From: cyberoid@milton.u.washington.edu (Robert Jacobson) Subject: IND: Kahaner Report: MITI Briefing on Real World Computing Date: 31 Jan 92 17:59:24 GMT Organization: Human Interface Technology Lab, Univ. of Wash., Seattle Crossposted from comp.research.japan, wherein one can find the summary text for the MITI Briefing. You can also ftp it from the U of Arizona archives, as indicated: [Dr. David Kahaner is a numerical analyst on sabbatical to the Office of Naval Research-Asia (ONR Asia) in Tokyo from NIST. The following is the professional opinion of David Kahaner and in no way has the blessing of the US Government or any agency of it. All information is dated and of limited life time. This disclaimer should be noted on ANY attribution.] [Copies of previous reports written by Kahaner can be obtained from host cs.arizona.edu using anonymous FTP.] To: Distribution From: David K. Kahaner, ONR Asia [kahaner@xroads.cc.u-tokyo.ac.jp] Re: MITI Briefing about Real World Computing Program, 29 Jan 1992 30 Jan 1992 . This file is named "rwc1-92.brf" ABSTRACT. A summary of the briefing given by MITI to foreign participants concerning the organization of the Real World Computing Program. >From October 1991, Japan's new ten-year advanced computing program has been renamed from NIPT (New Information Processing Technologies) to RWC (Real World Computing). See my report on that subject, nipt11.91, 11 Nov 1991, and references included there. On 29 Jan 1992, representatives of MITI held a briefing to explain to potential foreign participants their plans for how the program will be organized and managed. The briefing did NOT address itself to the technical aspects of the program. A draft of the technical issues (Master Plan) has also recently been circulated. A second meeting of the Feasibility Study Committee is scheduled for 18 Feb 1992, as is a two day workshop 2-3 March 1992, at which time the final Master Plan, including both technical and administrative aspects will be described. Two MITI representatives, Hidetaka Fukuda Deputy Director Industrial Electronics Division Machinery and Information Industries Bureau MITI 1-3-1 Kasumigaseki, Tokyo 100 Tel: +81-3-3501-1074, Fax: +81-3-3580-6073 and Yoshiki Mikami Director Information, Computer and Communications Policy Office MITI 1-3-1 Kasumigaseki, Tokyo 100 Tel: +81-3-3501-2964, Fax: +81-3-3580-6403 made the presentation, which was attended by Science Counsellors and/or their representatives from Australia Canada France EC Germany Korea Netherlands Singapore UK US Representatives from Italy and perhaps some other countries were invited by did not attend. It was emphasized that the briefing was for the purpose of explaining the organization, rather than for consultation, although it was impossible for MITI to prevent a dialogue from occurring, as there were many details which were not spelled out, and perhaps not even known yet. The US Science Councellor to Japan, Dr. E. Malloy, pointed out that the US has decided not to participate in this program. Instead the US and Japan are in the midst of a feasibility study (to be completed Sept 1992) about the possibility of cooperation under the Joint US-Japan Science and Technology Agreement, of a rapid prototyping optical fabrication facility, modeled somewhat along the lines of the MOSIS project, currently run by USC for silicon devices. The US considers this an activity which is independent of the RWC program. Japan on the other hand may use part of the RWC budget to support the program if it goes forward. This caused a great deal of confusion among the foreign participants, who wondered to what extent budgets will be siphoned off for this project, and whether RWC organizational plans will be affected by it. In any case the Japanese emphasized that the US-Japan program is still under study, and if it goes forward, will be a small project. There was a long discussion concerning how foreign governments would provide input to the organizational issues after the current briefing. A working group meeting has been scheduled, but at today's meeting it was announced to be only for MITI staff. Several foreign representatives raised the point that MITI's earlier documentation specifically included foreign participation in the working group. Eventually, MITI agreed to consider the possibility of allowing foreign attendees, possibly as observers or participants. It was obvious that they were concerned that a large meeting with many nations represented would not accomplish much, especially insofar as the topics relate to details of Japanese contract law. (The original agenda of the working group meetings does not mention the Japanese legal system.) One participant explained to me that his concern was that "no change has been announced at all to the people, to whom the original plan had been announced." On the other hand the Japanese have made it quite clear that RWC is considered to be a Japanese domestic program which is supposed to have international participation and cooperation. In other words the Japanese are putting up the money. My own opinion is that they have every right to set the rules. Other countries can participate or not, as they see these rules as being in their self interest, or at least fair to their country. A decision on foreign participation in the working group is to be made by MITI next week. BASIC PHILOSOPHY AND POLICY It was emphasized that the primary goal of RWC is NOT to develop a computer but instead to explore the basic technologies that are not yet established but thought to be significant to the general area of flexible information processing. On the other hand, several forms of new computer technology, including optical, neural, or massively parallel, are seen as providing the computational basis for the actual work that is going to be done. One important aspect of the computer work will be system integration. The computational hardware will provide the tools for higher level theoretical foundation work related to the basic theory of flexible information processing, or soft logic, which is viewed as the logic system carried out (unconsciously) by humans. This in turn will allow research in higher level (but still elemental) functions such as recognition, understanding, inference, problem solving, autonomous and cooperative control, simulation and human interface. At the top level it is hoped that these functions can be integrated in an advanced way to provide genuinely "flexible" information processing. Comment. The soon to end 5th Generation project made very sweeping promises ten years ago. Many of these have not yet been achieved. What has come out is more like the "computational basis", i.e., the bottom level. Will the same happen to RWC