IBM Data Processing Equipment (was Christie's Auction...)

From: mhstein_at_canada.com <(mhstein_at_canada.com)>
Date: Sun Feb 20 22:48:22 2005

Original Message:
From: Scott Stevens <chenmel_at_earthlink.net>
Subject: Re: Christie's auction and other computer history events

<snippage>
Well, technically, all the IBM equipment from those years was 'data
processing' equipment, not computers. The data was processed (as most
here probably know) with card sorters. The database consisted of the
cards, shuffled around in various ways to produce sorted decks. The
fields to sort on were determined with wire jumpers in plugboards.
There wasn't a computer in sight in most of those systems.

It highlights the difference between 'computer science/software
engineering' and 'IT'. A difference not so often discerned these days,
though it gets me riled up when a recruiter or HR person when I am
looking for work trys to shoehorn me into IT because I'm experienced in
embedded controller hardware/firmware development.
---------------------------------
Well, technically, the sorting was not really part of the "processing,"
more like preparation for processing, nor did sorters have plugboards;
columns to be sorted were selected with a manual dial.

The plugboards (aka "programming panels") were effectively programs,
telling the 402 "Accounting Machine," for example, which fields in
which cards to add, which fields to print and where, etc. Instead of
typing "Run Invoicing," (since there were no keyboards anywhere:) you
went to the cabinet holding the panels and "loaded" the panel into
the side of the machine(s).

A roomful of various machines would be *functionally* equivalent to a
small computer; a typical invoice run, for example, would take keypunched
and sorted input cards representing each item sold to each customer,
merge those cards with customer master cards and inventory cards,
print the invoices and prepare new customer and inventory cards with
updated balances for the next run. One machine would sort, another
would merge, another would calculate extensions, another would add
and print and finally one more would punch the new master cards.
Instead of file names, different colour cards were used for the
various card files, and moving the cards from one machine to the other
was done by the sneakernet method.

The programming (done with jumper wires) would go something like,
"If this is the first card of a new customer master, print the total
on the invoice, punch a new customer account total card, skip to a new
page, print the customer name and read the next card;
else ..."

Incidentally, for those who appreciate blinkenlights, a model 604
"Calcalulator" (sic - a typo from IBM's manual, actually an "Electronic
Calculating Punch") from the late 40's had over 300 of them and would do
a fine job of heating a small bungalow. Except for the electronic
calculating machines, the rest of the hardware was entirely relay-based.

Ahhh, the good old days; many happy hours spent in front of an 082 sorter...

mike
Received on Sun Feb 20 2005 - 22:48:22 GMT