« iepriekšējāTurpināt »
developing five different MIMD By no fault of ICOT's, the large systems as moving from conventional machines with differing architectures collection of different machines pro- computing (where numbers, documents, in order to compare and evaluate imple- duced at different sites is confusing for and data are processed) through Fifth mentation issues. All share one feature: observers who are not actively keeping Generation (logic-based) computing that about 10 central processing units track of the research.
(where knowledge is processed) to flex(CPUs) are connected by a single bus Please note that ICOT will host the ible information processing (where into a "cluster" within which main mem- International Conference on Fifth intuitive information can be processed), ory is shared. Clusters are interconnected Generation Computer Systems (FGCS), which is the heart of the RWC program. by a network that differs from one PIM 1-5 June 1992, in Tokyo. For informa. Examples of applications of RWC to another. The idea of a hierarchy is tion, contact
systems include incompletely specified simply to reduce the problems associated
(ill-defined) problems such as underwith trying to connect many CPUs with FGCS'92 Secretariat ICOT
standing of situations in a noisy envilow latency and high throughput.
Mita Kokusai Bldg, 21F
ronment, large-scaled problems such Five different industrial manufac- 1-4-28 Mita
as simulation of social and economical turers are developing the different PIMs. Minato-ku, Tokyo 108, Japan phenomena, and real-time problems Differences between them include Tel: +81-3-3456-3195
such as man-machine interface with machine instruction sets, number of Fax: +81-3-3456-1618
virtual reality and autonomous control CPUs per cluster, coherent cache meth- E-mail: firstname.lastname@example.org
of intelligent robots. ods, and network topology as mentioned
The main technological aspects of above. These are called PIM/p, /c, /m, --David K. Kahaner, ONRASIA
RWC are seen as follows: li, and PIM/K and were described in more detail in my report cited above.
(1) Development of the computational An earlier machine, Multi-PSI, was
bases. This includes research sepadeveloped in the middle of the project. REAL WORLD COMPUTING rately on general-purpose massively In addition to the PIM machines, there PROGRAM
parallel systems and specialare several associated projects only
purpose neural systems, optical loosely connected with ICOT. One set The Real World Computing (RWC) computing and optical devices, and is being performed at Keio University Program (alias, New Information Pro- system integration. under the direction of Profs. Anzai and cessing Technology (NIPT)) is to be a Amano, under the name (SM)^2 Sparse 10-year program by the Japanese Gov. (2) Theoretical foundations and the Matrix Solving Machine. (Alas, the name ernment (Ministry of International development of “novel” functions. notwithstanding, (SM)^2 does not Trade and Industry (MITI)] to lay the The former includes all aspects of appear to be appropriate for numerical technological foundations for the representation, storage and recall sparse matrix computation.) A single information society that Japan sees as of information, information intecluster of 20 MC 68000 processors began occurring in the 21st century. Informa- gration and evaluation, and learnoperation in 1986. More recently, the tion systems of that period are viewed ing and self-organization. The latATTEMPT (A Typical Testing Envi- as being based on several key technol- ter includes research into flexible ronment for MultiProcessor systems) ogies (as opposed to a single one), includ- recognition and understanding of is trying to find applications that run ing massively parallel computing, opti- multimodal information, flexible efficiently on thousands of processors cal computing, neural computing, and inference and problem solving on and quantitatively evaluate their behav- logic programming. The RWC program flexible information bases, flexible/ ior. ATTEMPT-1 is a commercialized aims to establish theoretical founda- autonomous control, and flexible version based on a Futurebus that tries tions for these technologies, explore interactive environment for manto emulate various types of cache proto- applications, and study how they can be machine interface. cols. The Keio professors specifically integrated. Flexible integration is seen want to study the kind of logic simula- as an important goal in order for infor- The program will be organized tion needed in designing digital cir- mation systems to be able to deal with around five fundamental policies: cuits for cell arrangement, wiring prob- real world problems. Program orgalems, and logical synthesis.
