Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1 6/24/83; site mit-eddie.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!allegra!mit-eddie!smh
From: smh@mit-eddie.UUCP (Steven M. Haflich)
Newsgroups: net.arch
Subject: Re: DEC RISC Development
Message-ID: <3271@mit-eddie.UUCP>
Date: Fri, 7-Dec-84 08:07:56 EST
Article-I.D.: mit-eddi.3271
Posted: Fri Dec  7 08:07:56 1984
Date-Received: Sat, 8-Dec-84 05:25:37 EST
References: <8@osu-eddie.UUCP>
Reply-To: smh@mit-eddie.UUCP (Steven M. Haflich)
Distribution: net
Organization: MIT, Cambridge, MA
Lines: 21

Indeed, a friend working on the West Coast had a job interview with a
DEC group developing a RISC machine.  I have no further details.  It
can't be all that much of a secret if DEC tells randoms what they are
doing during job interviews, but the project certainly does seem to have
a low profile.  I would guess that the effort is a research project and
probably is not immediately targeted as a product.

This raises interesting questions about DEC's product line of `real'
machines -- that is, ignoring the various PCs and the aging smallness of
the PDP11.  DEC's recent product strategy is to market any size, shape,
color, style, and price of computer that the customer might want, just
so long as it's object-compatible with a VAX and will run VMS.  But what
happens if after a few more years it becomes apparent that the `right'
way to achieve superior peformance/price ratios executing high-level
languages is RISC rather than VAX-style CISC?  Then DEC will have to do
some fast and furious back pedalling, in software as well as hardware.

Of course, IBM has been successfully managing to preserve some measure
of up, down, and sideways compatibility with the original 360
instruction set for almost twenty years now, and that instruction set is
hardly a model of how to do things in the '80s.