Re: What is "real-time"?

[mcsun!uknet!yorkohm!minster!mjl-b@uunet.uu.net]

In article <1992Oct15.153519.23717@fcom.cc.utah.edu> val@news.ccutah.edu (Val K
artchner) writes:
>I've given definitions of "real-time" from three different English language
>dictionaries.  Everyone seems to say that these are insufficient in a
>computer science context.  However, no one has actually posted an alternate
>third-party definition of "real-time".  I'll agree to use a formal definition
>as given by the FAQ of the comp.realtime group (sight unseen).
>
>: Consider the following example.  Suppose you are writing software for a
>: missile system.  Suppose you have a timing constraint where you must use
>: a backup sensor if the main sensor does not respond every 25 msec.  You
>: cannot express such a constraint in C without using an OS call or an
>: external routine.  The C language has no real-time facilities per se.

No-one seems to have spilt the real time issue into "hard real time" and
"soft real time". The definition of hard real time I was given (in "Real
Time Systems and Their Programming Languages", A. Burns and A. Wellings,
Addison-Wesley) can be paraphrased follows:

"A hard real time system is one where failure to respond within a given
time is as bad as the wrong response"

>Until we get a mutually agreeable third-party source (not biased toward
>Ada nor C/C++), properly cited for reference purposes (and fairness to the
>authors), then let's just "Stop the petty bickering!"*

The Burns and Wellings book writes off C fairly early on, because it has no
features for concurrency. Real-time and concurrency go hand in hand, and the
authors provide a fairly convincing argument.

Mat