[ih] byte order, was Octal vs Hex, not Re: Dotted decimal notation

[John Day]

That was what I always thought.  (We had one of the very early PDP-11/20s, when it was just a PDP-11.)  ;-)

> On Dec 29, 2020, at 23:49, Brian E Carpenter <brian.e.carpenter at gmail.com> wrote:
> 
> On 30-Dec-20 16:58, John Day wrote:
>> Isn’t little endian forced by the byte-order addressing of the PDP-11?
> 
> Exactly. But why did DEC decide to number the bytes that way
> round? I'm guessing that it just seemed natural: byte 0 is obviously
> less significant than byte 1, just as bit 0 is obviously less
> significant than bit 1. Much of the world has chosen to disagree.
> 
> Like much else, it's Turing's fault. He designed the first version
> of ACE, whose final version mutated into DEUCE. It was Turing who
> labelled the LSB #1 and wrote it at the left. I reckon Gordon Bell
> just followed along.
> 
> But of course, there is an alternative history, as so often:
> https://history-computer.com/ModernComputer/Electronic/PDP-11.html
> which states that the actual design was done by Harold McFarland.
> 
>   Brian
> 
>> 
>>> On Dec 29, 2020, at 22:30, Brian E Carpenter via Internet-history <internet-history at elists.isoc.org> wrote:
>>> 
>>> P.S. It happens that Bell learnt to program on an English Electric DEUCE,
>>> as a Fulbright scholar in Australia, and on DEUCE the bits were numbered
>>> from 1 to 32, with bit 1 normally written at the left because it came
>>> out of the serial memory first. Bit #1 represented 2^0; bit #32 represented
>>> 2^31. It wasn't a choice but a necessity, because a serial adder needs the
>>> LSB first. So serial machines were little-endian by nature, and extrapolating
>>> that to byte order might have seemed natural. The other thing about DEUCE
>>> is that it was a register-to-register machine, and Bell certainly copied
>>> that for the PDP-11.
>>> 
>>> Regards
>>>  Brian
>>> 
>>> On 30-Dec-20 15:40, Brian E Carpenter wrote:
>>>>> You'd probably have to find a PDP-11 designer...
>>>> 
>>>> There was only one, really, Gordon Bell. I'm sure he didn't work alone
>>>> but for a basic decision like endianness, I'm pretty sure it was him.
>>>> 
>>>> Allegedly, he's "GBell At Outlook Dot com"
>>>> http://gordonbell.azurewebsites.net/
>>>> 
>>>> Regards
>>>>  Brian
>>>> 
>>> -- 
>>> Internet-history mailing list
>>> Internet-history at elists.isoc.org
>>> https://elists.isoc.org/mailman/listinfo/internet-history
>> 
>> 
>