Path: utzoo!attcan!uunet!lll-winken!lll-tis!helios.ee.lbl.gov!pasteur!ucbvax!decwrl!labrea!sri-unix!quintus!ok
From: ok@quintus.uucp (Richard A. O'Keefe)
Newsgroups: comp.lang.prolog
Subject: Re: A question on semantics and program transformation
Message-ID: <380@quintus.UUCP>
Date: 12 Sep 88 01:02:54 GMT
References: <26045@ucbvax.BERKELEY.EDU>
Sender: news@quintus.UUCP
Reply-To: ok@quintus.UUCP (Richard A. O'Keefe)
Organization: Quintus Computer Systems, Inc.
Lines: 15

In article <26045@ucbvax.BERKELEY.EDU> vanroy@ji.Berkeley.EDU (Peter Van Roy) writes:
>Is it allowable for a program transformation to change the
>procedural semantics of a program while retaining the
>declarative semantics?

It is if and only if the person using the program transformation says
it is.  For example, in NU Prolog, the ":- pure" declaration is a
statement by the programmer that s/he only cares about the declarative
semantics.  A "recursion editor" such as Alan Bundy proposes (or even
my little unfolding toy) would (does) work solely at the declarative
level, so would be (is) allowed to do anything that preserves declarative
semantics.  But an optimising compiler would not be allowed to change the
operational semantics (in the absence of :- pure declarations).

This is a problem in other languages as well.  (For example, "lazy"
evaluation is *very* sensitive to the syntactic form of programs.)