Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: Notesfiles $Revision: 1.6.2.16 $; site ima.UUCP
Path: utzoo!linus!philabs!cmcl2!harvard!bbnccv!ima!compilers
From: compilers@ima.UUCP
Newsgroups: mod.compilers
Subject: denotational semantics and compilers
Message-ID: <136300038@ima.UUCP>
Date: Thu, 9-Jan-86 14:40:00 EST
Article-I.D.: ima.136300038
Posted: Thu Jan  9 14:40:00 1986
Date-Received: Sat, 11-Jan-86 06:12:05 EST
Lines: 18
Approved: compilers@ima.uucp
Nf-ID: #N:ima:136300038:000:759
Nf-From: ima!compilers    Jan  9 14:40:00 1986


[from ihnp4!philabs!sbcs!debray        (Saumya Debray)]

> When I read a denotational semantics, I think of it as an
> operational semantics written in Scheme ...

One big difference between denotational and opeerational semantics is
that the former lets us talk about nonterminating computations in a
useful way.  All nonterminating computations are treated as being
equivalent to "bottom", which essentially denotes "having zero
information content"; this isn't possible in operational semantics.
This equivalence forms the basis for strictness analysis, for example,
which enables us to replace call-by-name parameter passing by an
"equivalent" (but much more efficient) call-by-value parameter passing
at compile time.

-Saumya Debray
SUNY at Stony Brook