Re:
> I'd like to see this come up myself. WT(f) is MCP?
MCP = Master Control Program
It started life on the Burroughs B 5000, and has grown and
changed over the years.
On the B 6500, in 1970 (when I first saw it), it supported
multi processors, multi-programming, hierarchical file systems,
virtual memory. It was written in ALGOL (well, ESPOL, which
was Burroughs ALGOL with a couple more extensions).
Also, circa 1970, on the B 7500, it supported fault tolerance,
years before Tandem "invented" it. I remember a demo at
U.S. Customs agency in San Diego, where someone yanked a processor
board out while the system was up ... the system kept running.
(I can't recall if the process that was executing died, but
I do remember that everything else stayed up & running).
The neatest thing about some models of the 7700 was that if you
had a dual CPU and at least two memory modules and at least
two disk drives, you could "split" a running system into two
completely separate systems, without losing a single process on
the "original" system. (The processes and OS would be running
on one of the CPUs, and the other CPU would boot up into an
idle state.) We used this at night, in 1978, to split off the
second CPU so we could have a crash & burn development system
for awhile. When we were done, we re-joined the systems (again,
preserving all the running processes on the "original" system).
The B5000 (circa) 1963 ran something called MCP, but I don't
know the relationship between it and the B 5500. The B 5500 was
a next generation machine, and ran MCP (written in ALGOL). The
B 5500 was followed by the B 5700 (don't know the differences between
the 5500 and the 5700).
A great paper on the Burroughs line is at:
http://members.nbci.com/amayer/papers/B5000.html
(hope the cut/paste didn't cause my mailer to go nuts again)
Stan Sieler sieler_at_allegro.com
www.allegro.com/sieler/wanted/index.html www.allegro.com/sieler
Received on Mon Jun 04 2001 - 13:42:17 BST