A fact from this article was featured on Wikipedia's Main Page in the On this day section on June 8, 2004, June 8, 2005, June 8, 2006, June 8, 2007, June 8, 2008, June 8, 2012, June 8, 2014, and June 8, 2015.

This  level-4 vital article is rated B-class on Wikipedia's content assessment scale. It is of interest to the following WikiProjects:

Computing Low‑importance This article is within the scope of WikiProject Computing, a collaborative effort to improve the coverage of computers, computing, and information technology on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.ComputingWikipedia:WikiProject ComputingTemplate:WikiProject ComputingComputing articles Low This article has been rated as Low-importance on the project's importance scale.

Punch Cards (80 col, hollerith format) were also informally called "5081 cards". 5081 being the IBM part number printed in very small letter along the bottom of the card. This part number was likely for a standard punch card, with the stock printing across all 80 columns. Not certain, but I would expect a different number for other styles of printing (or possibly if the corner was cut on the other end).

Standard Form 5081 was a US Government Standard Form that adopted the IBM 5081 punch card standard. An example of the Standard Form 5081 is what is depicted in the 5081 card image on this page. These were in common use in the early '80s when I was a "Data Operator" at an Air Force Data Processing Center. Bowlingj (talk) 20:00, 19 March 2012 (UTC)Reply[reply]

I have a physical one in hand right now that my grandfather gave me from his time in the Air Force! 68.203.126.152 (talk) 16:00, 25 October 2022 (UTC)Reply[reply]

Cards were typically available in buff (sort of an off beige natural color), red, green and blue.

There were thousands of different card types, each with its own stock number. Large organizations, such as universities, would have cards with their logo in the background. Other punch card suppliers often used the IBM part numbers. I have 5081 cards, which are plain with no column dividers, from Globe and BSC. You can still buy 5081's from Cardamation.com for $32/box of 2000. The standard Fortran card was 888157. I have examples from IBM, CDC and JTC. I have some Princeton University Fortran cards marked IBM N26238 and Stevens Institute of Technology marked JTC 11121.

You could order cards in a wide variety of colors. It was more common, though, to get buff cards with a colored stripe along the top.--agr 03:23, 14 May 2006 (UTC)Reply[reply]

The cards typically had a small numeral 9 printed on the bottom edge and a 12 printed on the top. This was because cards were usually fed face down, 9 edge first into the reader, with a sticker on the hopper thus instructing the operator. This inevitably led to the question "which is the 9 edge?", since that isn't intuitively obvious to the casual observer since the bottom of 12 rows would be the 9 row (the top two had various names, the bottom ten were row zero through 9). -anon165.2.186.10 22:02, 1 December 2006 (UTC)Reply[reply]

I worked on an 1BM 1401 with an 1402 card reader in the '70's. The machine indeed would stop if there was a card in upside down. I think instead of looking at the cut, it would look at the code in the card. If it made no sense (not a valid character), then the program would stop. It would give you an error. You would then do a "runout" and change the orientation of the card. Other machines did the same, The CDC 3300 also would stop on an card being upside down. —Preceding unsigned comment added by 160.91.217.105 (talk) 21:15, 4 February 2008 (UTC)Reply[reply]

I don't think you could reliably detect an upside-down card this way. Many characters produce a perfectly valid (but different) character when turned upside-down. This is especially true of numbers, and it was quite common to have data cards that were entirely numeric. T-bonham (talk) 07:35, 10 April 2008 (UTC)Reply[reply]

The punched version of EBCDIC allows only one punch in rows 1 through 7. Turned upside down, those would be rows 6 through 0, so a card with punches in row 0 and one of rows 1 through 6 would fail the test. Those characters are reasonably likely to appear. Gah4 (talk) 19:31, 6 May 2015 (UTC)Reply[reply]

What is punch card 196.249.97.107 (talk) 09:18, 22 December 2021 (UTC)Reply[reply]

The article lists five examples of prior art then for some reason still assigns the invention to Hollerith. Even if you're drawing the, rather dubious, distinction between instructions and data; Babbage's and Carpentier's designs still fit the category and predate Hollerith. — Preceding unsigned comment added by 81.187.188.94 (talk) 14:53, 2 January 2022 (UTC)Reply[reply]

I believe that there is also a 96 column read/punch for the 360/20. There is a 6 bit code which uses the lower rows only, and 8 bit code which also uses the upper rows (where characters are printed). Gah4 (talk) 02:08, 12 June 2022 (UTC)Reply[reply]

I haven't seen any punched cards actually in use for quite a while. Is there any manufacturer who is still making new blank cards, or have they been completely discontinued everywhere? If they are still available (other than "new old stock", NOS), where? If no longer available, when were they discontinued?

A quick search indicated that some knitting machines still use punched cards. Should this be mentioned in the article?

Reify-tech (talk) 16:41, 6 April 2023 (UTC)Reply[reply]

Living Computer Museum had some made not so many years ago. I suspect that they just found someone who would print and cut them to the right size. The printing wasn't quite as close as IBM would have made. They have, last I knew, an actual 029 to punch them. Maybe Russia still makes them, in case they are still running old computers. Gah4 (talk) 20:35, 6 April 2023 (UTC)Reply[reply]

The article describes "mark sense (electrographic)" cards in the "Mark sense format" section. It may be worth mentioning that in the late 1960s Hewlett-Packard offered optical mark card readers for their HP 2100 series of minicomputers that could read cards marked with an ordinary No. 2 graphite pencil and did not require a special electrographic pencil. They were typically used for HP BASIC programming and had fields where a single mark would encode an entire key word, such as "FOR", "GOTO", etc. They had the same physical size as an IBM 80-column card and were mostly used in environments where multiple users shared a single computer, as in schools. These cards allowed users to write their programs almost anywhere without need for a keypunch machine, then have the computer read their deck of cards when they got their turn. When Hewlett-Packard introduced their time-sharing systems in the early 1970s, having magnetic mass storage on 14-inch disk drives and equipped with terminals such as the Teletype Model 33 that featured paper tape readers and punches, the need for the optical cards and readers quickly faded away. — Foxtrot1296 (talk) 17:04, 10 May 2023 (UTC)Reply[reply]

This is exactly the same system I remember where you certainly did not need a "special pen", it was a card reader attached to an HP2100A. I always figured it was common with any mark-reading card readers but maybe it was a special invention by HP. But I do think the text needs to be fixed to say that a "special pen" was not always needed.Spitzak (talk) 18:35, 10 May 2023 (UTC)Reply[reply]