Who introduced the standard that later became widespread for the 8-bit punch tape, and when? I think the 5-bit tape was an earlier standard.

  • 6
    It'll be a telegraphy thing rather than a computer thing. FWIW, though, there were 5, 6, 7, and 8 channel paper tapes. I skimmed the Teletype material at bitsavers, and it looks like the model 33 (1963-ish) was the first to support an 8-bit code, so that likely sets the approximate timeframe for 8-hole tape. Jan 8 at 17:07


5 hole (*1) tapes were developed as part of Murray's teletype system, the first (successful) to use typewriter keyboards as well as punch tapes. Wider (more holes) paper tapes are compatible enhancements thereof.

There is no sole 'inventor' per se, as the 8 (+1) hole punch with fixed width tape of 1 inch is the logical extension of existing 5 hole format. What can be said is that the Teletype Model 33 made it a cross platfom standard

The Long Read

The first media to store serial transmitted data on a punch tape was the two hole format developed by Charles Wheatstone (*2) in the 1850's. It was meant to store Morse code. Unlike later formats it didn't store symbols but transmission elements. Dots went into one column, dashes into the other and spaces were simply not punched.

The 5 hole encoding was developed by Baudot in the 1870s, but it wasn't until the 1900s that Donald Murray added a paper tape to his modified Baudot system (*3). Murray's system defined the paper tape as

  • 11/16 inch width
  • 1/10 inch hole spacing (horizontal)
  • 1/10 inch line spacing (vertical)
  • AWG 13 (0.072 inch) 'data' hole diameter (*4)
  • AWG 17 (0.046 inch) transport hole diameter (*4)
  • The 6 hole positions were arranged as
    • two 'data' holes
    • one transport hole
    • Three 'data' holes

Later 6..9 hole systems imply extended that tape with more holes, following the standard, to the 'left' - and of course used way different encodings.

Width of these tapes were specific to each extension. Manufacturer usually extended existing punch designs by adding more holes to the 'left', keeping all transport mechanic (and thus the transport holes) the same. It saved not only cost in design and production, but resulted in great compatibility with the existing 5 hole types. To enhance this some readers were made in a way to use no left stop, not limiting the tape width at all (*5).

In fact, some systems (like certain Friden Flexowriter punches) were intended to use arbitrary wide sheets to allow punching of data along files.

'The' Flexowriter is often mentioned as a first, but beside that there was no single model, but a family of different and often customer specific models, it did use many different formats. Even with an installed 8 hole punch installed, codes used were 7 bit, using the 8th hole at maximum for parity information - or special markings.

Examples with computers are:

  • LGP-30 Flexowriter
  • Alwac III Flexowriter
  • 7-level Flexowriter

There were in addition many codes that used constant bit numbers, like 3-out-of-7 (More, CCITT#3) or 4-out-of-7 (CCIR 476).

It wasn't until the 7/8 bit ASCII TTYs, more specific the Teletype Model 33, that made a compatible 1 inch wide tape the defacto standard for everything up to 8 holes. Not to mention that ASCII as well finally made numeric values the base of symbols.

While early computers used many various formats, next to everything after 1963 went for ASCII and 8 hole paper tapes for their control applications, new mini computers and later micros.

*1 - At that time there were no bits. Holes represented a sequential value, not a numeric one. Numeric interpretation only came with computers.

*2 - Yes, the very same man who developed the bridge named after him, well known to anyone who ever peeked into electronics.

*3 - Baudot's original code was made to be entered using a piano like keyboard with 5 keys, with the letters arranged in a way to ease typing. All text was a continous stream, with out any control characters (beside erase that is).

Murray added a typewriter style keyboard, mechanical code conversion and format characters (LF, CR, BELL) making the output (more) like a typed letter. To improve thruput, typing and sending, as well as receiving and printing was separated b use of a punched tape. That way transmission speed was always maxed put be feeding punch tapes, no longer dependent on typing speed. Similar investment in printer could be reduced, as a single, faster printer could be used to print out paper tapes from several receivers.

