Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!mips!spool.mu.edu!uwm.edu!bionet!parc!boehm
From: boehm@parc.xerox.com (Hans Boehm)
Newsgroups: comp.object
Subject: Re: A Hard Problem for Static Type Systems
Message-ID: <boehm.672515148@siria>
Date: 24 Apr 91 17:45:48 GMT
References: <1991Apr20.010347.28984@leland.Stanford.EDU> <DUCHIER.91Apr23205041@albania.cs.yale.edu> <1991Apr24.032543.11704@leland.Stanford.EDU>
Sender: news@parc.xerox.com
Organization: Xerox PARC
Lines: 25

It seems to me that this problem fundamentally has very little to do with object
oriented programming.  Expressing things in terms of a subtype hierarchy
confuses the issue and contributes nothing.  In Russell, the min function
has signature:

func [x,y: val T; T: type { < }] val T

It's a function of two values of type T, and of the type T itself.  T is
expected to include a < operation.  (I get to cheat slightly, in that
operations like < have the "right" type by default.)  The third (type)
parameter will be inferred for each call, and is not explicitly needed,
except in the declaration of min.

It would be nice if I could annotate "<" with some properties (e.g. the fact
that it should be transitive).  Languages like IBM's Scratchpad have
mechanisms for doing that.

Cardelli's Quest language combines this kind of mechanism with inheritance.
(Quest and Russell differ in other profound respects.  But those have nothing
to do with this discussion.)

Hans
(boehm@xerox.com)

Usual disclaimers ...