rick@cs.arizona.edu (Rick Schlichting) (04/09/91)
[Dr. David Kahaner is a numerical analyst visiting Japan for two-years
under the auspices of the Office of Naval Research-Asia (ONR/Asia).
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: TRON (The Real Time Operating System Nucleus) Comments
7 April 1991
ABSTRACT. Comments from readers about 4 March TRON report.
After my TRON report was distributed (4 March 1991), several readers sent
comments and amplifications to what I wrote. The two most substantial
were from Prof. J.D. Mooney (West Virginia Univ), who was cited for his
participation in the last TRON Symposium, and Prof J. Hootman (Univ North
Dakota) who was Editor-In-Chief of IEEE Micro at the time that the TRON
articles were published in that Journal (1987). Their comments are
quoted below. Mooney has also agreed to collect further comments for
distribution. Please send any additional remarks to him.
Hootman:
" It strikes me as if the TRON concept is an ideal one for AI, Neural
Nets etc. If one is to really model the brain etc it will require a
multitude of sensors and the interaction of many different types of
systems--an ideal type of situation for the TRON. I am really surprised
that no group has started to look at that. It would be interesting to
study this and just see what kind of information was generated.
In order to really convey the place and importance of TRON I think that
it is necessary to make a table and compare TRON with something like UNIX
or other operating systems and give the good and the bad points of both.
I think that we in the U.S. tend to look at ourselves and concentrate
on the good stuff we do. We don't spend time looking at others to see
what they are doing. I bet Ken S. looked around and just made the
considered decision to do something different. The other impressive part
of Ken's operation is the support that he has from the government and
industry [only industry as far as I can tell-DKK]. This says that the
IBMs of the world are going to have to do some serious looking at TRON
and other operating systems. "
Prof Joseph Hootman, Dept of Electrical Engineering
University of North Dakota
Tel: (701 777-4428), Fax: (701-777-3650)
Mooney
" I would like to follow up and comment on some of the points made by the
Kahaner Report on the TRON project. Dr. Kahaner cited me (with some
justification) as a "lonely exception" to the lack of participation by
Western researchers, especially academics, in the TRON project. He also
observed that the project is controversial, and the follow-up comments
certainly illustrated this: They all suggested that TRON was
uninteresting and should not be taken seriously, although none of the
posters had significant first-hand knowledge about the project.
I will speak as one who does have some first-hand knowledge. Dr.
Sakamura first contacted me in 1985 because of my work on the IEEE "MOSI"
standard (an operating system interface standard for small computer
systems). I have participated in discussions about TRON since that time,
and I have been an active participant in the CTRON subproject of TRON
since 1988. For the record, I do receive research funds in connection
with this project. I was an invited speaker at two TRON Symposia in
Tokyo, and a TRON researcher spent a year working with me at West
Virginia University.
I am not an apologist for TRON or for the Japanese, but I am often amazed
by the *uninformed* negative reactions to this project. A discussion
which I initiated about TRON on USENET two years ago led to a wide range
of criticisms, many based on inaccurate knowledge (and a few on outright
anti-Japanese bias). I later summarized this discussion in the TRON
special issue of Microprocessors and Microsystems (October 1989). There
was no interest in establishing a TRON newsgroup to continue the
discussion.
I would like to propose a more balanced view. The TRON projects are not
a panacea, but with all respect to Professor Tanenbaum, it is foolish and
shortsighted to call TRON "dead as a doornail."
First of all, it is important to remember that the TRON "project" is
actually a large collection of subprojects motivated by a common vision.
That vision is one of open, global networking, supporting everything from
worldwide communication to local networks of "intelligent objects" in the
home. It is fair to be skeptical or opposed to parts of this vision;
Americans, especially, do not want to live in an environment where
computers seem to have the upper hand. The total TRON vision may never
come to pass, or may be far in the future. But the TRON subprojects do
not depend on the vision, and are not waiting for it. Some of them are
already technically complete, and are quietly finding their way into
commercial products.
The TRON goals depend fundamentally on open participation. TRON
subprojects are aimed at developing *standards*, not products. Many
commercial interests are participating, and each standard is intended to
enable products of many vendors, although reasonably differentiated, to
work together. Western companies with no present involvement in TRON may
find advantage in offering products compatible with these standards.
TRON is, of course, of Japanese origin; in the U.S. view it will forever
be "not-invented-here." There are obvious cultural and language barriers
to foreign participation. But participants from any country have always
been welcome, and specifications for the TRON subprojects, although still
under development, are being openly published. A few TRON presentations
and workshops have been held outside Japan, and the TRON Association is
willing to help organize such events wherever there is sufficient
interest. The text of Dr. Kahaner's report suggests that only a handful
of Western companies have joined the TRON Association, but a detailed
scan of the list he provides shows a lot of familiar names [thanks for
the correction--DKK]. These companies may not all be actively
participating in development, but they will not ignore potentially
significant markets.
