To explain this photo a bit more: back then, most programming was done on paper (sometimes on special coding sheets), then punched onto cards/tape, then fed into a mainframe to be tested (& if necessary debugged). Counterintuitively, needing LESS computer time was the mark of a better programmer!
Key Punch Operators were indispensable
I always created infinite loops when I took a Fortran class at Ball State in 1973. Attention to detail matters!needless to say, I did not pursue programming.
😅💯
Do not worry. You will get there eventually.
Because if you could get your code right on paper and not have to use any (or much) expensive time on a mainframe to further debug and test, you were way ahead of the game both in terms of the time it took you to produce shippable code and also in terms of the expense of shipping that code.
Shirley’s mostly women programmers at Freelance Programmers were extremely good at this (I’m not just saying this, they kept stats on themselves & compared it to their competitors) and as a result were able to produce software for major companies & governmental agencies both on time & for less cost.
Not that many girls (Chicago public high schools) were interested (in mid-Sixties).
That may be one reason that you don’t see a ton of images of them with computers—ironically better programmers back then didn’t need to hang out in machine rooms as much!😅 bsky.app/profile/hist...
What's the baby doing now?
Legend has it she is still distracting programmers to this day… (Just kidding, I don’t actually know what her daughter does.)
I’m gonna link to info on both Moffatt’s memoir (which is really personal, interesting, and at times horrifying) and Shirley’s (which is more polished but also quite personal): www.fitt.org.au/the-it-girl-... www.steveshirley.com/books/
Thank you! I love reading about the pioneers in my field of work.
When I started in the 70s It could take an hour or more to compile and link your program before testing it. We worked through the design and logic thoroughly on paper first as well as closely proofreading for typos before submitting a job. Getting it working in one submission was the goal.
Although the IDE has replaced the paper, programmers in the 2020s would do well to adopt these practices and goal.
In my later career (just retired!) I was primarily in integration and test and I was amazed that testing was no longer considered part of the developer’s job. If it passed the little unit test it went on to me. Fair division of work but I would have been shamed to let anyone else debug my code.
My Dad used to write mainframe OS software, and in the early days had to rent time on other people’s machines to test it which meant working nights. (About 15 years ago he worked out the HDD he’d just picked up for £80 would have cost something like £3m back then… per month)
In my 1977 computer science class, we did our assignments on a time-sharing system with terminals. We had to go to the computer lab and type in and run our assignments. My standard practice was to write out and check my code on paper, then go to the lab and try it. \1
Many of my classmates walked in with nothing and starting composing at the terminal. I was usually finished and out of the lab long before those classmates. \2
This is very great words.hope you are doing fine
Wow! That is next-level code-jedi cool. It reminds me of being in middle school and writing BASIC programs in the back of my Louisiana history composition book so that I could make the best use of my time in the Apple II computer lab next period. Of course, said BASIC programs seldom worked.
The compute limitations and capacity of yesteryear meant that wasteful and negligent coding (as well as practices) was weeded out due to the cost of computations. Doing things properly the first time around, or at least attempting to do so, was the standard. It's jarring to know in the age of gAI.🙈
🫠🫠🫠
Spell-check causes people with good spelling skills to commit more spelling errors. Statistical software increases the misapplication of statistics. Quantity over quality errors?
Part and parcel of “the tyranny of numbers”?
I used punch cards in college I wrote a.chat bot in 6th grade
Interesting
This is how my first Computer Studies teaching went. Took a week to get the results back! “Error at line 1”…and repeat
The fortran compiler on the IBM 360 was even better. Put in a 1000 line fortran program and get back "syntax error" with no indication of where the error was.
lmaohno
My class was a Vth form secondary. Task: input two numbers, add them and output the result. Could take the entire six week course at one period of access per week via acoustic modem (which was also problematic)
Wow!
I did know of one school in the 1980s which taught a GCSE course on ICT whilst having no computers for the whole of the first two terms of a two-year course, and made do with cardboard keyboards!
Aha. That's why my FORTRAN class used WATFOR (FORTRAN compiler from the University of Waterloo) instead of the IBM compiler. I always wondered.
My dad got a job by being asked to "code" something, writing down the code on paper, and it running without any eras when they plugged it in
*errors argh
Classic.
Cool. Here's an assembler coding sheet. I still have a pad of COBOL sheets somewhere in the basement.
COBOL will be around for a long time to come
I still have somewhere a pair of FORTRAN coding sheets
That’s so cool! Thank you for sharing a picture!
The transition to interactive coding is neatly illustrating by this teletype printout on Wikipedia. en.wikipedia.org/wiki/Teletyp...
