A question about punchcards

Factual Questions
1

Inspired bythis thread.

So, I know punchcards were once used to program computers, right? Before there were keyboards and intuitive interfaces and whatnot (I’m so not a computer person, so stop laughing 'till you get to my real question!)

What did y’all do with all the chads from the punchcards? Were the punchcards punched over a collection bin? Is that where all that confetti for parades used to come from? Was it all thrown out? Recyled as firelighters? Inquiring (young) minds want to know.

And, as an adjunctive question, did you have to punch all the little holes out by hand? With a little stylus or something? Were dimpled chads a problem back then?

Going to hide in a corner from the snickers and derision about to ensue.

2

Punch cards were mostly used for storage, the actual input and output was usually accomplished via a typewriter like device and roll paper. Most punchcards were punched by the computer itself, but you could do them manually if you wanted. Most machines had a bucket to collect all the bits.

3

Is there more info on punchcards? How big were they? How small were the holes? How many bytes of info could you store on a punchcard?

4

When I started programming I wrote out my program on sheets with boxes marked on them. I sent them away and back came a stack of 80 column cards that I then had to check for transcription errors.

5

Everything you’ve ever wanted to know about punch cards. The photos on that site should bring back fond memories for some of you!

6

I used to use an IBM 029 to punch cards for my programs, back in the dinosaur era.

See http://www.columbia.edu/acis/history/029.html

The chads fell into an internal box. You wouldn’t want to use them for confetti. They had sharp corners and were difficult to clean up if spilled.

7

Believe it or not, there were actually machines that read and created punch cards. At the time there was also electric lights, radios and automobiles. The holes were not punched with duck quills.

The holes were punched and the little squares were collected in a bin. Those with access to the computer center on campus took a bag of the stuff to the football game and used it as confetti. The fans that had no idea where it came from wondered how they managed to print a number on each of the little confetti squares.

Punch cards were crude by today’s standards but they got the job done and were high-tech at the time, much like the Apollo space capsule.

8

WhyNot:

Here’s a good photo of the standard late 1970s keypunch I started out on. http://ed-thelen.org/comp-hist/ibm-029-keypunch.jpg (The rest of that site has a lot of interesting info on older equipment as well.)

You can see the fresh unpunched cards at the upper right of the machine. When you pressed a key similar to the [Enter] key on a PC, a card was pulled down from the upper right hopper and its left edge slid under the punch heads. The punch heads are the left edge of the rightmost of the 3 card-holding stations in a horizontal row. (ie just above the area that’s glared out in the photo)

As you typed, the hole pattern for each character was punched into a column on the card and the card slid one column to the left. Each card held 80 columns of information. When you got to the end of the card, it was ejected to the left and pulled upwards into the stacker on the far left of the machine. You can also see a few cards sticking out of the staker edge on in the photo. Meanwhile, a new card would feed from the upper right hopper and you’d keep on typing.

The machine had a bunch of crude programming or macroing features, controlled by, you guessed it, a punch card. That control card was wrapped around a drum visible inside the window at the top center of the machine.

The machine and its built in desk weighed about 200 lbs.

The chads dropped into a hopper down by your knees. They were very sharp-edged and folks were told not to use them as confetti because the little buggers were sharp enough to cut people. So in answer to your questions, they were thrown away as ordinary trash.

Voting machine punch cards, their punching machines and counting machines are a totally different animal.

On preview I see I’m too slow for this bunch, but what the heck, I spen the time to type this so you can spend the time to read it if you care …

p.s.

Friedo: Depends on the system. When I was doing mainframe stuff in the 1970s, 100% of programmer input and user data entry was via punch cards. We had no terminals yet. Yes, teletypes and their descendants, as well as proto-3270s (I forget the modle number) existed, but not everywhere had them yet.
Dog80: Each punch card held exactly 80 bytes. In the case of program source code, by convention the first 8 or 10 columns and the last 8 columns were reserved for other uses, so each line of code was at most 62-64 columns.

Likewise, data files often used the first few columns to indicate what type of file and record they were, and the last few columns were often a sequence number, so the effective data width per card was also 60-70 bytes.

Sequence numbers were essential since the decks of cards were handed by people which means they were occasionally dropped. Imagine taking a large program shuffling the statements at random. Not good.

Having sequence numbers meant that if a deck was dropped, you could pick it up and run it through a mechanical sorter to get it back into correct seqeunce and then it was usable again.

In those days most data was kept in some particular sorted order and the using programs depended on that order. Random access to records wasn’t feasible with punch cards. So a dropped / shuffled data deck was also useless. So they too had to have sequence numbers so the “database” could be “reindexed” if somebody dropped it on the floor.

Cards came in boxes about 2 feet long which held 2000 cards each. A good-sized COBOL or assembler program could be several boxes long.

As narrrow as cards were, a data file may have consisted of several cards per logical record times several 10s of thousands of records might be 100,000 punch cards. Pretty quickly everyone was switching over to tape or disk for all but input and output, since handling temp files the size of Volkswagens was impractical.
Don’t ask: Aaahh, yes. Desk checking. Always my favorite way to spend an afternoon.

9

This brings back memories: How MUCH is a prime-condition, IBM 360 computer worth now? A junk dealer north of Boston has one…you can ahve it for free!

