Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!csd4.milw.wisc.edu!cs.utexas.edu!halley!tjd
From: tjd@halley.UUCP (Tom Davidson)
Newsgroups: comp.arch
Subject: Re: Why 36 bits.
Summary: CDC 6x00 and character sets....
Message-ID: <544@halley.UUCP>
Date: 29 Jul 89 16:11:35 GMT
References: <355@torsqnt.UUCP> <3490016@wdl1.UUCP> <1451@l.cc.purdue.edu>
Organization: Tandem Computers, Austin, TX
Lines: 22

In article <1451@l.cc.purdue.edu>, cik@l.cc.purdue.edu (Herman Rubin) writes:


> The CDC 1604 and 3600 were 48 bits, and the CDC 6x00 was 60 bits.  These
> machines only had upper case, and 6 bit character sets.  They did use the
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> more efficient column binary.

At least in the case of the 6x00, 6 bit character sets were the norm.
The character set was called display code and came from the fact
that the system console understood 6 bit characters. (get it? display code)

Under the timesharing operating systems KRONOS and NOS, there was two
other character sets which were used.  Extended display code (or
more commonly called 6/12 ascii) reserved the codes 74(8) and 76(8)
to escape into an expanded set in the next 6 bits.  This made programming
rather fun, since not all characters were the same size.  Also, real
ascii(yup!) could be had if you didn't mind using 8 bits out of 12.
Packed ascii (8 in 8) was(is) used in the network interfaces where you
had 7.5 characters in one 60 bit word!

Alas, most compilers understood only the 6 bit and 6/12 characters sets.