Xref: utzoo comp.lang.scheme:2506 comp.theory:1992 comp.lang.functional:785
Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!wuarchive!udel!haven.umd.edu!socrates.umd.edu!socrates!rockwell
From: rockwell@socrates.umd.edu (Raul Rockwell)
Newsgroups: comp.lang.scheme,comp.theory,comp.lang.functional
Subject: Re: ML-like type-checker for Scheme subset?
Message-ID: <ROCKWELL.91May15223913@socrates.umd.edu>
Date: 16 May 91 03:39:13 GMT
References: <5710@goanna.cs.rmit.oz.au> <7480@rex.cs.tulane.edu>
	<1991May13.164558.3703@news.cs.indiana.edu> <7504@rex.cs.tulane.edu>
Sender: rockwell@socrates.umd.edu (Raul Rockwell)
Organization: Traveller
Lines: 19
In-Reply-To: fs@caesar.cs.tulane.edu's message of 15 May 91 20: 14:15 GMT

Frank Silbermann:
   If types are domain names then a type-clash is mathematical
   _nonsense_.  Under this paradigm, a functional language can include
   _only_ well-typed programs.  Ill-typed programs have no meaning.
   This severely hampers the usefulness of runtime typechecking, since
   the program is invalid _even if_ the type clash occurs in an
   unevaluated subexpression.

Now I'm confused.

Simple example: x % (y - z) where % indicates division.  It seems to
me that (A) this program has meaning, and (B) there is the potential
for a type clash (y - z might be outside the domain of %).  It also
seems to me that this might be an unevaluated subexpression for
conditions where the type clash occurs.

Color me perplexed.

Raul Rockwell