10

There were also “binary” cards that stored more than one byte per column. These were created by overpunching the card. They were used for special applications like bootstraps and loaders. Yes, you could boot a computer from a card reader.

11

You might want to grab it, melt it down, and recover the gold from the contacts. These days, gold-plating techniques are so good, the gold on circuit board contacts is so thin, it’s not worth recovering. But those old mainframes had a fair amount in them.

12

Speaking just to IBM-style cards and 360/370 machines:

Each card column had 12 rows and hence 12 bits. I know there were cards with binary data as well as the more common EBCDIC, but I’m pretty sure they still only put one byte per column. Clearly it’d be possible to store 1-1/2 bytes per column, but that’d take a lot more circuitry to decode, and hardware was dear in those days.

The format of an object file (ie output of a compiler and input to the link editor / loader) had, IIRC, the characters “OBJ” in the first 3 colmns, and binary for the remaining 77 columns. There may have been a few columns of sequence number ahead of the “OBJ”

I agree the format of a boot file would almost certainly be pure instructions without any punctuation (like “OBJ”).

I don’t recall the internal format of a fully linked program ready for load, but it’d be pretty close to a pure memory image.

13

With the demise of punched cards, a popular practical joke has faded into history as well. Throwing some punched card “confetti” into someone’s bed was good for a laugh (except for the victim). Those itsy-bitsy squares are pointy. Makes for an unpleasant night until you manage to get rid of the stuff again. :smiley:

14

I also seem to recall someone preparing a special “program card,” which, when used to prepare a new card, tapped out the rhythm from The Lone Ranger.

In the back of my mind, I can still hear a punchcard machine in operation, including sounds like the dropping of the metal plate that sat on top of the stack of completed cards, when you pulled the stack out. Funny how things like that stick in your memory for decades. Still nothing like the ear-shattering clatter of a paper tape machine, though!

15

I haven’t read the whole thread, but I wanted to say that LSLguy’s photo link brings back memories.

In the mid-1980s I worked for an aerospace contractor in L.A. Although we could easily create JCL on the terminals, the boss insisted that all JCL be punched onto cards. To submit a job, we would have to take the cards to the computer room, where the operator would load them into the reader and run them. The rationale was that we would have a physical record of the job. :rolleyes:

After a couple of years, some people started using JCL to run some jobs; but the punch cards were still there. Finally, around the time I was laid off in 1992, IBM said they would no longer service the machines and the boss was forced to accept terminal-entered JCL.

When I worked at Edwards AFB in the early-'80s, we had to use punchcards on the Cyber 77 mainframe. In that case, we typed the code onto Lee Data terminals and they were punched in the computer room.

16

I’ve only ever had to manually punch a few control cards to insert in front of decks and so on. That was trying enough. The Sigma machine I was doing this for also supported a timeshare environment using model 33 teletypes, and some Tektronix storage tubes. Those are also good topics for memory lane.

Somebody already provided this link:

http://www.columbia.edu/acis/history/029.html

You should take a look at the picture of the card halfway down the page to get an idea of the character coding. Although there are 12 columns on the card, only 64 characters are represented, making use of only a small number of the 4096 possible combinations of punched holes (including all unpunched). This derives form use of cards before the days of actual computers, to allow easier handling by electromechanical devices - characters are easily classified on the basis of very simple patterns, so mechanical sorters, collaters, tabulators, etc. can operate on them.

A “binary card” was done by simply reading down the columns as a stream of bits, 1 and a half bytes per column, for 120 bytes per card. Siggy could handle those, as could many mainframe environments of the era, but I don’t think they were used much.

Card readers towards the end of the punched card era were pretty fast for mechanical devices. I don’t know the actual rates, but an operator could shove a couple feet worth of cards into the hopper, and it just sucked them through and read them in about the space of a breath.

I also knew a guy who had picked up an obsolete card punch somewhere (he had a whole garage full of crap like that). I kept trying to convince him to write a driver for it so that he could punch card decks on his PC, but he never did, for some reason.

17

A couple of rules about punch cards:

(1) Always use TWO rubber bands to hold your stack of punch cards together

(2) Always number each card so you can quickly re-arrange the cards if you forgot rule (1) and the one rubber band you are using broke

(lesson learned the hard way)

The card punch I used just emptied into a trash can. I have no idea what ultimately became of all the confetti out of it.

I used to have a cartoon of a computer with a confused look on its face thinking “Oops! How do you erase a hole?”

18

Apparently these pieces of paper really were called chads – I hadn’t heard the term before 2000. Merriam-Webster gives the origin as ‘perhaps from Scots, gravel’.

19

Wow, a thread about punch cards! So I open a drawer in my desk and pull out a bunch of old punch cards. Haven’t looked at these for a long, long time. I had punched my initials (with the pattern in the holes) on these cards ages ago and have used them as bookmarks. At least for me, punch cards had a lot of uses beyond carrying my old job cards and Fortran programs, such as for grocery lists, index cards, etc. Punch cards became obsolete when people can enter lines of code via teletypes and later CRTs.

20

I still have the entire data set used for my master’s thesis on punch cards, just in case anyone might want to use it for subsequent research. :dubious: