Re: ASCII control codes in sequences of multibyte character sets

From: Philippe Verdy <>
Date: Tue, 3 Sep 2013 07:36:20 +0200

I think that microcontrolers with 1-bit, 4-bit and 8-bit bytes are still
used today. The size of their bytes do not vary but are in separate address
spaces, each of their bytes are single units of addresses, with their own
dedicated instructions. Microconstrollers like this include the Intel 8044
family and its numerous derivatives (oten integrated within other
programmable chips). They are powerful enough to support C programming, or
to support a full TCP/IP protocol stack, IEEE 32-bit or 64-bit floating
points. They are not generally well suited to support compluser UI, so they
will generally not have to handle encoded text except in very limited way
for numeric conversions. But anyway these devices will communicate with

2013/9/3 Christopher Vance <>

> The last machines I used with bytes not being octets were a Univac 1182
> with bytes of 6, 9, 12, and 18 bits, and a PDP-10 with bytes of variable
> size from about 2 (or 1?) up to 35 (or 36?) bits. Both machines had 36-bit
> words.
> The Univac probably didn't call these smaller chunks bytes, but the PDP-10
> definitely did.
> The Univac used 9-bit quarter words to hold ASCII strings, while the
> PDP-10 packed 5 7-bit bytes into a 36-bit word to represent ASCII strings.
> Both systems also used (different) 6-bit character sets for some purposes.
> On 3 September 2013 11:58, Asmus Freytag <> wrote:
>> On 9/2/2013 6:47 PM, Doug Ewell wrote:
>>> In any case, there is nothing about "multi-octet" versus "multi-byte"
>>> that makes one fixed-length and the other variable-length.
>>> Yep.
>> A./
> --
> Christopher Vance
Received on Tue Sep 03 2013 - 00:38:58 CDT

This archive was generated by hypermail 2.2.0 : Tue Sep 03 2013 - 00:38:59 CDT