Re: backward architecture, The Design of Design

According to Michael S  <already5chosen@yahoo.com>:

The /20, /44, and /67 were each for special markets. ...

But programs (or OSes) that utilize the features of /67 would not run
on anything else, right?

How about programs that depend on precise floating-point exceptions?

As I think I said in the message you quoted, those three models were
for special markets and they weren't fully compatible with other
models. Having run my share of Fortran programs on a 360/91, I can
report that you used the same Fortran compilers as you'd use on any
other model of 360. The imprecise interrupts were a pain for debugging
but in practice no useful Fortran programs depended on catching and
recovering from floating point exceptions so it didn't matter.

I also wrote some programs for a /20 and considering that it was a 16
bit machine with 8 registers, it was surprising how similar the
programming was to a real 32 bit 360 with 16 registers.

The other eight models were extremely compatible including the exceptions.

What about various vector facilities that they were adding and removing
seemingly at random during 1970s and 1980s ?

Those were defined as special features. If you wanted to run your
program, the machine needed to have the features your program used.

More like 1957. The IBM 705 was mostly backward compatible with the
702, and the 709 with the 704. But only in one direction

>
"One direction" is synonymous with "backward compatible", is it not?
But the word "mostly" is suspect.

It was close enough that you could run 704 programs on the 709 or
7090. Read the 709 and 7090 manuals at Bitsavers if you actually care
about this.

Can you, please, define the meaning of upward and downward
compatibility?

See page 5 of "IBM System/360 Principles Df Operation" published in
1966. This is not new.
--
Regards,
John Levine, johnl@taugh.com, Primary Perpetrator of "The Internet for Dummies",
Please consider the environment before reading this e-mail. https://jl.ly