My understanding is that Dave Ungar did the SOAR work
under Dr. David Patterson at Berkeley.  I also understand that
a couple of years later, Sun Microsystems came to Dr. Patterson
and told him, "The 68040 has reached the end of its lifetime.
Can you recommend a new processor architecture to us?  A RISC,
maybe?"

	The results of the SOAR study showed that remarkably
little had to be implemented in hardware to give good Smalltalk
performance, as long as the processor had about a zillion registers
to use for recursive subroutine calls.  SOAR had that in spades.
And because they'd only added about one Smalltalk-specific thing
to the hardware architecture, which had turned out not to help
all that much, SOAR became the basis for the SPARC.

	L. Peter Deutsch knew this when he wrote the "just-in-time"
native code method compilation stuff for ParcPlace Smalltalk.  He
was delighted with the performance he was able to achieve.

	So, if y'all want hardware that was designed to run
Smalltalk well, y'all have it.  It's called SPARC.  Or so I'm told.

Mike O'Brien