Subject: Re: 8-bitters and multi-whatever
From: "Chuck Guzis" <cclist at sydex.com>
Date: Thu, 13 Sep 2007 15:59:59 -0700
To: "General Discussion: On-Topic and Off-Topic Posts" <cctalk at
classiccmp.org>
On 13 Sep 2007 at 17:36, Roy J. Tellason wrote:
And
I'd also disqualify device-sharing, such as a MAC (multiple
access controller) between an I/O device and two computers. Those go
WAY back...
I'm not familiar with such stuff.
Ah, showing my age. "Back in the day" you might have two otherwise
independent mainframes and, say, two printers. To have a system do
nothing but wait for a print job to finish while only one of the
printers was busy is a huge waste of resources. A card punch might
be an infrequenly used piece of equipment, so why have two? Or a
printer could be offline for maintenance, but why take the machine it
was attached to down also? Or, instead of having two banks of 8 tape
drives for two machines, why not whittle that down to, say 12, and
allow the drives to be shared? IIRC, most vendors offered some sort
of a MAC facility, even if it was a QSE. CDC certainly did.
Most MAC setups again were master/slave. An example was BOCES/LYRICs
PDP-10/TOPS-10 timeshare system that had a PDP-8I as the comms frontend
(switch). The PDP-8 served as an intelligent peripheral but it's
dectapes were not available to the 10.
Of course, direct coupling of computers was also done,
either via a
special I/O device or even shared bulk core. But we never called
that "networking".
Mostly because it was a clear master slave lashup. Generally/loosely
networking implied more than two machines and a more general ability
to transfer/communicate as needed be it files, shared devices or some
combination of both with any machine being able to initiate and
communicate as a peer to others that could do same or similar.
Allison