Path: utzoo!utgpu!news-server.csri.toronto.edu!clyde.concordia.ca!uunet!mcsun!ukc!stl!tom@nw.stl.stc.co.uk
From: tom@nw.stl.stc.co.uk (Tom Thomson)
Newsgroups: comp.lang.functional
Subject: Re: What is strong typing? (Was: I like strong typing)
Message-ID: <3282@stl.stc.co.uk>
Date: 6 Aug 90 00:42:20 GMT
References: <4387@uqcspe.cs.uq.oz.au> <1689@opal.tubopal.UUCP> <4424@uqcspe.cs.uq.oz.au> <3263@stl.stc.co.uk> <5448@castle.ed.ac.uk>
Sender: news@stl.stc.co.uk
Reply-To: "Tom Thomson" <tom@stl.stc.co.uk>
Organization: ICL Manchester M12 5DR UK
Lines: 23

In article <5448@castle.ed.ac.uk> nick@lfcs.ed.ac.uk (Nick Rothwell) writes:
>0 is an integer, just like 17 is. Milner-style typechecking is structural,
>so it knows about, say, integers, but not about positive integers, primes,
>and non-structural entities of this kind.
>
I think this misses the point completely.  The reason for introducing
zero-division as an example was precisely to point that Milner-style
type checking doesn't cover this case - - in fact any type checking
system that insists on operating at compile time and not at run time
(while leaving it possible to write a reasonable range of programs)
will fail to cover this case.

Also, what does structural mean in this context?  In terms of structural
types, I can easily define positive integers as a type distinct from
natural numbers including zero.  I have never come across a functional
language that would stop me doing this.  

I think it might be useful to think about why we have type systems 
(to limit the damage when we apply a function to arguments outside its domain)
rather than what a particular type system (or class of type systems)
can do.

Tom Thomson   [tom@nw.stl.stc.co.uk