Discussion of large systems

[M H Stein]

------------Original Message:
From: "Chuck Guzis" <cclist at sydex.com>
Subject: Re: Discussion of large systems

>On 2 Nov 2006 at 10:25, M H Stein wrote:

> And considering the cost of CPU time and the fact that input data was
> still on punched cards in most cases, it was usually more efficient and
> cost-effective to still sort the cards off-line and then copy/merge to/with
> tape.

I suppose that relibility was about the same--the chance of creating 
a mangled card in unit-record gear was probably about the same as 
that of mangling it in a 1442.  

------------Reply:

Reminds me of my very first night working in a unit record shop:
Was running an invoicing job on a 402 with a 514 punching the new
summary cards; loaded the feed hoppers with cards and went upstairs 
for a smoke and a chat with my new boss. Came back down to reload
the hoppers about a half hour later, but there were only two or three
cards punched in the 514 instead of the 200 or so expected, and one
card sticking out from under the machine cover. Opened the cover
and the other 197 or so cards literally exploded out of the machine, torn,
accordioned and with nice black stripes where they'd passed through
the drive gears into the machine's innards instead of into the output stacker.

Figured it was going to be a good place to work when the boss joined
me in the keypunch room and helped with finding and putting the torn
fragments back together in order to manually repunch them. Got pretty
good at reading cards that night...

m