Punched Cards | Hacker News Reader

Punched Cards

computerhistory.org

44 pointsli4ick22 comments

I am old (I have seen things). I still have my first program from college, in FORTRAN, on a card deck. I also worked as a "computer operator" (an extinct species) while going to college, putting card decks in the readers, running decks that were output from the punches through "interpreters" (which printed the characters across the top of cards), and remembering that you always (ALWAYS) drew a diagonal line across the top of a deck of cards with a pen, in case you dropped them (colloquially known as a "floor sort"). If you wanted to be really fancy, you'd use different colored cards for different sections of a program, but that was rarely worth the effort. Good times.

Besides card punches, readers and interpreters, other quaint machines I dealt with on a daily basis were chain printers (with carriage tapes - look it up), reel-to-reel tapes (which required cleaning the heads with isopropyl alcohol and cotton swabs once a shift), and the most evil of all, decollators (again, look it up). All controlled via consoles that didn't have cursor keys, so to this day I have the TERRIBLE habit of backspacing to nuke and fix typing mistakes rather than cursoring and surgically correcting them. I bet the Backspace key on my keyboard is probably in the top 10 keys in my usage profile. :)

I'm old, too. I learned to program on punched cards, and remember the move to high-speed paper tape. That was awesome!

One of my favorite stories is from those punched card days. I was a careless teen, and knocked one of the developer's card decks off of a desk, sending all the cards scattered widely across the floor.

The sysadmin made me gather them up and manually sort the several thousand cards into the proper order. It took a over a day. When I was done, he took them from me and dropped them into the hopper of a machine in the corner to "check my accuracy".

What he didn't say, and I didn't know until weeks later, was that the machine he used was a card sorter and could have put the entire deck into the correct order automatically. The dev made me sort the cards manually as punishment -- and it worked, as I was much more cautious from then on.

As someone who first started using computers in the 90s, I lacked the tactile experience with cards that the generation before mine had. I long suspected that I had missed something by not having used punched cards, little realizing just how much I had missed.

What was shocking to me about punched cards is the indelible mark that they left on modern day computing. Why do terminals default to 80 columns? That’s the width of a punched card. Why do tabs have variable length? The key used to move to the next “field” on a punched card, by way of a physical tab stop. Why did FORTRAN reserve the first few columns to line numbers? So you could put your FORTRAN deck in a card sorter if you dropped it.

The link above seems to mirror the first part of the exhibit at the Computer History Museum in Mountain View, which is well worth seeing in person.

Even better are the live demonstrations of the IBM 1401 that are also done at the Computer History Museum on Wednesday‘s and Saturday’s: http://ibm-1401.info/

Why COBOL has that 80 character limit. But for far reaching legacy hangovers, you can't beat the QWERTY keyboard when it comes to tech. A layout born to slow down typists in a way to prevent the hammers jamming upon mechanical keyboards of old.

But the legacy of punched cards is deeper with that 80 character limit and bank card data formats and many other things had that character limit consideration come into play.

So if you ever see a data field that seems to arbitrary have around 70-80 characters(rest used for transaction ID, EOR char, checksum char(s) to pad the 70+ to the full 80 characters), it may well be down to punched cards if you go down the historical rabbit holes.

  A layout born to slow down typists in a way to prevent the hammers jamming...

Myth.

https://www.smithsonianmag.com/arts-culture/fact-of-fiction-...

That article, and the source it references [1], are no more convincing than the sources that support this "myth".

[1] bases it's conclusion primary on the assertions and opinions of the authors.

> At least, in our opinion, T, the most frequently-used consonant, should have moved to the center of keyboard, W, the other semivowel, to the upper row, Q, less frequently-used letter, to the edge, and I, also used for thenumeral 1, near by 8 to type “1870” or “1871” rapidly ...

After reading that article, I don't know why I should trust the opinion of those authors over the findings of every other author (which support the "myth").

It's pretty clear that the most common digrams are well separated in the typebar basket, with the exception of e+r, which is because 'r' and '.' were swapped by the manufacturers, not the designers.

[1]: http://kanji.zinbun.kyoto-u.ac.jp/~yasuoka/publications/PreQ...

  every other author (which support the "myth")

Which are?

Please see the list of references at the end of the paper.

This is not quite accurate. The first few columns were reserved for an optional numeric statement label that could then be used as a target for an IF or GOTO statement (and we used lots of those!) Labels did not have to be in numerically increasing order.

On the other hand, the last few columns of the card were ignored by the compiler, and could indeed be used to hold numerical values. These could be used to sort the deck of cards if you dropped it. The choice of numerical values in this column was not trivial because of the way programs were edited to fix bugs: you manually removed some cards and inserted new ones into the deck. You never repunched the whole program.

And normally in Fortran the statement numbers used different ranges.

From memory we used to use a common standard where

6000 series for GOTO etc

7000 for input FORMAT

8000 for output FORMATS

But isn't the idea the same? If you have explicit line numbers then you can just load the cards in any order. No need to sort it.

Another important market for IBM punch card technology during the 40's https://www.theatlantic.com/magazine/archive/2001/04/hitlers...

Mentioned on page 9 of the article.

The layout for this page is so interesting and useful! Each section divided into little bookmarks, with each bookmark having its own bar denoting how much content is held in it. Never seen anything like it before, and it works great for this format.

I came to say the same thing. Very interesting design indeed.

The titles appear onmousover for the "bookmarks" -- although I don't think the are bookmarks but stacks of punchcards.

Didn't the Lisp Machine have a graphical interactive punched card simulator?

  /usr/games/bcd

https://www.masswerk.at/keypunch/

Hmm, no mention of edge-notched punched cards - https://en.wikipedia.org/wiki/Edge-notched_card

Punch cards were also quite durable, you could dump them in a punch in 'dup' mode to make a backup, and usually they had printed across their tops the text on them so it made reading them pretty simple.

Oh wow what a huge amount of content. This is fantastic. Really nailing the goal of being an online museum.

The joy of reading your card deck, having it compile and produce the correct output, the very first time!

Reply on news.ycombinator.com