nizers view the evolution of computing
(1) A central laboratory (probably at information database, etc., as well previous conference, the second in the
or near the Electrotechnical Labo- as to provide an electronic notice series, was held at the same venue in ratory (ETL) in Tsukuba) at which board, electronic mail, and elec- January 1991. These conferences were common research will be per- tronic meeting capabilities. organized by the National Institute of formed, students and postdoctoral
Science and Technology Policy candidates trained, etc., and sev- The RWC program schedule is as (NISTEP) of the Science and Technoleral distributed laboratories (which follows. The Master Plan was completed ogy Agency (STA). In recognition of may not all be in Japan) for more by the end of March 1992. Also during the fact that science and technology highly individual research. March a RWC mission went to Europe (S&T) transcends national boundaries,
to explain the program and to discuss the conference reviewed science and (2) A competitive principle. During participation in the program either as technology policy from a global point
the first half of the program, a large members of the research partnership of view. It considered the questions number of different approaches will or as subcontractors. By the end of “how can science and technology best be supported, and during the sec- April the Feasibility Study Committee be advanced for the promotion of the ond half research themes will be will authorize the Master Plan. Some- global economy?” and “how can the evaluated and concentrated. The time during this summer (June/July policy contribute to the challenge?” program deliberately has a clear 1992) the RWC Partnership will be The meeting last year focussed on the description but only vague, multi- established. The partnership will actually issues of what should and what can be ple subtargets.
begin activities in October 1992, with a done in S&T policy research. It seemed
call for participation, subcontractors, to this observer that the papers this (3) Interdisciplinary and international etc., and with an application deadline year were better integrated with each
cooperation. Active joint research near the end of 1992. Thus, substantial other. with ETL, universities, etc. and money will not enter the program until subcontract research from Japanese late in 1992; the 1992 fiscal year budget General and foreign research organizations. is only about Y900M. MITI feels that this program repre- For general information contact The proportion of participants from sents a real change of direction for
overseas at this conference was relaits support of science: from one NIPT Workshop '91 Secretariat tively large--they numbered 33 of the that invested in near-market devel- Kikai-Shinko Bldg
140 attendees. A great many of the opment to one that was fundamen- 3-5-8 Shibakoen
participants knew each other, which tally research oriented.
Minato-ku, Tokyo 105, Japan made the discussions and interactions
much more animated than is usually (4) Publication of research achieve- Fax: +81-3-3431-4324
the case at such meetings. Thirty-two ments to promote true interna- E-mail: email@example.com
papers were presented during 2 days; tional cooperation and open report
they were divided into six sessions: ing via conferences and symposia. --David K. Kahaner, ONRASIA As much as possible documenta
• Science and technology policy tion of RWC results will be in
research English. However, it is expected that many informal and internal THE THIRD INTERNATIONAL • Public policy documents will be in Japanese. CONFERENCE ON SCIENCE
AND TECHNOLOGY POLICY • Innovation process (5) Establishment of infrastructure for RESEARCH research by setting up a high-speed
• Internationalization network. A significant network Introduction linking more than 10 locations in
• Research and development strategy Japan and a few overseas is planned The Third International Conference to give researchers access to cen- on Science and Technology Policy • Recent developments on science and tral laboratory facilities, such as a Research was held in Oiso, Kanagawa technology situations: general trends massively parallel computer, shared Prefecture, on 8-11 March 1992. The and environment
The conference closed with a round- advantage in international competition. industry with larger firms will show a table discussion at which Prof. Ikujiro While that may be correct, her treat- more rapid rate of technical advances Nonaka of NISTEP presented a sum- ment of Cray Research as a company on the approaches to innovation. Dr. mary of the highlights of the meeting. comparable with Hitachi, Fujitsu, or Hariolf Gruppofthe Fraunhofer Insti
NEC gives a very wrong impression of tute for Systems and Innovation Discussion
the competitive edge held by the Research, Karlsruhe, Germany,
Japanese. These three Japanese com- reviewed the dynamics of science-based In the science and technology policy panies are huge, broad-based companies, innovation in North America, Japan, research section the papers mainly dealt whereas Cray is a single product busi- and Western Europe. In his analysis he with models and paradigms. Drs. Kinji ness with much less capital. Dr. Kinji primarily used patent data. Dr. Diane Gonda and Fumihiko Kakizaki of Gonda of NISTEP studied the effect of Hicks of the Science Policy Research NISTEP proposed an “L” model to industrial location policies on regional Unit, University of Sussex (she spent a relate science, technology, and indus- R&D in Japan. He reported that regional year as a research fellow at NISTEP), trial competitiveness. Drs. Fujio Niwa, R&D activities are shifting from applied