in ten years? MITI has established five policies for the RWC program. * Flexible design of research organization. MITI realizes that the program has many vague points. Thus to pursue a wide range of research targets they will establish one central laboratory and some distributed laboratories, along with a flexible collaboration between them. The central lab will engage in research in important common themes, as well as integration of the results from research conducted at the distributed labs. * Promoting cooperation and interaction between researchers working at the bottom level (providing computational base), theory, elementary functions, and those at the very top level will be strongly encouraged. The disconnect between the topic areas was widely commented on during the November meeting in Yokohama. MITI promises to promote smooth liaison between the groups during this program. * Stimulate competitive research. MITI admits that they do not really know the right approach to many of the problems they want to solve, so at least for the first half of the ten year program, competitive approaches to the same target will be tested by different groups and after five years, the results will be evaluated and research topics will be selected and reorganized. * Promote international and interdisciplinary cooperation. Again it was emphasized that this is basically a Japanese domestic program. However international and interdisciplinary cooperation will be necessary. MITI expects that the central lab will play an important role in implementing international (horizontal) and industrial-governmental- academic (vertical) research cooperation. * Better publicizing of the research results. Progress and results of research should be openly reported to the public at conferences, and symposia and workshops for distributing and exchanging information will be held frequently. A question came up about the language of technical results. MITI pointed out that as much as possible they would encourage researchers to use English and that they would provided translators and translations, but that many of the results would be expected to be in Japanese (again Japanese domestic program). ORGANIZATIONAL SCHEME MITI views the organization as one that describes how both money and information flow. * MITI will provide money directly to the ETL (Electrotechnical Laboratory), which is the government arm of the research program. * MITI will help set up a RWC Partnership. Partners can be either foreign or Japanese individuals or corporations. Partners will pay an initiation fee. The role of the Partnership will be to generate annual research plans, provide the R&D infrastructure including the facilities of the central laboratory, networks, etc., as well as manage procedures regarding sub-contracting research and joint research. Members of the Partnership will also carry out research. MITI will provide money to the Partnership. The Partnership can sub-contract and fund either foreign or Japanese companies, research organizations, universities, etc. Budget. I wrote about budget plans in my report on the Yokohama meeting. The MITI staff at the current meeting objected to these figures, saying that they represented people's personal opinions. Perhaps, but this question came up repeatedly in Yokohama and the figures I gave were repeated to us by numerous Japanese in responsible positions. Expecting people to discuss such a program without budget information is like expecting kids to go to a birthday party and not watch the birthday child open the presents. Fukuda explained that for FY92 (April 92-March 93) MITI has requested 900M Yen for RWC, including 200M Yen for the optoelectronics program. If these figures seem low, please recall that actual start of the program will not begin until spring 1993. * MITI feels that joint research can occur between ETL and the Partnership, between ETL and subcontractors, and between subcontractors and the Partnership. However, there is no funding for any of these joint research possibilities except to the extent that subcontractors will be funded by the Partnership for specific research. Joint research is thus viewed as basically sharing of information. Several issues have not been considered, such as will visitors to the central lab be charged for use of facilities, or how to handle jointly funded projects (i.e., perhaps a foreign lab funds its researchers to work on RWC-related activity, how is collaboration to be managed). PROMOTING COOPERATION AND FRUITS OF RESEARCH As I reported several times last year, in May 1991 a Japanese law, referred to informally as the "NEDO Act" was amended to encourage international cooperation. In the context of RWC this applies only to research consigned to the Partnership. (ETL is a government lab, and thus all the research performed there is owned by the government.) The changes in the law say that rights belong only partially to the contract consignee, and that details are to be decided by a separate regulation. The interpretation seems to be that up to 50% of the research will be owned by the Partnership. MITI explained that the "up to" part requires that MITI negotiate with the Ministry of Finance for the exact percentage. In addition, use of the government's share of the research can be either free or at a low price (again MOF/MITI will have to develop the details). The Partnership can make its own rules about how ownership is to divided among the Partners. There were many questions centered around these issues. For example, we were told that only overseas corporations were entitled to make use of this law, and that there seemed to be a distinction between branches of foreign companies in Japan, and XX Corp Japan. Finally, MITI pointed out that the law did NOT cover software copyrights, which apparently are still to be owned by the government. There were also questions raised about license fees, possession of patent and utility rights, and veto power over co-ownership of patent rights. I prefer not to dwell on these, because it was clear that many minor (although perhaps important) permutations need to be considered. SCHEDULE Current plans are as follows. March-April 1992: Final Master Plan announced with technical and administrative details and rough budget. April 1992: Partnership solicited June-July 1992: Partnership established (participation as a Partner after the Partnership is established is also possible). Oct 1992: Partnership invites subcontractors and recruits joint researchers, invites proposals, preparation of research facilities, labs, network, etc. End of(?) 1992: Deadline for application for subcontract proposals. April 1993: Subcontract agreements and joint research agreements are concluded.