In an answer to Why did IBM System 360 have byte addressable RAM I wrote regarding the choice of byte size:

7 bits would be a perfect match for ASCII, but engineers would instinctively recoil from basing the word sizes on a prime number.

(Okay, someone else pointed out that EBCDIC actually uses 8 bits, but that would only be a concern for IBM in particular.)

Now it occurs to me to wonder, did anyone ever build a computer with a 7 bit byte? Or with a 14, 28 or 56-bit word?

  • 2
    28 bits (as 32 - 4) as addresses appeared in a few machines. And computers were not designed to process 7-bit ASCII, but to do calculations; so that would be a more important consideration than prime numbers or not. – dirkt Jul 8 '20 at 4:40
  • 2
    There were a number of computers with 6-bit bytes early on. – Hot Licks Jul 8 '20 at 13:14
  • 2
    @smci In my era (1970s) the eighth bit was used on paper-tape and comms channels as a parity bit, to detect (but not correct) single-bit errors on the media. Wrong parity would stop a papertape reader on the faulted row, but for comms it would send ASCII NAK to elicit retransmission. I worked on ICL 1900 series mainframes, which had a word size of 24 bits (4 x 6-bit chars) so 7-bit data used (effectively) wide chars to distinguish upper and lower case and control characters, which did not exist in the native 64-char set. – Paul_Pedant Jul 8 '20 at 15:01
  • 2
    @smci, Re, "I thought ASCII was a kludge to fit 7 bits of alphanumeric data inside an 8-bit byte." ASCII wasn't created for computing. ASCII was created for the telegraph system. en.wikipedia.org/wiki/ASCII#History It was the wide availability of telegraph equipment (especially Teletype model 33 machines) that could be adapted as computer I/O devices that dragged ASCII into the computing world. – Solomon Slow Jul 8 '20 at 15:30
  • 6
    @Paul_Pedant: I think the idea behind parity was that in many cases, having a character be recognizably unreadable would be tolerable, but having it appear as the wrong character would not. If an "error character" appears in a place where t█e meaning is obvious, the recipient of a transmission may just fix it, but if such characters would render the meaning unclear, a retransmission can be requested. – supercat Jul 8 '20 at 16:33

The PDP-10 had 'byte instructions' that could process a sequence of bytes of size 1 to 36 bits. The byte pointer was a word containing an 18-bit word address (and the usual index/indirect indications) plus position and size of the byte within the word.

It was common to use 7-bit byte sequences for ASCII text, which gave 5 characters per word and one (usually) unused bit.

There were monitor calls (system calls) that accepted strings of 7-bit characters packed in this way.

So: at the hardware level, bytes were variable-sized. At the software convention level, a byte was frequently 7 bits.

See section 2.3 in the KA-10 system reference manual.

  • 4
    PDP-10 flexibility in this way was very useful as operating systems also used sixbit encoding in places, e.g., for representing filenames in the file system in TOPS-10. – davidbak Jul 8 '20 at 11:35
  • 5
    See also RFC 4042. :-) – Lars Brinkhoff Jul 8 '20 at 16:08
  • 1
    Side note: In a BASIC program, the line number was stored as 5 7-bit bytes in a word, and the left over bit was set to 1 to indicate this was a line number. I discovered this when I tried to create a BASIC program with a conventional text editor. It looked fine, but it didn't work because that extra bit wasn't set. (Wow, this was like 40 years ago.) – Jay Jul 8 '20 at 20:15
  • 1
    Sure. SOS used to set line numbers too. I think most of the language processors knew to ignore them. – another-dave Jul 9 '20 at 0:32
  • 1
    The same answer applies to the PDP-6. It had byte manipulation instructons like those of the PDP-10. – Walter Mitty Jul 9 '20 at 11:42

The VT52 text terminal certainly doesn't qualify as a full computer, but it does have a processor running software out of a ROM. The RAM holding the displayed text is 2048 7-bit bytes. The character generator ROM is also 7 bits wide.

