Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84; site wdl1.UUCP
Path: utzoo!watmath!clyde!cbosgd!ihnp4!zehntel!dual!amdcad!fortune!wdl1!jbn
From: jbn@wdl1.UUCP
Newsgroups: net.unix-wizards
Subject: Re: Re:  Optimization and language level
Message-ID: <193@wdl1.UUCP>
Date: Tue, 15-Jan-85 16:43:24 EST
Article-I.D.: wdl1.193
Posted: Tue Jan 15 16:43:24 1985
Date-Received: Sun, 20-Jan-85 06:21:20 EST
Sender: jrb@wdl1.UUCP
Organization: Ford Aerospace, Western Development Laboratories
Lines: 13
Nf-ID: #R:harvard:-22700:wdl1:17100045:000:624
Nf-From: wdl1!jbn    Dec 21 12:57:00 1984


     Exactly right.  The problem is language design.  I have written a
longish paper on this subject as it relates to C, and will send it on
to anyone interested.  (Much of it is a formal treatment of C arithmetic
and how it could be defined in a portable way).
     It is possible to design a language in which the semantics are
totally nailed down, but it is unusual to do so.  Such a language is
completely portable.  The UCSD P-system P-code can be thought of as
such a language.  But no programming language defined to date and widely
used meets this criterion; even Ada is defined sloppily in places.

					John Nagle