Re: Systems Integration -- a query [ please cc me on any replies .. djf]

[David Farber <farber@central.cis.upenn.edu>]

Date: Mon, 1 Nov 93 08:23:35 -0800
From: touch@ISI.EDU
To: Brian.Randell@newcastle.ac.uk
Cc: touch@ISI.EDU, farber@central.cis.upenn.edu,
        interesting-people-send@eff.org,
        Manny Farber <farber@systech.tik.ethz.ch>,
        f-troup@aurora.cis.upenn.edu


I'll take a stab at this... (being Halloween, it seems appropriate to generate
'straw men' today :-)


> I'm pursuing the idea of taking the topic "Systems Integration" as (perhaps
> part of) the focus for our 27th Annual Newcastle International Seminar on
> the Teaching of Computing Science at University Level.


So far as I've seen, 'systems integration' means different things,
depending on
who's using it:


To consulting companies, SI means 'selecting the hardware and software
components of a system so that they perform well as a synergistic whole'.
I.e.,
don't pick a high performance CPU and a disk with hugh seek times for a
disk-intensive system.


To computer manufacturers (e.g., IBM, Apple, etc), it means 'selling
"turn-key"
systems composed of only their company's components, such that removing any
single component defeats the configuration. This includes the use of
'proprietary' cabling and connectors, and very odd component characteristics
(odd size, shape, power requirements, etc). Basically, it's the same as the
above, constrained within the resources of a company. The main problem is that
the pieces may work well together (i.e., hiding each others' flaws), but that
component replacement or upgrade isn't possible.


Oh well - just a first stab.


Joe