In early 1990s IBM had product plans for a PowerPC based computer designed to a spec called CHRP (Common Hardware Ref Platform). It was to be backed by a consortium of Taiwanese board makers.

In 1994 we had demo versions of the hardware boxes. Apple had agreeded to the CHRP platform and had a design using PowerPC chips, that was supposed to conform to CHRP (but Apple lost their nerve & backed out of the deal - in all fairness, history suggests they did the right thing - back then Apple were probably in the same boat Sun is today (sinking!) and had they proceeded, they may have sunk completely).

Anyway, the OS that was going to run on this was a microkernel (Mach 3.0) based software system called Workplace OS. The subsequent OS/2 based GUI called 'Workplace Shell' was to be the OS/2 personality. In addition there would be a Windows, & Posix(AIX/Unix) personality. Apple was supposed to be developing an Apple personality that also ran on the Workplace uKernel.

[link|http://www.byte.com/art/9401/sec8/art10.htm|http://www.byte.com/...01/sec8/art10.htm]

The PowerPC chip supported both big-endian & little-endian modes of operation on a page basis. The challenge was the I/O adapters & the i/o driver software. They never did get that part to work to acceptable levels of performance.

The above failures are highlighted in this link ... [link|http://www.eecs.harvard.edu/hotos/scribes/notes/implement-1-discussion.html|http://www.eecs.harv...1-discussion.html]

It was estimated by one person, that the failure of the Workplace OS / CHRP project cost 1 Billion $US (ouch).


Doug M
(who was to be the product manager in Australia, for the box)