Thread by @Foone: "OK SO let's say it's 1962 and you're lucky enough to be a programmer working somewhere that has an IBM 7090."
At secondary school I chose Computer Studies as one of my subjects, and we used punched cards which were sent to the local university (I think) for processing. The method was to mark the cards with a 6B pencil, reading the fine print to understand the coding of letters digits and punctuation. The cards were returned punched, and I soon learnt that it was a mistake to correct just one card and resubmit, because a repunched card could be damaged and fail - it was better to fill in all the cards of a fresh deck even for a single change. I think the turnaround was about a week, but I suppose it might have been a bit less - we probably had two lessons a week in that subject.
Some many years later at my first job, I rather surprised my boss by speaking of a SPICE deck - SPICE was a circuit simulator, and it seemed natural enough to refer to the input as a deck, even though in our case at that time it would have been a text file on a VAX.
In college, in the Engineering Breezeway lab, we had and “RJE” system. RJE stood for Remote Job Entry.
Essentially, this was a terminal, a card read, and a line (band) printer.
This allowed students and faculty to directly run their decks.
The printer was used both for routine printing from the timeshare system as well as folks running their decks, so there was no guaranteed correlation that when the printer lit up after a deck was run that it was, indeed, the job from the deck just submitted.
But it was the fastest on demand printer available.
Drawing a diagonal line across the deck (as in the photo) was a simply way to see if any particular card is drastically out of order, but if you transposed cards, it didn’t help much.
It helped dramatically if you were able to add line numbers to your cards in case the cards got dropped. I have dropped my share of decks (and not only mine). Hopefully they fall and just fan out, which is quite different from “52 card pickup”. But in the end, you likely needed to correlate your deck to a recent printout.
The largest decks routinely seen in that lab were no larger than, say, 8" thick. But I had seen some decks which were several trays, several feet long. This was likely mostly raw data, in contrast to source code.
Cards were transitioning out at school when I got there. There were still some legacy courses that still used them instead of the time share option. Many of them were engineering or science courses leveraging software, not so much for learning computer programming. (SPICE decks for example, SPSS data, etc.)
But, appreciate, today sure, this is a glacially slow process. But back then? Honestly? No, not really. EVERYTHING was slow back then. If you’re were bulk entering data, typing in to a card is just as fast as typing in to an editor. EDITING was bothersome as you had to replace the entire line (card), but even that isn’t untoward. There are folks that bind their DELETE keys to do a DELETE WORD, as for many it’s simply faster to retype a word than fix a mistake.
There’s a nice bit of process reviewing the deck, thumbing through it, just like we do now re-reading code on the screen. You still have to do that, it’s just done in your lap instead of on a monitor.
The biggest issue was simply the waste of it. Not the cards, but the printer paper. It’s hard to imagine how many reams and reams and reams of greenbar the computer center burned through each day. We were never lacking for moving boxes, I’ll tell you that.
But for the time, it was a viable and productive mechanism for using the computer. As with everything else we find horrible today, a lot of valuable stuff was done using these techniques.
I don’t miss the era, but I don’t think it should be looked at necessary as walking up hill in the snow both ways either.
Something I have become fascinated with lately: single card binary maintenance programs. E.g., a single card which would prompt for an address range and then dump that portion of the memory nicely formatted onto a printer. There’s actually ample space in 80 words (up to 12 bits) to accomplish something useful – and, if it comes on a card, it’s certainly much more impressive than a one-liner on a monitor.