Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!uunet!munnari.oz.au!brolga!uqcspe!batserver.cs.uq.oz.au!brendan
From: brendan@batserver.cs.uq.oz.au (Brendan Mahony)
Newsgroups: comp.realtime
Subject: Unbounded loops and recursion (Re: Software primitives)
Message-ID: <5005@uqcspe.cs.uq.oz.au>
Date: 26 Sep 90 00:48:21 GMT
Sender: news@uqcspe.cs.uq.oz.au
Reply-To: brendan@batserver.cs.uq.oz.au
Distribution: comp
Lines: 24

I am a little confused about the exect meaning of unboundedness in the
discussion of recursion. Do you mean absolute bounds are required rather
than function bounds? Non-determinism was mentioned, so you might be
complaining about truly unbounded execution times. 

In any case how can a compiler be expected to distinguish the bounded
loop from the unbounded? Do you wish to remove loops entirely?

I'd say many real-time (embedded) systems have a fervent hope that they will not
terminate until the holocaust of the FINAL POWER DOWN. So logically-infinite
loops are needed. You know things of the sort

	while true do
		keep that reactor running

I don't really see why loops are any more dangerous in RT. You need to
do more thorough analysis of their behaviour, but that is true for all
RT programming. Non-terminating calculations are not appreciated in any
program.
--
Brendan Mahony                   | brendan@batserver.cs.uq.oz       
Department of Computer Science   | heretic: someone who disgrees with you
University of Queensland         | about something neither of you knows
Australia                        | anything about.