On 2017-12-04, at 15:55:20, Charles Mills wrote:
> I believe someone (Harlan Mills? Fred Brooks?) said that he felt the only (or
> most significant?) *error* in the System 360 design was the 24- rather than
> 31- or 32-bit addressing.
>
Gordon Bell?:
http://gordonbell.azurewebsites.net/computing%20versus%20corporate%20platforms%20011204.ppt
> Anyone who has wrestled with legacy control blocks in the modern era would
> probably agree.
>
Yes. But what was modal memory size and address size across all architectures
in 1965? Very few years after that, a co-worker struggling with a CDC 3600 with
15-bit address registers extended by bank (segment, AS, ...) switching snarled,
"I'm an 18-bit address man!"
Cost of storage for long pointers, and microcycles for carry propagation in
address resolution. Brooks rejected DAT because of the cost of address
resolution.
Some OS/360 control blocks had to be addressed with 16-bit pointers.
20-bit addresses suffice for 640K.
I've heard similar arguments that the cost of Y2K conversion was fully
justified by the capital investment in data storage saved by using 2-digit
years rather than 4-digit.
-- gil