We used sheets similar to this when I took computer classes in the 70s
Here's an 80-column card. This was used for material transfer between warehouses, but it could also be used for other purposes, such as coding COBOL programs. I use it as a bookmark :)
Nice! I remember the RPG II pages coming in a giant variety pack. I don't miss those days. Nor teletypes.
I just learned assembler and this is really cool! I had no idea about those sheets.
Yep. I started programming in NEAT/3 on legal pads, which I then keypunched into cards and fed into an NCR mainframe. Very little of the time spent actually involved being at the computer.
I remember writing code on paper, and sending it to the keypunch team. Clear hand-printing was important. (State technologies often lagged behind the times.)
I remember that time as "the bad old days". Some comp sci professors would grade projects based on how many runs we needed to debug it. 3 or less was an A, 4 runs = B, 5 runs = C, etc. A far cry from today's world of cloud servers, VMs, and containerized apps!
The whole domain has changed. What used to be the study of what happens on a single computer has now become trying to master a small part of a large distributed system.
That baby is adorable tho
When I was first learning to code, we wrote our COBOL on coding sheets which were collected by courier at the end of the day. Sheets delivered many miles away to computer centre where they were punched onto 80 column cards & fed into a mainframe. Cont…
Telex ;) Fax machines ;) Carbon copy 8 page toll documents! IBM electric typewriters with memory. Those were the times.
Compile & test results returned to us by courier next morning. I never saw that mainframe. End.
The hand punch was your friend :)
Yep, if only a few corrections needed, we hand punched replacement cards.
PS, I remembered 45 years later: we were in Southampton, mainframe was in Windsor, 60 miles away.
We were on-site with the IBM mainframe, but you still didn’t know when your job would run because it was in a queue.
That’s how it was when I was an operator. Except the programmers we liked tended to get bumped to the front of the queue.
UCC7?
Nope. No idea what that is 😆
It was a mainframe job scheduler.
So you never learned to operate a keypunch? Once upon a time, that was a skill I was absurdly proud of.
Yes, I used a keypunch when I was an operator, and later as a programmer to punch corrections.
Tim Berners-Lee’s mum taught us BASIC programming at school. The results came back from a local Honeywell office on paper tape. Unexpectedly, I became a COBOL programmer after doing an Economics degree.
Yes, that’s how I wrote COBOL in 1980. Probably still have a pad of coding sheets somewhere. Definitely have some print with sprocket holes and a 360 reference card.
Can confirm how that was how you used to code - in High School did punch cards.. I'm not that old...
One thing worth noting about these days: Cards were actually far more convenient to use for editing programs because you could do things like move blocks of code around by physically moving cards, and insert characters by "duplicating" the card in a cardpunch while holding the source card down.
Love learning this part of history. One time a moth got stuck in the paper and string of mainframe; that's why computers have "bugs"
This was still true when I was first learning to code in the 1970s. You were supposed to debug your code on paper before even submitting it on punched cards. I'm now flashing back on the distinctive sounds of an IBM keypunch machine.
In 1963 I was coding in 8k memory. Yes I remember writing code, using punch cards or tape for input. Voyager I is still out there running with 68k of memory.
Back in the 80s I had a colleague who had done a PhD in Queensland when there was only one available computer for academic use in Australia. 2,000 km away. She sent boxes of punch cards to Melbourne, it was a two week turnaround for each run. She was VERY careful on what she sent off.
Im Sara from Gaza😭🍉🇵🇸 My baby need pampers and milk Can help my baby by link in bio please 🙏🏻
My dad trained programmers at IBM—this was way before any such thing as a CS degree—and he said when they switched from coding templates and once-per-day assembly runs to writing code in a terminal environment, his students consistently failed to complete their assignments.
My experience was the opposite. When my stack of cards had one trivial typo in it so failed to run, and I had to wait 1 or more days for a rerun, it was not helpful. In a terminal environment, I fixed it quickly and moved on, launching a decades-long software career.
I can empathise with a collection of programming students, used to a particular method of working, finding adjusting to a new method of working difficult. I found the transition from compiled C++ to interpreted python tricky and only really “got it” once I found new techniques to help me (eg TDD)
He was extremely critical of what he called “trial-and-error programming.”
But, per my other comment, being able to fix things and rerun quickly isn't necessarily "trial and error programming". Even when it was, we LEARNED from those errors, didn't repeat them, and progressed to writing solid code. Getting to that point far faster than with the ancient turnaround times.
Hey I’m just relating the observations of a cranky geezer. One certainly couldn’t operate under that kind of structure in this day and age. Those guys had like 16kbits for the whole operating system.
Totally agree. And no disrespect to my predecessors - they were giants.
I coded my data set (the 1849 Paris Chamber of Commerce report) for my MA thesis on Parisian women's labor. Those green sheets! & the errors I made typing those cards!