مدیریت و بررسی اثرات پروژه "تحقیق و توسعه" در MITI : مطالعه موردی از پروژه ابر رایانه

This report analyzes the results of the Supercomputer Project, which was executed as a government-sponsored R&D project starting in 1981. The conclusions of this paper are as follows: (1) There was not much need for the government to carry out the Supercomputer Project on a national-scale because three companies had already decided to introduce supercomputers and were ready to implement R&D for practical use when the national project was inaugurated. Therefore, there was little room for the government to intervene in this matter. (2) One possibly appropriate way to evaluate the quality of this project would be on the number of paper citations resulting from it. There were fewer for this project than for similar computer projects. We can therefore judge that this project had relatively little effect. (3) A high-speed computer with 10GFLOPS, one of the objectives of the project, was successfully made, but the devices that were developed to replace silicon have never been applied to computers. As stated above, we cannot say that the project has proven to be successful. However, we discovered that devices to replace silicon, such as JJ devices and HEMT devices, were not suitable for computer use. They were, however, used with mobile phones and high-speed devices for satellite broadcasting instead, resulting in a large profit. When evaluating an R&D project, it is important to evaluate the accomplishment of the objective set before the project is launched, but the indirect effects, which could not have been anticipated, also have to be evaluated.

Many people believe that Japan's sharp economic growth in the past was due to MITI's industrial technology policies.1 The policies include government-sponsored R&D projects, which attract a great deal of attention.2 The project involving the High-speed Computer Systems for Scientific and Technological Use (herein referred to as the Supercomputer Project) is one of the government-sponsored projects, established with the objective of improving and speeding up the computer's performance. This project ran from 1981 to 1989, and with the Scientific Computer Research Association as the leader, companies that joined the association and the Electrotechnical Laboratory carried out the R&D. The MITI coordinated the overall project and provided capital.

There was not much need for the government to carry out the Supercomputer Project on a national-scale because three companies had already decided to introduce supercomputers, and they were ready to implement the R&D for practical use when the national project was inaugurated. There was not much space in which the government could intervene. However, the R&D about devices, such as JJ devices to serve as substitutes for silicon, had a high uncertainty, and so there was space for the government to intervene in the R&D in this area. An analysis of the quality of the project based on the number of paper citations shows that the Supercomputer Project was quoted fewer times than other similar computer projects, and so the effectiveness from the project could be judged relatively small. We can also see the spillover of this project by examining where the patents gained from this project are actually being used. As a result of the project, 271 patents were registered, but contracts to use only four of them were made, and all with one company that provided researchers to the Supercomputer Project. Therefore, as far as can be seen from the patents, there was almost no spillover to other companies that didn't take part in the project.