DIYbanter - View Single Post

J. Clarke[_2_]

In article , says...

On 10/18/2011 12:08 AM, m II wrote:
Sorry Gates was not writing any O/S for the 6800 or anything back then.
He certainly didn't write OS-9, CP/M or MP/M.
Execution and access traps were non-existent in microprocessors back
then and self modifying code would not violate the traps back then. I
believe the 80386 was the first acclaimed "real" microprocessor that
could error out on violations of memory boundaries similar to mainframe
machines. Self-modifying code would get you kicked out of any recognized
University for heresy at any time in history. Real coders just don't do
it and keep a job.

Gate wrote very little code as he wasn't very good at it. He was a
marketing genius in the right place at the right time. Another time in
history he might have been a flop doing the same thing. His stuff was
impressive from the outside but very dirty inside.
...

If he wasn't the code author, then hardly fair to blame him for being
the author of whatever, is it?

You're just nuts...

An aside, but FWIW, the '286 had a full set of protection mechanisms--
what it lacked was the ability to virtualize itself and run code written
for a machine with protections disabled in a protected virtual machine.
Unix System V ran fine on the 80286 with all the protections in place,
but there wasn't a way to run a DOS box under Unix other than by
switching the CPU to unprotected mode and back. And there was a bug in
the hardware that caused problems with that switch--Novell, AT&T, and
others managed to work around the bug, but somehow Digital Research
never did and lost a lot of momentum as a result.