Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!sdd.hp.com!spool.mu.edu!munnari.oz.au!goanna!ok
From: ok@goanna.cs.rmit.oz.au (Richard A. O'Keefe)
Newsgroups: comp.lang.scheme
Subject: Re: ML-like type-checker for Scheme subset?
Message-ID: <5808@goanna.cs.rmit.oz.au>
Date: 17 May 91 09:36:16 GMT
Article-I.D.: goanna.5808
References: <5710@goanna.cs.rmit.oz.au> <1065@sys.uea.ac.uk>
Organization: Comp Sci, RMIT, Melbourne, Australia
Lines: 22

In article <1065@sys.uea.ac.uk>, jrk@sys.uea.ac.uk (Richard Kennaway) writes:
> In <5710@goanna.cs.rmit.oz.au> ok@goanna.cs.rmit.oz.au (Richard A. O'Keefe) writes:
> >    -- the idea that programming language implementations are entitled
> >       to quietly deliver incorrect results (Pascal implementations
> >       are often like this, and C implementations almost always are)
> 
> What are you referring to here?

In most Fortran, Pascal, C, ... implementations I have used,
you can add two positive numbers and get a negative number.
If you read the Fortran, Pascal, and C standards with sufficient
care and attention, you will see that quietly delivering the
wrong answers to integer arithmetic operations like this is _legal_.
Implementations are _also_ free to report overflows (nice one, MIPS!)
but often don't.  Indeed, some systems makes it hard to _notice_ the
overflow.  (For example, one Pascal compiler I used on a 68010
_did_ generate code to trap on overflows, but the support routines
for 32-bit multiplication and division forgot to set the V flag.)

I like rationals, but I *love* bignums.
-- 
There is no such thing as a balanced ecology; ecosystems are chaotic.