news:387fc37f.949660@news-server...

What on earth makes you think that there are customers out there desperate
to put CLIX boxes back on maintenance:-) HEck if they want a UNIX box to
play with, they can pick up a LINX/x86 box for less than a years maintenance
on a CLIPPER/CLIX box. In addition, I'd strongly suspect that field service
has reduced the capability to support CLIX boxes (in terms of spare parts
and expertize) back to a minimum level consistent with the number of boxes
still in active use by customers. The last thing they would want would be a
bunch of people asking to put such boxes back on maintenance. I know it's
hard to accept, but as far as Intergraph is concerned CLIPPER and CLIX are
something from the dim and distant past, like InterAct 68Ks.

To the best of my knowledge this was never even discussed and I was privy to
some pretty harebrained discussions in the first half of the nineties :-)

This has an element of truth in it. Part of the original deal between Ingr
and MS was that Ingr could put NT on a CLIPPER platform. The platform it
would run on, bore no resemblance to any CLIX machine ever shipped. The
prototypes of the CLIPPER/NT platform were publically demonstrated at AEC
tradeshow in Atlanta in 1993. They were very "alpha", NT 3.5 didn't so much
run on them as stumble drunkenly from BSOD to BSOD. That was about as
advanced as the port ever got. The hardware had a PC bus structure (EISA)
and was packaged inside a 2xxx series chassis. There were a number of issues
that were never satisfactorily resolved, these included:

1. Hardware paging management, the low level paging algorithms of NT at that
time matched quite closely the hardware architecture of the MIPS 3000/4000
CPU, this was radically different to CLIPPER and lead to serious performance
issues.

2. Microsoft had no intention of porting any of its applications to the
CLIPPER/NT platform (just like they didn't for other RISC platforms)

3. Intergraph by that time had recognised that it did not make financial
sense to continue throwing money at CPU development for a platform that
would, in a good year, sell maybe 20K CPUs.

4. The CLIPPER C400 was rapidly being overtaken in performance by Intel x86
processors like the 66MHz Pentium launched in 1993.

With these and other issues in mind, the CLIPPER/NT port was inevitably
stillborn. However at the time the need for a RISC based platform was still
something that seemed important. So when the CLIPPER development on C500
architectures was canned, the engineers from APD (Intergraph's Advanced
processor Division) were effectively sold to Sun. The original intent was to
develop a byte sex switchable SPARC architecture that could host a SPARC/NT
port. As part of the deal Intergraph would then sell these SPARC/NT boxes.
The problems with this included:

1. Doing anything at Sun with NT was a political minefield, a situation that
didn't improve with time.

2. The work was being done on a 32bit SPARC architecture that was rapidly
being abandoned in favor of the 64bit UltraSPARC by Sun.

3. Assuming everything went perfectly, at least problems 2&3 from the
CLIPPER/NT port still existed and arguably problem 4 was still an issue,
particularly for a 32bit SPARC platform.

4. By the time anything started to happen on this, Intergraph Computer
Systems was aggressively pursuing the idea of x86 based workstation
products.

With both Sun and Intergraph lukewarm about the whole thing, it came as no
great surprise when Sun canned the project before it even got as far as the
CLIPPER/NT platform.

The source for a CLIX/x86 or CLIPPER/NT port would have exactly the same
copyright ownership problems as the CLIPPER/CLIX port.

Unfortuantely, it wouldn't! Because even if it were something that could be
released, the hardware architecture of the target for the CLIPPER/NT port
was radically different to any CLIX/CLIPPER platform. About the only thing
they had in common was the processor, the bus structure and everything else
was totally different.

--
Nik Simpson