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

[Brian E Carpenter]

>...Stecker once gave us a lecture/seminar about the PDP-11 project in the early/mid-70s when I was a student. The question was tossed to him about why the 11 was byte-swapped and he replied that it had saved 10ns in the carry look-ahead logic on what would be the 11/20 processor design.  My memory is that it was Bill Wulf who countered -  "but you cost us hours in programmer time."

A little bit disingenuous though. The 1970 AFIPS paper "A new architecture for mini-computers - The DEC PDP-11" was authored by G. BELL, R. CADY, H. McFARLAND, B. DELAGI, J. O’LAUGHLIN, R. NOONAN and W. WULF.

Regards
   Brian

On 31-Dec-20 04:46, Clem Cole wrote:
> 
> 
> On Tue, Dec 29, 2020 at 9:41 PM Brian E Carpenter via Internet-history <internet-history at elists.isoc.org <mailto:internet-history at elists.isoc.org>> 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
> 
> His two main grad students at CMU were Harold McFarlandand Bill Strecker, and much of the original work on the PDP-11 was done there. I was always under the impression that McFarland was the higher bit, but I believe Strecker designed a great deal of logic also.   IIRC it was Dave Cane (who did the SMI and later the BI and of course lead the 750 team), once told me that McFarland's most important contribution was the Unibus protocols (Dave had joined the HW team to work on the memory system for the 11).
> 
> FWIW: Wulf and his students work on compilers and his paper on instruction sets regularity were definitely in the team's mind and of course the original BLISS-11 compiler was done at CMU during that same time.  So some credit really needs to go to those folks too on the ISA.
> 
> As for byte swapping, I remember that Stecker once gave us a lecture/seminar about the PDP-11 project in the early/mid-70s when I was a student.  The question was tossed to him about why the 11 was byte-swapped and he replied that it had saved 10ns in the carry look-ahead logic on what would be the 11/20 processor design.  My memory is that it was Bill Wulf who countered -  "but you cost us hours in programmer time."
> ᐧ