Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84; site peora.UUCP
Path: utzoo!watmath!clyde!cbosgd!ihnp4!mhuxn!mhuxr!mhuxt!houxm!hjuxa!petsd!peora!jer
From: jer@peora.UUCP (J. Eric Roskos)
Newsgroups: net.arch
Subject: Re: Integer Division; interesting approach!
Message-ID: <1973@peora.UUCP>
Date: Tue, 18-Feb-86 08:46:57 EST
Article-I.D.: peora.1973
Posted: Tue Feb 18 08:46:57 1986
Date-Received: Wed, 19-Feb-86 03:50:26 EST
References: <9559@amdcad.UUCP>
Organization: Concurrent Computer Corporation, Orlando, Fl
Lines: 46
Keywords: Divide Shift Rounding

> Comparing complex instructions to hieroglyphics is what I like.

This is an interesting analogy. (I, also, like comparing things from
different but similar domains in order to find similarities, in the hope
that they have something common underlying them.)

But are CISC instructions comparable to heiroglyphics?  Maybe so.  The
heiroglyphics do express "higher-level" ideas, just like CISC instructions.

On the other hand, how do Heiroglyphics compare to a phonetic alphabet?
The phonetic alphabet conveys *no* information at all other than sounds.
Thus, for example, Spanish and English both use the same alphabet, with
a few exceptions, yet you can be the world's expert on English and not
be able to figure out anything at all about a Spanish sentence.  This
is because, really, a *word* in English or Spanish is the semantic unit;
if you break a sentence into words, each word still means something, but
if you break a word down into letters, it doesn't mean anything any more.
Each letter is just a sound, and actually *that* isn't even true, since
the sound of the letters actually depends on what word they're used in,
so the letters are somewhat context-sensitive.

So really you'd have to compare the number of different heiroglyphs to the
number of different words in the English language, applying some measure
of "expressive power", to compare the languages.

On the other hand, RISC and CISC instructions are both meaningful units
of their own.  Comparing the two is more like comparing two types of
heiroglyphics to see which is more expressive.

The strange irony about this whole RISC/CISC debate is something analogous
to the above.  If you read the papers that have been coming out of what
has recently been labeled the CISC community for years, you find theory
that is quite comparable to the theory advanced by RISC advocates -- they
don't always agree, but at least they provide some solid means of
analyzing instruction sets.  Read, for example, M. J. Flynn's "Computer
Organization and Architecture" in _Operating_Systems:_An_Advanced_Course_
(Springer-Verlag; ISBN 0-387-09812-7).  The RISC/CISC dualism is
essentially artificial, I think, because really there is a continuum of
instruction set "complexity".

Thus I don't think the comparison to heiroglyphics is that useful a
comparison.
-- 
UUCP: Ofc:  jer@peora.UUCP  Home: jer@jerpc.CCUR.UUCP  CCUR DNS: peora, pesnta
  US Mail:  MS 795; CONCURRENT Computer Corp. SDC; (A Perkin-Elmer Company)
	    2486 Sand Lake Road, Orlando, FL 32809-7642