Log in

No account? Create an account
Previous Entry Share Next Entry

Card deckery in the '60s

Back in the bad old days the only method of inputting anything new into a computer was via punched cards. The IBM 026 Key Punch machine was the way to do it. They rented for $50 a month (this was before the IBM antitrust suit whose result was that IBM had to sell as well as rent equipment). The machine was actually slightly programmable - there was a drum with a specially prepared punched card on it that could specify things such as that certain columns had numbers only or have the equivalent of tabs. Most programming languages had column specific meanings such as column 7 being the start of an actual statement in Fortran and columns 72-80 were for sequence numbers.

Tired of waiting in line to make corrections to my card decks I tried to convince management to get another key punch for programmers to use. My economic argument about the $50/month rental of a key punch machine vs the $600-$1200 a month salary of a programmer which showed that if there was at least one person waiting line for a spare machine to make corrections for as little as a cumulative 45 minutes a day it was worth getting another machine for programmer corrections still wasn't good enough because of various bureaucratic reasons; eventually they gave in and for a while our project got a temp worker and a key punch machine.

The Key punch machine printed on the top of the card what was punched so they could be read. Woe to the person who dropped a deck of cards that didn't have sequence numbers. Card decks could be sorted on a sorting machine and printed on an IBM 407 Accounting Machine. The Accounting machine was programmable by plugging in wires on the control panel board and was used in businesses for considerably more complex operations than printing a deck of cards. It could print at 150 cards a minute; for assembly language programs which could be a whole tray of cards this was slow.

  • 1
I am proud to say that I used punch cards in 1978 to RJE GPSS to UMC. By the time I was an undergrad there the cards were all gone.

I still have a nice little stack of blank cards!!

I'd been wondering if there were still some around!

Actually, IBM still kept them in the supply closets as late as 1994 when I left the company - they might still. They make great note cards that fit perfectly into that inside pocket of a suit jacket.

I left with a stack and got regular infusions for years from former co-workers. These days I do not use with abandon because my source has dried up but they are still quite handy for lists.

Somewhere I have a deck of a Fortran english-to-piglatin translator I wrote. One of these days I'll type it in and get GNU Fortran to compile it.

As I might have mentioned, I ran into card sequence numbers in the year 2001, in (what else?) the source code of a Fortran program. Some of the lines (as they had become, though the program itself referred to "cards" in some of its output messages) had numbers, and some did not.

I remember a slight later version of this; I got to play with them while my mom was taking computer classes (the lab was empty enough at night when she was typing her cards for it not to be a nuisance if I sat around punching cards too).

I was always fascinated by the machine but *hated* the process.

What I remember most about those machines are the very gratifying sounds they made while being operated.

  • 1