🚀 go-pugleaf

I don't know if any Australians get this newsgroup. I hope they forgive
and correct my bad memory.

At the University of Sydney they were using big Cyber mainframes for
student jobs in the 77-79 timeframe. This was B.W. (Before
Workstations) remember. The department was physically separated from
the Computing Center by about 500 meters and jobs were submitted via a
RJE station. The way Unix got a foothold on Australian territory was as
a replacement for the CDC RJE software. Ian Johnstone and colleagues of
the UNSW had heard of this software called V6 which universities could
get for a pittance. What the heck, why not? After 6 man-months (I
think) of C hacking, they had a standalone Unix image that handled all
the RJE functions and was easier to maintain. The Sydney Univ. RJE
system was derived from the UNSW code. Our station was a 11/34 with two
line printers, a card reader and a teletype console. It read student
jobs from the hopper and generated the commands and delays required to
talk to the Cyber (no readahead in NOS).  Piers Dick-Lauder, our
resident guru, had dumped the OS image onto a pack of cards which was
used to boot the system from the card reader. Each card had the
starting address in memory of the contents so it didn't matter if you
spilled the deck!

This system served us well. Besides supporting Pascal teaching, there
was an assembler and simulator for a PDP-11 written in Fortran used for
teaching machine architecture. Imagine running a machine simulation in
a batch job.

In 1980 the department got a 11/780 and the RJE system was retired.
They did some amazing things with Unix V32 (BSD was at Release 1 or 2)
on that machine to support 60-80 users running ed and pi jobs.  Some of
that scheduling software was described in The Fair Share Scheduler (?)
in a recent CACM issue.  But that machine has a bunch of other
stories.