R&D activities are shifting from applied examined the extent to which scientific Hiroyuki Tomizawa, and Fumito to basic research. Dr. Xiang-Hao Heof research of Japanese corporations is Hirahara of NISTEP conducted indi- the Research Center of Management globalized on the basis of the number cator analysis of Japanese science and Science, China Association of Science of published papers. She concludes that technology. They argued that Japanese and Technology, who spent last year as Japanese companies actually rely heavily basic research and scientific infra- an STA fellow at NISTEP, analyzed on Japanese sources. An attempt to structure are poor in comparison with how Japan gained international com- identify the techno-economic paradigms other developed nations. A method petitive advantage. He concluded that through analysis of the innovation for technological and new products Japanese export is vital to their eco- process was reported by Prof. Masaaki forecasting called Delphi Technological nomic growth, that their industrial Hirooka of the Faculty of Economics, Forecasting was presented by Dr. system is divided into those that are Kobe University. One of his conclusions Akitoshi Seike also of NISTEP. Science export oriented and those that are based on the analysis of the technologand technology indicators were also domestic market oriented, and that ical lag of Japanese industries was that topics of a paper by Dr. Giorgio Sirilli Japanese industry considers R&D efforts Japan reached the level of Western of the Institute for Studies on Scientific as a main driving force for gaining nations in the early 1970s. Profs. Research and Documentation of the competitive advantage. The system for Don E. Kash, Institute of Public Policy, National Research Council of Italy. His the education and professional devel- George Mason University, and Robert main indicators were research and opment of engineers was studied by W. Rycroft, Center for International development (R&D), innovation sta- Prof. F. Karl Willenbrock of the Depart- Science and Technology Policy, George tistics, patents, technological balance ment of Engineering and Public Policy, Washington University, discussed of payments, trade in high tech prod- Carnegie Mellon University. In the ideal “simple technologies,” which can be ucts, and bibliometrics.
system, he said, there would be a strong understood by individuals and whose The effectiveness of policy measures emphasis on science and mathematics
emphasis on science and mathematics value results from revolutionary innowas reviewed in the session on public in pre-college, a multi-route system in vations and complex technologies,” policy. Prof. Marie Anchordoguy of the college so that options can remain open which are difficult for individuals to University of Washington, who is cur- longer for the student, and multiple understand and whose value results from rently conducting research at entry points permitting people to re- incremental innovations. For the forHitotsubashi University, presented a enter to upgrade their skills.
mer, they said, public policy must focus summary of a study she conducted for In the session on innovation process on generating new ideas that lead to the Office of Technology Assessment Prof. Wesley Cohen of Carnegie Mellon innovations, whereas for the latter the on Japanese industrial policy and their University considered the tradeoff public policy needs are greater and more supercomputer trade and targeting between the size of a firm and its diver- complex. Prof. Fumio Kodama of policies. She argues that “government- sity in promoting technological progress. NISTEP, who is currently a visiting sponsored cartels” and cooperative He concluded that having a large number professor at the Kennedy School of R&D projects and subsidies give the of small firms will result in a greater Government, Harvard University, Japanese supercomputer makers an diversity in innovations but that an reported on the effectiveness of tech
nology transfer in large and small firms. The session on internationalization Dr. Yutaka Kuwahara of Hitachi, Ltd., concepts, and a minimal understandwas led off by Justin L. Bloom of Tech- discussed global R&D management and ing of the impact of science on society, nology International, Inc. Mr. Bloom suggested “inclusive-interactiveness” Miller reported that scientific literacy had served as the Counselor for Scien- and "visibility" as new concepts. He
