Return-Path: <John.Harrison-request@cl.cam.ac.uk>
Delivery-Date: 
Received: from antares.mcs.anl.gov (no rfc931) by swan.cl.cam.ac.uk 
          with SMTP (PP-6.5) outside ac.uk; Fri, 23 Apr 1993 18:06:37 +0100
Received: by antares.mcs.anl.gov id AA21535 (5.65c/IDA-1.4.4 for qed-outgoing);
          Fri, 23 Apr 1993 11:11:36 -0500
Received: from netcom4.netcom.com by antares.mcs.anl.gov with SMTP 
          id AA21527 (5.65c/IDA-1.4.4 for <qed@mcs.anl.gov>);
          Fri, 23 Apr 1993 11:11:33 -0500
Received: by netcom4.netcom.com (5.65/SMI-4.1/Netcom) id AA05976;
          Fri, 23 Apr 93 09:11:46 -0700
Date: Fri, 23 Apr 93 09:11:46 -0700
From: val@netcom.com (Dewey Val Schorre)
Message-Id: <9304231611.AA05976@netcom4.netcom.com>
To: qed@mcs.anl.gov
Subject: Re:Machine Math
Sender: qed-owner@mcs.anl.gov
Precedence: bulk


Back in the old days, nobody specified languages, they wrote 
compilers. It had to be that way because nobody knew whether the 
language constructs contained all the information needed to compile 
into machine language, unless they actually had a compiler that did 
it. 

We are in this same situation today. We might design a RMachine MathS 
language that was perfectly acceptable to mathematicians, and later 
find that it lacked information needed for translation to a lower 
level.

The design of the language that mathematicians will use, and its 
translation to the lower level language needs to be an iterative 
process.

Val