The TRON standards are not developed in a vaccuum. They do not conflict
with existing international standards, and they interface to these
standards where appropriate (e.g. the OSI model, the Ada Language). TRON
representatives participate in international standards activities, and
the various TRON specifications are likely to be proposed for ISO/IEC
JTC-1 standardization when completed.
TRON is funded purely by an industrial consortium; it receives no
government support (is there a surer recipe for success? :-). TRON is
also not a trade barrier; nothing in its nature suggests that it could be
anything but a trade facilitator. In May 1989 the U.S. government
*proposed* TRON for possible inclusion on a list of sanctioned products.
There was a clear misunderstanding of the nature of the TRON project.
Part of the concern centered on the rumor that MITI would mandate use of
TRON-based products in schools, creating a supposed obstacle to U.S.
suppliers. This did not happen, although the U.S. government certainly
mandates widespread use of many American standards. This
misunderstanding was soon resolved, and TRON was never listed, but the
bad press continues.
The TRON project was conceived from the start to include five principal
subprojects. It is not correct to say that the project has "branched"
due to growth. It is also misleading to confuse the name TRON with a
particular subproject, or to form opinions or draw conclusions about the
TRON Project as a whole based on views about only one subproject.
Four of these subprojects have been well developed to date: ITRON, BTRON,
CTRON, and the TRON CPU (or CHIP). Each of these has already led to both
detailed specifications and products. ITRON, BTRON, and CTRON are each
families of operating system interface specifications. The TRON CPU is a
family of microprocessor architecture specifications. The fifth
subproject, MTRON (for Macro TRON), is aimed at developing an intelligent
distributed control for a complete network. It is in a much earlier
stage of development.
These specifications were designed to work together; the TRON CPU is
envisioned as the usual processor for ITRON-based embedded systems and
for BTRON-based workstations. However, they surely do not depend on one
another. In practice, most ITRON and BTRON products to date have used
other processors (Intel, Motorola, etc.) while TRON CPU systems often run
other types of OSs, including UNIX.
The TRON CPU has received the most criticism. I will not try to defend
this architecture, but even if it is not admired it will soon be found in
many Japanese products. Moreover, the OS specifications are being used
without the CPU. ITRON is the basis for embedded systems in applications
such as robotics, mobile communication, and consumer products -- not to
mention the TRON house, which does exist and apparently works. BTRON
workstations to date have been specialized for Japanese input, which may
limit their usefulness in the West. However, I have seen (in 1988) BTRON
systems that include effective multilingual processing, high-level data
management, multimedia output and *input*, and (a special concern of Dr.
Sakamura's) integrated support for disabled users. This could be
effective competition for some existing workstations.
CTRON is in a special class, designed for larger environments and
optimized especially for communications and information processing
applications. It is likely to find application in telephone and
communication systems, in Japan and elsewhere.
The TRON projects are not only feasible; they are developed and maturing.
Annual international conferences have been held since 1987, with
presentations in both Japanese and English and simultaneous translation.
Papers in the first conference focused on TRON concepts and development
of the specifications. In 1988 and 1989 increasing numbers of
implementation reports were presented. The 1990 conference was concerned
with topics such as performance, reliability, and validation. The CTRON
committee has begun a series of formal portability experiments involving
CTRON products of a number of companies, to validate the ease of porting
software in CTRON environments. This project was launched with a
symposium on Software Portability in September 1990. My paper in the
1990 TRON Symposium Proceedings, to which the Kahaner Report refers,
emphasizes portability because it was originally presented at the
portability symposium. A slightly revised version was then reprinted in
the later proceedings.
In summary, I strongly agree with the conclusion drawn by Dr. Kahaner in
his very objective report, that TRON is indeed a force to be reckoned
with. You may like or hate the project, but each TRON specification
deserves to be evaluated on its own technical merits. Many companies are
doing this, and some are adopting TRON elements. Like it or not, these
elements are already appearing in Japanese products and systems, and
understanding them will be important for international commerce.
Following these comments I am posting a copy of my TRON information
summary, which I try to keep reasonably up to date. This document
includes brief technical project summaries, an English bibliography, and
relevant names and addresses. I will be glad to assist anyone
interested in contacting the TRON Association or obtaining more
information about any aspect of the TRON Project."
Prof James Mooney Dept. of Stat. & Computer Science
Tel: (304) 293-3607 West Virginia University
Morgantown, WV 26506
INTERNET: jdm@a.cs.wvu.wvnet.edu
[End of Mooney's comments]
------------END OF REPORT-----------------------------------------------