TRON Project Leader's Opinion

Ken Sakamura

The University Museum, The University of Tokyo

We will soon be greeting in the 21st century. The computer was born in 1946, in other words the middle of the 20th century. With the 21st century imminent, computers are rapidly penetrating society and they exist as indispensable items in daily life. One hundred million personal computers are produced annually, and, according to one view, 50 billion embedded microcomputers are in use. When we consider this over a vast span of time, the penetration of computers into human society probably occupies a very important position in human history along the lines the beginning of agriculture, the invention of printing technology, or the Industrial Revolution of the 18th century. The exact time for changing to a computer society has overlapped with the transfer point from the 20th to the 21st century.

The unassuming computer that everyone is using today has undergone great changes over the last 10 years or so. In the 1980s, the personal computer was mainly used independently, in other words, as a stand alone machine. However, today connecting personal computers to networks is the norm. In the 1980s, personal computer communications was something that curious people did, but today no matter what the kind of personal computer, it is connected to the Internet. Connecting personal computers in a company via a local area network (LAN) has become natural, and there are more than a few people who have installed a LAN at home. Up until the last five or six years, even among persons concerned with computers there were not many who possessed Internet addresses, but today it has come about that everyone has a mailing address, including even people with no interest in computers.

However, computers themselves were developed with emphasis on independent operation. The microcomputer also was like that, and the personal computer too came into being under the assumption of independent operation. Microsoft's operating system was also very late with network support. It is well known that Novell grew by patching up that gap. Ethernet, which is a synonym for LAN, was commercialized in 1981, the same year the IBM PC, the ancestor of the personal computers now coming into wide use, was born. Because the IBM PC was launched into the market as a machine for processing individual, undefined jobs, there was nothing that could be done with the level of technology at the time, but when we consider it from the present, I think it would have been been better if it had been constructed with a network-oriented architecture from the beginning. The Xerox Star office workstation, which took network support into consideration, was launched into the market simultaneously with Ethernet, but for a variety of reasons, it did not become the ancestor of today's personal computers (that concept was taken over by Macintosh and Windows).

When the computer takes on an indispensable existence like this in society, numerous problems that have not been experienced up to now begin to burst into the open. There are often cases of things that were not a problem in the previous age not being in agreement with the new age. In an age when memory was expensive, the years of the western calendar were handled using two digits in order to conserve memory inside the computer. Because this was passed down to the present, there is the computer year 2000 problem in which computers will become unable to distinguish 2000 from 1900. Since this is also something that will happen a half year from now, it is being reported on a daily basis in the mass media. In February, when Intel announced it was introducing an unique ID number called a processor serial number (PSN) to each processor from the Pentium III, an opposition movement was unfolded by people who were anxious about protecting user privacy. Although Intel has explained that the ID number function can be turned off, the voices of opposition have not stopped. As for electronic mail also, there has never been anything so convenient as a means of data exchange among researchers up to now, but today, when it has come to be used by people daily, various problems have appeared one after the other. Strange sales mail arrives daily. Where there are people to whom extraordinarily large files come attached, mail with viruses also gets sent along nonchalantly. And with the mailing list also, if someone improperly handles it, many people are annoyed by the receipt of large amounts of mail.

When new problems are created in this manner, it is a matter of what to do. Since the social systems are different in the U.S. and Japan, both the approach to and methods of solving problems are different. If you ask where, in the U.S., a problem solving method often taken is to get rid of a the causes of a problem with drastic action as quickly as possible after a thorough analysis of matters. Court cases are also casually launched there in daily life. In the U.S., as expected, many lawsuits have been launched due to the year 2000 problem. It has been reported that the amount of compensation when taken together reaches as much as $1 trillion. So how is Japan in regard to problems such as these? In contrast to the U.S., where they thoroughly squeeze out the pus and peel off the skin, in Japan, we prize overall harmony, and thus we frequently take an incremental approach that is the least painful. Things occasionally go well with this. However, with the financial crisis resulting from the collapse of the bubble economy, while the U.S. quickly cleared up the problem with harsh methods, in Japan, the problem ended up becoming worse due to putting things off to the future. Although the U.S. on occasion takes a firm approach, generally there is the intention to solve problems in a forward-looking manner. If you ask what is the situation in Japan with computer-related problems, the situation is that we put all our energy into desperately pursuing what is taking place in the U.S.

By the way, in the world of computers, real time systems have been getting more and more attention with the aid of the information appliance and digital appliance boom. Today, Sun Microsystems' Jini is drawing attention as an architecture that makes it possible to connect all sorts of equipment into a network. However, this was not originally the case in the U.S. With TRON, from the earliest stage of the project, we have proposed a vision of cooperative distribution in which computers would be put in everyday items, which we would then link together with a real-time control network. This is coming close to practical application, but when all sorts of things are connected in a network, just as with the year 2000 problem and the processor ID problem, problems appear that must be taken into consideration. For example, when many nodes are connected together, there will be breakdowns, so how do we insure fault tolerance of the system in its entirety? What do we do if we start a loop operation through the mutual action of networked equipments? What do we do when networked equipments are the cause of a fire that breaks out or an accident that occurs? And if the cause of that is an instruction from the outside, what do we do? Furthermore, in the case of networks of things, if the way we send data is dependent on something like TCP/IP, which is used for sending data via the Internet, then it will not necessarily be efficient. There are a mountain of problems. Because TRON proposed the networking of things from early on, I believe that we should not pursue the situation in the U.S., rather we should take the initiative and go on to solve the various problems.

The above opinion piece by TRON Project Leader Ken Sakamura appeared on page 1 of Vol. 57 of TRONWARE. It was translated and loaded onto this page with the permission of Personal Media Corporation.

Copyright © 1999 Personal Media Corporation

Copyright © 1999 Sakamura Laboratory, University Museum, University of Tokyo