  • But were those 7-bit bytes or 7-bit words? ;-) – another-dave Jul 8 '20 at 12:16
  • 3
    @another-dave Yes. ;-) – HiddenWindshield Jul 8 '20 at 15:27
  • @another-dave, maybe and? – Lars Brinkhoff Jul 8 '20 at 16:04
  • 1
    According to the VT52's maintenance manual, the machine had an 8 bit data bus and used the full 8 bits internally for its microcode instructions. So it could be said to have an 8-bit word size. However, only 7 bits were used for its character set. – Paul Humphreys Jul 9 '20 at 10:07
  • 1
    @PaulHumphreys, I see 7-bit data buses in the block diagram in figure 4-1. Yes, the instructions are 8 bits. Program addresses are 10 bits. – Lars Brinkhoff Jul 9 '20 at 11:52

The second-generation Soviet computer Minsk-32 (the series size is 2889 machines, 1968-75, civilian use, one of the rare early mainframes noted for use in Antarctica) used a 37-bit word and 7-bit representation of alphanumeric characters (5 in a word). Yes, the concept of "bytes" is difficult to apply to a similar old computer (which continued the line of vacuum tube machines), but special commands for the convenience of operations with 7-bit blocks took place in the command architecture.


Yes; there have been several (although, to my knowledge, none in the most simple sense where seven binary bits are treated strictly as as a base-7 system of Peano-like numbers). Instead, they are systems in which at least one (typically, two or three) carry are treated as separate state-modification bits.

The most oldest/most simple example (although it may not meet the definition of a Turing-complete computer) is the ancient 5/2 abacus.

More recent examples generally are cases where some form of binary-coded decimal is used, particularly those that use Chen-Ho encoding (which fit a better conception of the system being "7-bit", as Boolean logic/operations can still be (relatively) easily applied, as opposed to more packed (or packed/padded) 7-bit numbers, which require a variable number of instructions to ascertain certain binary/two's complement values.

Of these, the "two of seven" approach is most common. Examples: the IBM 650, the FACOM 128, and the "IBM 370 compatibility feature" (hardware emulation) built in to the IBM 7070/7074.


The well-known IBM 1401 technically had a 7-bit byte (plus parity). It was designed around the common format of IBM punched cards, which it was designed to process; these had ten "digit" rows and two "zone" rows, of which one digit and optionally one zone (for which the zero row also counted as a third zone) could be punched simultaneously in each column.

This essentially meant that each column of a card contained a value that could be encoded in 6 bits - 4 representing a BCD digit, and two more indicating the zone (no zone, first zone, second zone, or zero row). To this, the 1401 added a "word mark" bit for internal use, for a total of 7 bits.


The Norsk Data ND-505 had a 28-bit address bus.

  • 2
    So it did! But only physically not architecturally: "The only significant 28-bit computer was the Norsk Data ND-505, which was essentially a 32-bit machine with six wires in its address bus removed." – rwallace Jul 8 '20 at 1:13
  • 4
    @rwallace six? not four? – OmarL Jul 8 '20 at 6:41
  • 11
    It has taken 14 years for the arithmetic in en.wikipedia.org/wiki/Special:Diff/56879895 to be challenged. – JdeBP Jul 8 '20 at 13:41
  • 1
    The question here talks about the bit size of bytes, not how many of them could be addressed. It seems to me that a machine having a 28 bit address bus wouldn't be particularly relevant to it. – Paul Humphreys Jul 9 '20 at 10:18
  • 2
    @OmarL - I haven't looked into this machine specifically, but machines with 32-bit wide memory interfaces frequently lack A0 and A1 address lines, i.e. if maximum address size is also 32 bits, they have 30 address lines. Presumably, this system had address lines A2-A27, hence the suggestion of 6 lines being removed. – occipita Jul 10 '20 at 1:23

The ADAU1701 is a 28-/56-bit DSP for audio processing. CHAR_BIT is probably 28 on that platform like most odd-sized DSPs but I'm not quite sure since I couldn't find its programming manual

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