In the early decades of the industry, computers used punch cards for data storage and transmission, partly because they were already widely used for pre-computer data processing; indeed, a major reason for the ascendance of IBM as the number one computer company was that they already had relationships with businesses using IBM tabulation (electromechanical punch card processing) equipment, so it was straightforward to upsell them to IBM computers.

Was it ever possible to take an actual existing punch card reader, and hack a connection to hook it up to a computer? (By analogy with the way existing audio cassette decks could be hooked up to some models of 8-bit computers to use cassettes for data storage. Or for that matter, with the way existing teletypes could be connected to computers, complete with paper tape readers and writers.) Or was it the case that yes the existing cards could be reused, but to use them with computers required designing and building all new equipment?

  • 3
    Before computers, a "punch card reader" didn't exist on its own. It would always have been a sub-component of a machine that did something with the data that was read (e.g. a printer, or card sorter). Obviously a "computer card reader" would use existing designs of the parts for card unstacking, transport through the read head, stacking after reading, etc.
    – alephzero
    Apr 12, 2021 at 18:55
  • Considering how many times IBM changed their character sets, I am expecting an answer that it was physically possible, but required special decoding of the characters.
    – DrSheldon
    Apr 12, 2021 at 18:55
  • 3
    @DrSheldon LOL. Just, for the real important parts punch card code was quite stable. the WERE the data banks of their time.
    – Raffzahn
    Apr 12, 2021 at 18:57
  • 1
    @Raffzahn: I can imagine an IBM salesman telling a customer, "Sorry, you're going to have to buy a whole new card reader because it's a different kind of EBCDIC." ;^)
    – DrSheldon
    Apr 12, 2021 at 19:05

2 Answers 2


Was it ever possible to take an actual existing punch card reader, and hack a connection to hook it up to a computer?

No, why? The existing tabulating machine was all about parallel processing.

Or was it the case that yes the existing cards could be reused, but to use them with computers required designing and building all new equipment?

No. Of course not. But 'using' equipment is more than just plugging it in.

It's no either/or, but one as well as the other. All are parts of a process. The process is more than a single machine or computer and like with every change, a gradual one, possibly were bottlenecks are, is preferred over a green field approach. In turn one must look at the gradual changes that happened.

Before (electronic) computers tabulating offered process steps like this:

  • Data gets collected on punch cards, either
    • by being outputted by previous runs or
    • by being keyed in with a key punch
  • Processing happened
  • Data gets outputted
    • Into new cards by a card punch
    • Onto paper, by a printer, which in turn could be handled by

So there are a lot of tools and steps for data processing to be used. And while most (andfinally all) can be replaced by a computer and additional devices, a new computer does also ft nicely into the existing landscape by first only replacing the tabulator. To do so he needs to be able to input and output cards, nothing else.

Data gets still collected by halls of young girls typing them in (or grumpy old men in an autoparts store), stacks geht moved to the computer, processed there and then moved over to a printer (printing tabulator). So he could replace all, or most of the machinery, couldn't he?

Well, yes, but why occupying precious processing time of that machine with a mundane task like

  • sorting cards, when there's still the existing sorter?
  • searching for records, when there is still a dedicated sorter?
  • searching more convenient when there is still a collator?
  • not to mention merge by simply stacking cards by hand?

Why blocking the computer with handling an (additional to pay for) printer, when that could still be done with the old printing tabulator, now plugged for straight printing out whatever is on a card?

Any reasonable manager would only use the computer to replace the most strained device, the bottleneck - keep in mind, if there is no bottleneck, one doesn't need a new computer in the first place - and leave all the other parts of the process as they are.

Of course, with faster machines and management seeing the benefits of buying a computer, it got additional card readers and card punches and sorters as well as collators went out of use (*1). Interpreters got replaced not by computers, but by card punches that could as well print and new key punches with the same ability.

The last one is an important point. with having a computer, it first extended and later gradually replaced the more complex separate machines, but need for key punches dod not vanish, but increas, so much that IBM (and others) introduced new models not only with the /360 in 1964 (029), but new designs as late as 1971 with the IBM 5924 Kanji keypunch and the transistorized and programmable 129.

*1 - Still, I've seen sorters used, or at least on stand by as late as 1978.

  • 3
    In the company I worked for, while the IBM mainframe was normally accessed via text or graphics terminals, there was still an ancient IBM 1130 for the times when somebody needed to duplicate, sort or print some punched cards. IRC it finally died from old age sometime in the 1980s - by which time there were only 2 or 3 people left who could remember how to use it!.
    – alephzero
    Apr 12, 2021 at 19:02
  • 2
    I saw a card sorter still in (occasional) up through late 1982 (but at a college, not a commercial setting). Apr 12, 2021 at 21:08

Reading a single punched card is not very difficult. You just need enough photodiodes or spring pins to sense all the rows in one column, and a mechanism to advance the card precisely from one column to the next. The tricky part is handling many cards in sequence, and keeping them in sequence in the output tray for later reuse.

IBM had already solved these problems when building high-volume tabulators and sorters, a business it inherited from the Hollerith company which it had absorbed some time before. It was a relatively natural thing for IBM to adapt its existing card reader and card punch hardware from tabulator to computer interface. These were the primary means of getting programs and data into the 1401 business computer, there being a card reader mounted on the side of the main CPU cabinet.

Companies that bought rented a 1401 were usually already using punched cards extensively. Indeed, there were usually a card sorter and several manual reader/punch terminals in the computer room, as they were useful for operating the computer even though they weren't directly connected to it.

As well as punched cards, another data storage medium widely used in the immediate pre-computer era was punched paper tape, most commonly for handling telegraph traffic. This was adapted for data storage for several early computers, as early as the wartime Colossus (a big loop of tape held the Lorenz messages to be cracked) and as late as the Altair 8800.

Many teletype terminals included paper tape reader/punch units, partly as a hangover from their original use in telegraphy. The computer they were attached to could load data from the tape simply by dint of the user switching his teletype to read the tape instead of the keyboard.

An early computer that relies heavily on its tape reader is the Harwell WITCH, built largely from surplus telephone equipment; the only way to load a program into its memory is to execute a program from paper tape. Most of its peripheral devices considerably predate the computer itself. It has been restored to working order by the Computer History Museum, and there are several videos on YouTube showing it in action.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .