Memory Models - Thanks for the info

[mcdonald at uxe.cso.uiuc.edu]

>From article <10744 at smoke.BRL.MIL>, by gwyn at smoke.BRL.MIL (Doug Gwyn):
> People who aren't "wedded" to the *86 architecture generally don't
> seem to think it was necessary to cause memory models to be visible
> in higher-level programming languages.

Nobody, even those like the '86, would claim that memory models
need to be visible. They didn't. The compilers could have made
everything large model. BUT - this would have put compiled
programs at a very serious performance disadvantage. The use
of memory models, especially mixed ones, is a code optimization,
pure and simple.

Due to the competitive nature of the 80x86 market, if Microsoft/IBM
had offered only a fixed (large) model compiler, someone else would\
have sold one. A similar situation, but in reverse, exists with
80386's - Microsoft and IBM refuse to sell a 32-bit, fixed-mode compiler 
(small model - 32 bit integers and pointers; large has 48 bit pointers).
So others sell them. They are not smash commercial successes ONLY
because of their onerous licenseing policies (i.e. you can't distribute
whole, runnable programs without paying them extra BIG bucks. Thus, while
I have several nice 386 programs I would like to distribute (and can)
they would be useless except to people whao have paid $495 (gag) for
the necessary stuff from Phar Lap).

Doug McDonald