Specific technologies of individual research and development themes have already been reported in many forms including by this RWC NEWS, so I shall not address them in this article. What interested me was a common trend in how research is conducted, which I perceived by observing the results and progress of various research programs. This trend could be called "Commodity-Orientation". Allow me to explain my views.

What are Commodities?

The word "commodity" entered common usage in Japan quite recently. While its origin might be "things useful", its primary meaning in modern English should be "sellable goods". It means essentially goods rather than service, i.e., physical goods.

However, in the computing community in recent years, the word often means "goods easily available in the market" including intellectual property. Fruits of forefront research naturally take long before being used practically, but many terms like commodity operating systems, commodity database systems, commodity programming language systems, and so on are now being used. Commodity programming languages (not the systems) may even be used. Programming language systems can be considered as goods, but the "languages" themselves cannot. Thus, "commodity" here is not limited to goods in the market, but rather means "things readily available and usable because of their widespread nature" in a broader sense. Phrases like "commodity theory" or " commodity paradigm " may soon emerge.

The largest benefit of commodities is, as they are defined, that they are available at any time needed. Physical goods can be bought in the market if they are available, and even if not, those that will become available within a few months manufactured with readily-available production systems may also be included in the definition of commodities. Another benefit is that they are relatively inexpensive due to mass-production, as they are widely used. Even for applications with only integer operations, ready-made microprocessors in the market are preferred over custom-made processors without floating-point operators with fewer circuits because of their lower cost.

The benefits of commodities do not stop there. Since they may involve "language" or "theory", their benefits do not solely rest on their availability or prices. Such intangible things without ownership placed on them derive merits from the fact that they are widespread. Because many people already have knowledge about them, there is a common background knowledge, thus enabling smooth communication and eliminating much of the need for new education about them.

Commodities and Forefront Research

How should commodities be positioned in forefront research and development activities? The benefits of commodities discussed so far concern cost. Since the cheaper the research cost, the better, it is natural to use commodities in most research environments and experiment materials used routinely in laboratories. For instance, using expensive workstations where regular PCs would suffice is like ordering customized pencils for taking notes for research.

What about commodities as targets of research and development efforts? Small improvements in so-called commodities may not be considered forefront research. Those available in the market have established production technologies and are expected to be improved gradually without any efforts from forefront research institutes. Special attention should be paid on research concerning so-called "high-tech" commodities to avoid losing sight of their original aim.

In contrast, it is important to explore potential commodities, especially in engineering-oriented research. To narrow down the targets of research for this purpose, we must predict what and how technologies will develop in the future with a wide perspective and deep insight, which may not be easy but is essential for research that will have a big impact in the future.

At this point, unlike commercial research and development activities, forefront research does not have to focus on commodities that will soon become useful. However, even the development of good technologies with a reasonable potential of becoming widespread some day but expected to become commodities in the far future should have lower priority from an engineering point of view. Even for the far future, research itself is important, yet during this long period, the underlying technologies and R&D environments are improved, so research should be conducted in the future rather than now for much lower cost. Furthermore, wasted research investment can be avoided because as time goes by, the certainty whether the technology becomes commodities or not will increase. Pioneering research should not necessarily be rated highly, rather, timeliness is essential. Da Vinci's research on flight can not be considered to be beneficial to many in modern times.

Research on How Commodities are Used

Another trend in the relation between commodities and forefront research is research on how commodities are used. Even if commodities themselves are not advanced, their uses may be. In particular, the uses of those positioned as new commodities in recent years have not been studied fully. For information processing technologies, there are many goods and technologies that are recently becoming commodities because of their widespread nature, therefore forefront research in this direction is important.

As long as forefront research remains advanced, it takes a long time for it to enter practical use. Meantime, commodities used will change. Therefore, it is important, in research on how commodities are used, to conduct research by clearly understanding which current commodities are essential and will remain essential, and then to utilize this understanding. Research conducted without such insight and with fortuitous characteristics of the current commodities as their core is not of significance.

Commodity-orientation in the RWC Project

The RWC Project entered its second five-year period in FY1997. At the start of this period, the research direction was to narrow down the exploratory research subjects from the first period in order to establish key technologies for developing the next-generation fundamental information technologies.

During this visit to RWCP, I felt that there was a strong relation among the many researches in the project with commodities. The objective of creating commodities in the future is quite natural as the policy of the project. I also felt that existing commodities are being exploited as much as possible and many different technologies are being built on those commodities. Using commodities as a tool may be motivated by research cost requirements, while research on how they are used may also be considered, especially in research on PC Cluster, system software, and application software running on it.

Such research has the risk of being considered as a compromise with commercial technologies, and may not progress if conducted without full awareness of the direction of the research. However, the effective use of commodities is critical for the next-generation information technologies. In particular, forefront research should have a clear understanding of the future direction of commodity technology, and should develop technologies for effective use thereof, unlike those spontaneously nurtured by the market. I hope that the RWCP will continue research based on appropriate prediction of the future with a clear understanding of the essence of commodity technologies.