From info-hol-request@ted.cs.uidaho.edu Thu Aug 13 12:07:48 1992
Received: from brwbf.inmos.co.uk by frogland.inmos.co.uk; Thu, 13 Aug 92 12:07:48 BST
Received: from inmos-c.inmos.com by brwbf.inmos.co.uk with DNI-MTP [1.1]
        id inmos-c-2703; Thu, 13 Aug 92 12:07:49 BST
Received: from ted.cs.uidaho.edu by relay2.UU.NET with SMTP
        (5.61/UUNET-internet-primary) id AA15332; Thu, 13 Aug 92 05:52:51 -0400
Received: by ted.cs.uidaho.edu (16.6/1.34)
        id AA28670; Thu, 13 Aug 92 02:42:04 -0700
Sender: info-hol-request@ted.cs.uidaho.edu
Errors-To: info-hol-request@ted.cs.uidaho.edu
Precedence: bulk
Received: from swan.cl.cam.ac.uk by ted.cs.uidaho.edu (16.6/1.34)
        id AA28665; Thu, 13 Aug 92 02:40:49 -0700
Received: from dipper.cl.cam.ac.uk by swan.cl.cam.ac.uk
          with SMTP (PP-6.0) to cl id <25216-0@swan.cl.cam.ac.uk>;
          Thu, 13 Aug 1992 10:37:43 +0100
Received: by dipper.cl.cam.ac.uk (4.1/SMI-3.0DEV3) id AA01376;
          Thu, 13 Aug 92 10:37:12 BST
Date: Thu, 13 Aug 92 10:37:12 BST
From: Mike.Gordon@cl.cam.ac.uk
Message-Id: <9208130937.AA01376@dipper.cl.cam.ac.uk>
To: Wai.Wong@cl.cam.ac.uk
Cc: info-hol@ted.cs.uidaho.edu
In-Reply-To: Wai Wong's message of Thu, 13 Aug 92 10:05:08 +0100 <"swan.cl.ca.334:13.07.92.09.06.00"@cl.cam.ac.uk>
Subject: Conditional rewriting


                        Historical Note
                        ===============

The HOL rewriting engine, which was written by Larry Paulson for LCF,
originally did have conditional rewriting. When I was porting his code
from LCF to HOL I unwisely jettisoned the conditional parts.

There were two reasons for this; a good one and a bad one: the bad one
is that I had some problems modifying the code to work on terms rather
than LCF formulae. I was impatient, so I gave up. The good one was
that I anticipated that conditional rewriting would be much less
important for HOL than for LCF, since one no longer had to reason
about "bottom" all the time. Conditional rewriting was essential in
LCF to handle things like:

   ~(y = UU) ==> (HD(CONS(x y)) == x)

Due to recent performance enhancing changes in rewriting, it would
probably not be sensible to try to resurrect the original code. A new
implementation would be best.

Mike Gordon
