Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!mailrus!iuvax!cica!ctrsol!ginosko!uunet!mcvax!ukc!harrier.ukc.ac.uk!rjf
From: rjf@ukc.ac.uk (Robin Faichney)
Newsgroups: comp.cog-eng
Subject: Re: Definition of Integration
Message-ID: <1906@harrier.ukc.ac.uk>
Date: 26 Jul 89 11:43:30 GMT
References: <54787@aerospace.AERO.ORG>
Reply-To: rjf@ukc.ac.uk (Robin Faichney)
Organization: Computing Lab, University of Kent at Canterbury, UK.
Lines: 44

In article <54787@aerospace.AERO.ORG> abbott@aerospace.aero.org (Russell J. Abbott) writes:
<Does anyone have a good definition for "integration" in systems?  What
<I'm looking for is a specification that requires system components to
<work together.  In an integrated system, for example:
<
<(a) If a component C2 takes as input the logical (from the user's perspective) 
<output of another component C1, then it should be possible to compose the two 
<components as in C2(C1(X)).
<
<(b) Similarly, in a mouse-based environment, given the input/output
<relationship between components C1 and C2 described in (a) but where
<input and output are provided interactively by the user it should be
<possible to point to C1's output as a way of specifying C2's input.
<
<What I'm looking for is a formal/rigorous way of saying this--and any
<other useful requirements about integration.
<
<-- Russ

This is not exactly what you're looking for, but:

I've been working on the practicalities, rather than the formalities, of
this sort of thing.  In particular, the redirection of graphical user-
interaction.  There is a paper which is going to be given at the European
Unix systems Users Group (EUUG) conference in Vienna in September.  The
proceedings are published by the EUUG.  There is also software (crs), but
unfortunately that probably will not be ready for distribution when the
project contract ends in October.  On the other hand, some related
software (dp), basically a high level programmers interface to existing
IPC, with some applications including a graphics-interfaced conferencing
utility, is available now. Dp requires BSD IPC, and the applications need
Sunview or X11.  If you are really keen, we might be able to give you an
alpha version of crs but you'd be strictly on your own (though its not
hard to handle).

If you're interested in any of this, reply by email or phone me (Robin
Faichney) on +44 227 764000 ext7681, or write me at

Computing Lab
The University
Canterbury
Kent
CT2 7NF
UK