and "visibility” as new concepts. He in the United States is 6.9%. Among tific and Technological Affairs at the defined the former as "internal exis- Nagahama's findings were (1) the fracAmerican Embassy in Tokyo for 5 years tence of the subject which performs tion of young males interested in “topics in the late 1970s. He reported on judgment of value and does activities and news on science and technology” is American corporate investment in based on intentions" and the latter as falling, (2) the understanding (i.e., Japanese research facilities and among “clearness in concepts, procedures, and “awareness”) of scientific knowledge the points he made were (1) direct practices.” Drs. Yasunori Baba and such as "DNA" or "lasers” reflected investment is much greater than imag- Ikujiro Nonaka of NISTEP discussed school education content, and (3) attiined, (2) U.S. firms have not encoun- the creation and transference of manu- tudes of the Japanese, Americans, and tered much Japanese Government facturing knowledge in the die and mold West Europeans towards science and resistance in making technological industry; this industry was selected scientists varied considerably. entries into Japan, and (3) Japan tends because, while it is a shadow industry, to buy R&D while the U.S. sets up it is absolutely indispensable for man- Conclusion facilities from scratch. Orlando ufacturing and thus a good indicator of Camargo, Research Associate, NISTEP, trends. They state that the Japanese As we all know, the relationship reported on foreign firms conducting model of knowledge management can between the United States and Japan, R&D in Japan. His study was based on systematically integrate “two opposing which former Ambassador Mike the results of a questionnaire survey of types of knowledge into a powerful Mansfield often described as the most senior R&D managers and senior repre- source of knowledge creation/ important in the world, has sailed into sentatives of these companies; 52% of advancement.” Prof. Kei Takeuchi of turbulent waters. Recent statements the 132 firms surveyed responded. The the Research Center for Advanced by Japanese political leaders and the biggest problems these managers faced Science and Technology, University of very acrimonious rhetoric by American were the hiring and retention of per- Tokyo, discussed “technological con- leaders have added fuel to the fire. sonnel. China's policy on international glomerates.” These conglomerates are There are a number of underlying causes cooperation of science and technology characterized by an existence of a core to this friction. Among them are the was discussed by Liu Yong-Xiang, technology and a continuity in person- hardships suffered by the Americans Counselor of Science and Technology, nel and management.
because of the recession, the trade Embassy of the People's Republic of The final set of papers was pre
perception among China. He strongly emphasized that sented in the session on recent develop- Americans that the Japanese do (or the policy of China was to actively ments on science and technology, which will) not play by our rules, cultural promote international cooperation with was divided into general trends and differences, and language difficulties. all countries with which it has friendly environmental issues. Of particular As I listened to the presentations, two relations. Prof. Jon Sigurdson, Research interest to me were the papers by Prof. other equally important factors were Policy Institute, University of Lund, Jon D. Miller of Northern Illinois apparent. One is that observers often Sweden, reported on the impact of the University, who discussed the scien- have preconceived ideas and look for internationalization of corporate R&D tific literacy and public attitudes towards and find evidence to support them. An on science and technology policy. international competition in scientific observer can make correct observations
Five papers were presented in the research and manufacturing of but can report them in such a manner session on R&D strategy. Prof. Richard Americans, and by Dr. Hajime that a wrong view or impression is Gordon, Silicon Valley Research Group, Nagahama of NISTEP, who presented conveyed. This was clearly evident in University of California at Santa Cruz, the characteristics of public understand- one of the papers presented by an based on a study of the U.S. semicon- ing of science and technology in Japan. American participant. Another is that ductor industry, reviewed linear and Using the definition of scientific liter- we Americans tend to blow hot and evolutionary models for innovation acy as having a minimal understanding cold on issues. This year one of the and global competition and suggested of the processes of science, a minimal major Japanese “bad guys” is the keiretsu, a third, “social organizational” model. understanding of scientific terms and
which is a term for the Japanese industrial cartels or conglomerates; some of the famous ones are Mitsui, Mitsubishi, and Sumitomo. We feel that the manner in which the keiretsu operate, e.g., the group bank making capital available to a member company at low interest rates, gives that company an unfair advantage over its foreign competitors. Many American speakers used the term keiretsu often; last year that term was not used at all. A few speakers gave the impression that the keiretsu evolved relatively recently, that it was nurtured by the Japanese Government to help their industry. Actually, though, the system came into being in the 17th century and has existed in one form or another ever since. Prior to and during the Second World War, it was called zaibatsu. These two examples were disturbing because the American participants at this conference were scholars who are knowledgeable about Japan and, yet, one of them reported observations that conveyed a wrong impression and several jumped on the keiretsu bandwagon. The Japanese are tough and often difficult competitors. In order for us to compete successfully with them, it is imperative that we learn about their ways and plan our actions accordingly. This requires that we study their ways objectively with an open mind.-Sachio Yamamoto, ONRASIA