On Fri, May 4, 2012 at 9:24 AM, Denys Beauchemin <[log in to unmask]> wrote:

> I do not believe that Vision was the PA-RISC-based system, that was
> Spectrum.
>
> Vision was never released, but I understand some of its technology did go
> into Spectrum.


If memory serves me correctly, Vision was the 32 bit version of the 16 bit
Classic stack based architecture.  Since HP had been competing successfully
against 32 bit DEC Vax, IBM 3x, and other minicomputers of the time, a 32
bit version might have been a giant killer.  But, I suspect that there are
only so many 6 bit opcodes (assuming the same as the Classic architecture)
that you could pack into a 32 bit word (4 or 5) before you end up with
filling with mostly NOPs (no operation instructions), especially with
business applications with the many branches.  As a side note,
architecturally, the HP3000 was a better design for a scientific computer
with its stack architecture which RPN using HP35 calculators that engineers
understood very well at the time.

But, with the hiring and then influence of Joel Birnbaum, Vision was
dropped in favor of RISC architecture development.  I liked the idea at the
time, and thought Alfredo was off his rocker when he passionately argued
for continuing the Vision project.  In hindsight, I think Alfredo was not
fighting against change, but understood the advantages of a stack based
architecture, though I do have reservations about its ability to scale to
larger word addressing architectures efficiently.

Pete

* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *