On 9/14/2016 2:45 PM, Califbill wrote:
wrote:
On Mon, 12 Sep 2016 11:02:07 -0500, Califbill
wrote:

I went to school for 36 weeks to
learn to fix mainframe computer systems for NCR.

Yikes, the longest mainframe school I ever went to was 8 weeks. They
sent you back to the field for a while, then you went back for a more
advanced course. Once you got a feel for the culture of the various
families, (Endicott, Kingston or Rochester) you usually did not need
much additional education to figure that stuff out anyway. If you were
trained on a 168, it wasn't hard to figure out what a 3033 or a 3090
was doing. The Rochester machines were even more so. If you understood
any AS/400, you understood them all. The hardware may have been
different but the maintenance package was the same and they all ran
the same software.
When I moved to Florida I waived training on about 400 boxes because
the technology of virtually all of the "industry systems" (ATMs, Cash
registers, teller terminals etc) was robbed from the UC.5 support
processor we had been using on Endicott mainframes for over a decade.
The rest was just belts, pulleys and wheels. Same with the 3890 check
sorter. It took me a few weeks to get a feel for the ink ****er and
some of the adjustments on the feed but it is just a paper pusher, run
by a 360/25 processor and I was a region specialist on the 25.



The 36 weeks was basic electronics for some weeks, and then the mainframe,
and all the peripherals. They we still discrete transistors back then, and
we learned to fix individual boards.

When I got into the game Burroughs was using parallel plate packages and
memory cores hand wired in Brazil. No monitors, just TTY with paper tape
or cards for input. Man, I'm old.