Xref: utzoo comp.realtime:114 comp.os.misc:963
Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!cs.utexas.edu!uunet!mcvax!unido!iraun1!Katmandu!grunwald
From: grunwald@Katmandu.ira.uka.de (Grunwald Betr. Tichy)
Newsgroups: comp.realtime,comp.os.misc
Subject: Re: TRON (a little long)
Summary: Maybe TRON is system for production plants
Keywords: Japan, TRON, standards, networks, operating systems
Message-ID: <926@iraun1.ira.uka.de>
Date: 6 Jul 89 12:51:02 GMT
References: <382@h.cs.wvu.wvnet.edu> <4628@ficc.uu.net>
Sender: news@iraun1.ira.uka.de
Reply-To: grunwald@Katmandu.UUCP (Grunwald Betr. Tichy)
Organization: University of Karlsruhe, West-Germany
Lines: 27

I think the Japanese have seen that the Automatation of the Production is
one major goal in the near future. There is a big need for realtime
systems, which can stand the conditions at production places. But
there is also need for a good program development platform.
Unix can only serve the second goal, because it is to big to build
costcompetitive systems with it. On the other hand there are very few real-
time systems with a good programming environment (the only exception I know
is OS9/680x0). 
Normally you develop under Unix and download the program, but this process is
errorprone and tedious, and you need an extra development system.
The next point is, that you have a hierarchy of computers in production
plants, from the machine motor control, up to the plant control system.
If TRON could cover all of it, it would be the system of choice, because you
have only one system, which simplifies maintenance very much.
If you think of machine motor control, the system should be realtime and
rombased, which seems impossible for UNIX. You could use embedded systems for
motor control, but at one point in the hierarchy the software cost will
raise to much and UNIX will demand a very exepensive hardware even at that
point.
To my opinion, you should forget about realtime UNIX, because UNIX can not
cover the whole terrain of Automation computing at a reasonable price.
(Using diskbased 68030 systems for the control of a single NC-Axis is not
reasonable)

The opinions expressed here in are totally mine. If I miss at some points
please correct me, but don't come with the hardware costs will diminish
sometime argument only.