Re: Translated IUC10 Web pages: Experimental Results

From: John Cowan (cowan@locke.ccil.org)
Date: Fri Feb 07 1997 - 10:49:47 EST


unicode@Unicode.ORG wrote:

> Actually, I programmed 360's, and they were (and are) big endian. I
> believe little endian was introduced by DEC. I can't remember if it was
> the PDP-8 or the PDP-11.

The PDP-11 was a little-endian 16-bit system, AFAIK the first
little-endian system anywhere. Oddly, though, the software-only
support for 32-bit integers stored them in wordwise big-endian, thus:

0x1234 = 0x02 0x01 0x04 0x03

sometimes called "mixed-endian". This was only a convention, of course,
but it was the standard one used by DEC's libraries and inherited by
PDP-11 Unix.

The PDP-8 was a 12-bit computer and had no byte addressing, so no
hardware enforced "endianism". Bits were conventionally numbered in
big-endian order, though, and two 6-bit characters (truncated ASCII
with no lower case or `{|}~ available) could be stored in a 12-bit
word, first character in the high-order 6 bits.

The conventional PDP-8 storage algorithm for 8-bit bytes in 12-bit words
was:

        first byte in the low-order 8 bits of the first word;
        second byte in the low-order 8 bits of the second word;
        4 high-order bits of third byte in the high-order 4 bits of first word;
        4 low-order bits of third byte in the high-order 4 bits of second word;

and repeat indefinitely, giving 3 bytes per 2 words. Painful to
pack and unpack, especially since OS/8 made the application
programmer do the work.

-- 
John Cowan						cowan@ccil.org
			e'osai ko sarji la lojban



This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:34 EDT