To reduce stress on his 5 hole punch tape, he rearranged the letters in a way to give the most common letters the least number of holes, so Baudot's ITA1 became Murray's base for the later ITA2 (which includes changes from Western Union, the first adaptor of Murrays system as well as some CCITT ones).

*4 - Yes, these are not the usual fractional measurements, as the pins used to punch were made from 'wire', so AWG it is.

*5 - In this case usually two transport sprocket wheels were used to avoid misalignment.

  • 2
    Wikipedia supports the position that teleprinters switched to a 7-bit code when ASCII came along, and I suppose that would be the natural point for 8-hole tape (no reason before that). Why 8 rather than 7? (Sprocket holes don't count). Future-proofing. The Teletype model 33 tech manual describes the device as supporting an 8-level code, though "at the present it utilizes only the first seven intelligence elements (=bits!)". Jan 8 at 20:46
  • 1
    "8-bit bytes" is an entirely separate issue, and that has a well-known answer: IBM System/360. It is independent of 8-hole paper tape carrying 7-bit ASCII, and independent of transmission of 7-bit ASCII in 8-bit data frames (plus 1 start bit and 1.5 stop bits) Jan 8 at 22:00
  • 1
    @Anixx No, the 8 bit byte was introduced by IBM with the system /360 - although the machine was originally intended to use ASCII - but ASCII design took way too long (as it's known with committees), so IBM went with EBCDIC. And yes, the while I would not call Teletype the 'inventor' of 8 hole paper tape, their Model 33 created the de facto standard everyone followed - not at least because it did so nicely fit the freshly established 8 bit byte.
    – Raffzahn
    Jan 8 at 23:59
  • 1
    Per this link from Wikipedia, the 8-channel M35/M33 appeared in 1961, though the quoted "6 years of development" puts some fuzziness around the date that 8-hole tape was adopted. Jan 9 at 21:36
  • 1
    @another-dave it was common with 7-bit ASCII to use the 8th bit for parity, giving you a rudimentary form of error detection. Jan 11 at 16:13

The earliest contender I can find is the Friden Flexowriter, from 1958. It was an electric typewriter with an attached paper tape reader and punch. You can see a video here. This is a restored Flexowriter, but the visible sign lists 1958 as the year it was presumably built. As you can see, this particular model uses 8 channel tape. The Flexowriter would have been suitable for offline preparation of source code for computer programs.

Soon after, Digital Equipment Corporation introduced the PDP-1 computer, the first computer built by Digital Equipment Corporation. (prototype 1960). The PDP-1 had 8 channel (fanfolded) paper tapes with a reader and a punch.

The PDP-1 was patterned after the TX0 computer built in the 1950s by Lincoln Labs. I am not sure what kind of paper tapes the TX0 worked with. It did have a magnetic tape drive.

As other answers have pointed out, ASCII was a later development.

  • KDF9 used 8-channel paper tape with 6-bit + parity encoding (bottom of page here). We could use a Flex to prepare Algol 60 input tapes, so I suppose the Flex was producing the same code -- though this is not necessarily true, since device- and even application-dependent code was not unusual then. Internal code, printer code, paper-tape code, and teletype code were all different. Jan 9 at 15:21
  • If that particular model was indeed a 1958 model, and if the 8-hole tape unit was original on that device, then Friden has a better claim to 'first' than Teletype. Jan 9 at 15:32
  • The first time I ever saw a Friden Flexowriter was in late 1962. It definitely had 8 channel tape at that time. The video looks to be of the same model, but I can't tell. I think the one in the video is 8 channel tape. For the year 1958, all we have to go on is the sign in the video. Jan 9 at 21:08
  • 1
    You can read the MIT produced operations manual for the Flexowriter here. Jan 9 at 21:19
  • The PDP-1 Handbook from 1960 has Flex codes, so 8-channel readers on Flex go back at least that far. And the photoelectric reader on the PDP-1 is also 8-hole, as you mentioned. So we're back to 1960. It would be good to find original Friden docs to be sure that 8-hole readers were in use on the 1958 models. Jan 9 at 21:48

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.