, although some collaborative ventures had not lived up to expectations. Therefore Japan's position in advanced computing hardware could become dominant unless new initiatives are undertaken.