• Telecommunication Systems, • Information Processing Systems, and • Electronic Devices. Oki is one of the largest manufacturers of telemetry and telecontrol equipment, including aircraft and marine control systems. Also, they are leaders in navigation systems for automobiles as well as for marine applications. The Japanese submersible Shinkai is equipped with Oki developed acoustic navigation systems and underwater imaging sys tems. Software activities are associated with many aspects of Oki's business, but especially with communication software such as PBX, networks, switching systems, ATM systems, etc. Many of these systems must work in real-time. We were told that a recent product had about 1.2 million lines of code, and another PBX software project had 400,000 lines. Currently Oki has about 6,000 employees that are considered software engineers. (Willey explained that the corresponding figure for Motorola is about 8,000.) We visited the research activity of Oki Electric Industry Co., Ltd. at 10-3, Shibaura 4-Chome, Minato-ku, Tokyo 108, Japan and we were hosted by Mr. Yasufumi Nanae Director General Manager, Corporate Software Planning and Engineering Div. Mr. Teruo Uchida Ms Atsuko Fujitaka Planning Staff, General Planning Office Tel: +81 3-3454-2111 Fax: +81 3-3798-7685 Willey presented an overview of Motorola's activities and a general summary of an earlier Motorola team trip that he made to Japan to study software production methods. Mr. Nanae, who also had a copy of Michael Cusumano's book, expressed his opinion that the Motorola teams' overall findings with respect to the companies they visited in Japan were essentially the same as those Dr. Cusumano expressed in his book. He also confirmed that measured software productivity improvements ranged 7 to 13% annually, but that such gains would not continue without new tools. Willey feels that what is now needed is to look beyond the now-famous software facto ries described by Cusumano to discover what else is happening in Japanese software. Mr. Uchida presented materials that described OPENS (Oki Software Production Environment Enhancement Strategy). Essentially, he presented an update to materials that were originally presented in May, 1991 in the OKI Technical Review, which was summarized in the report mentioned above. Oki's Software Improvement Strategy Mr. Nakae reported that in April 1992, he had been appointed to his current position as General Manager, Corporate Software Planning & Engineering Div. at Oki. This was a newly created position to serve as the focus for software improvements. He was responsible for the Software Production Engineering Center oper ated by Mr. Uchida, and also for the technical training and education program for software professionals at Oki. Willey concludes that this indicates that Oki, like many other major Japanese corporations engaged in software development, has made a decision to commit significant resources for the improvement of their software production resources. As a Oki corporate Director, Mr. Nakae would have significant visibility, and appears to have been given a charter with enough resources to make it happen. For example, we were told that Oki's goal is one workstation for each engineer. Oki Software Production Environment Enhancement Strategy (OPENS) Oki has adopted a very ambitious long-term plan to develop appropriate environments for its software engineers. It is clear that several years, thought, have gone into the evolution of the ideas contained in OPENS, and real corporate commitment exists. OKI's presentation materials covered the following: OPENS Goals: - Establishment of Advanced Software Engineering Environment - Distributed and Seamless - Effective Management OPEN for OKI group OPENS is built on the SENNA Oki provided us a graphic that demonstrated a 3-D view of an environment for engineers. The SENNA Model resembles, in many respects, MCC's Leonardo-in fact, Les Belady has clearly had a significant impact on their thinking. (Belady has an article on Leonardo in the Oki Technical Review mentioned above.) Willey worries that "we seem to be in danger, here, of having the good ideas we paid to have MCC develop for us used by our global competitors before we have effectively implemented them ourselves." OPENS features a "development support system" The development support system is tailored for four styles 1. Universal Method Environment: Constructing with general purpose concepts 2. Reuse Method Environment: Customizing (Modifying) similar systems 3. Mass-Production Method Environment: Domain (Target)-specific Automated Systems 4. Porting Method Environment: Porting to other machines While Oki talked about the different Method Environments that would be supported by using the SENNA Model tailored for these styles, our impression was that these were conceptual, and had not yet, in fact, been implemented for many of Oki's software developers. We were given some examples of the kinds of projects that each of the four styles would be used for. No. 1 was used for medium-sized projects such as switching or trans nel (NWT) project and Computational Fluid Dynamics (CFD) at the NAL, Japan". It is an excellent historical description of the work at NAL including their computing experiences and general directions of CFD research. This is in English, so I will not summarize it here except to point to NAL's ideas about a NWT, a project to enhance their computing power with a system focused on their specific needs. (Readers can write to me or to Hirose for full copies of this paper.) Requirements estimation for NWT: 1. Main memory a. Grid points: (1) For full configuration analysis with engine, flaps, and SRB. 5-15E6 points (2) For practical application of LES 150E6 points b. Data (1) Conventional TANS analysis (perfect gas, alge RESEARCH CENTER FOR ADVANCED SCIENCE AND TECHNOLOGY ACTIVITIES IN AI AND EXPERT SYSTEMS AI and Expert System activities at the Research Center for Advanced Science and David K. Kahaner This report is another spinoff from my March 1992 visit with the Japanese Technical Evaluation Center (JTEC) team, see [hitachi.ai, 25 August 1992]. RCAST was established in 1987 in an extraordinary effort to develop a new (for Japan) kind of research organization. RCAST was to grow under four guiding principles, 1. interdisciplinary studies, 2. international cooperation, 3. mobility and flexibility of staff and research areas, and 4. openness to the public and to other organizations. It has four main departments with 19 regular chaired professors, as follows: • Advanced systems department Biomechanics chair • Advanced devices department Optical devices chair Biomedical devices chair High-speed functional devices chair Quantum microstructure devices chair Intelligent sensing devices chair <-- earlier (Tachi) visit • Advanced materials Chemical materials chair Molecular information • Socio-technological Ethics of science and In addition there are four guest chairs (currently functional materials dealing with interface phenomena, new laser devices, system tero-technology, and technology assessment), and eight endowed chairs, as follows: Frontier chemistry (Hitachi) Telecommunications (NTT) Future system analysis Lab facilities are augmented by a 3-m wind tunnel, a process center, which contains epitaxy and lithography rooms, a bio-clean room, and a running test center. In addition to fundamental research, RCAST is also involved in beginning an interdisciplinary graduate course that will be open to the entire public. There are about 80 faculties at various levels (53 professors), about 150 graduate or research students, and about 140 research fellows. The 1990 budget was approximately 1.8 Billion Yen, almost US$15M. RCAST is adjacent to one of the standard University of Tokyo campuses just outside the city's central ring. It has an attractive, weathered, academic appearance. Because its location is less congested and much greener than that of the University main campus, RCAST would be a very pleasant place to spend a sabbatical or research visit. RCAST publishes two useful periodicals, the RCAST News (twice a year), that contains introductory |