BL> ...... A Pentium is just a Z80 on steroids, moving more and
 BL> more shit about faster. 32-bits have proved a dead loss, mostly
 BL> using twice the memory with a size handicap and no great speed
 BL> advantage so what happens?

 RM> The Skymap program I use was ported from Win31 to Win95 a year
 RM> or so ago, with a claimed doubling of speed in numerical
 RM> calculations.

  Did you notice any difference? It's a sad reality in this wicked
world, that if I make a change, I can nearly always find a selling
point where something doubles. It is very rarely that you find a
genuinely honest person who tells it like it is... so mostly we have
to do our own analyses and come to our own conclusions.

  Did you notice any difference?

  My own experience of Win95 is that is runs slower than Win31, but
that experience is restricted to someone else's computer. Of course, I
could probably write a loop where either one would run faster than the
other...

 BL> I think the next step is going to be a much smaller software
 BL> coupled to the CPU itself, directly... a specialist GUI CPU
 BL> that lets us send much smaller instructions.

 RM> Won't that become obsolete slightly faster than the current
 RM> crop of too-slow general purpose cpu's?

  Yes... but obsolescence is not a sensible criterion, is it? Rapid
advancement is what we want. If you built a specialist GUI CPU that
used objects or token or whatever, then those tokens might change and
obsolete the hardware, but each step would have to be backwardly
compatible unless the new model justified the hardware change.

  It occurs to me that the new VB/RAD software GUI development systems
could be done in hardware too, and minimise 99% of the crap travelling
telephone wires.

Regards,
Bob
  
___ Blue Wave/QWK v2.12
@EOT:

---