How CPU's work (was Re: Hi, I'm new...)

[Chuck Guzis]

On 8/6/2006 at 10:39 AM Don wrote:

>The "problem" with the MPU was putting it in a "computer".

...and that's really the nub of my argument.  I have no quibble with the
application of an MPU to play MP3s, run a scanner or traffic signal, but
the huge mistake IMOHO, was trying to turn them into "Gemeral Purpose
Computers".  I'm not proposing that large-scale integration was wrong--it
alone would reduce the size of your "washing machine" to a deskop box.
Heck, my mouse has a PIC in it--but I'm not about to try to put a word
processor or a DBMS on it.   We're still living with the legacy of the 8080
and CP/M in Windows.

Not your post, but Dave's.  While it's true that there were a significant
number of major differences between the 8008 and the 8080, the architecture
from a software point of view was set in the 8008.  Accumulator, B,C,D,E
registers; H,L handled as a pair to address memory; basic instruction
layout, etc.  The addion of SP and 16-bit addressing was a welcome
addition, as were 16 bit adds and loads and 256 I/O ports, but to my eye,
the 8080 looked like a tarted-up 8008.  One valuable aspect of the 8008
lost was the treatment of 00 and FF as HALT instructions.  

Also, I'm not saying that the CP1600 was a speed demon--AFAIK, aside fromt
he Activision games, it found no general application.  Yet, from a
programming standpoint, the closest thing to it prior to the 68000 was the
LSI-11 (not a speed demon either).

Sometimes, the actions of the people inside the micro business made me
wonder.  When I put my first hard disk (a 14-inch Shugart) on an 8085, I
noticed that there was no one offering disk backup utilities for micros.  I
still have the rejection letter (along with the 8" floppy)  from Lifeboat
saying that they didn't think there was any market for a hard disk backup
package.  What I found incredible was the reinvention of mistakes.  How
long was it before the micro database people realized that atomic logging
of database operations was worse than useless--that one had to log entire
transactions?  Mainframe people knew that before there were even micros.

Cheers,
Chuck