Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1 6/24/83; site pur-ee.UUCP
Path: utzoo!linus!philabs!seismo!harpo!eagle!mhuxl!ihnp4!inuxc!pur-ee!iuvax!cjl
From: cjl@iuvax.UUCP
Newsgroups: net.lang
Subject: Re: Strong Typing and Ignorance - (nf)
Message-ID: <1290@pur-ee.UUCP>
Date: Tue, 3-Jan-84 03:38:24 EST
Article-I.D.: pur-ee.1290
Posted: Tue Jan  3 03:38:24 1984
Date-Received: Fri, 6-Jan-84 01:29:19 EST
Sender: notes@pur-ee.UUCP
Organization: Electrical Engineering Department , Purdue University
Lines: 27

#R:utcsrgv:-289200:iuvax:11800007:000:1122
iuvax!cjl    Jan  2 22:37:00 1984


I am against Tim Moroney's definition on "strong typing" which can also
be found in p.38 od Horowitz's "Fundamentals of Programming Languages".
Expressions consist of operators and operands which are just syntactic 
sugar of functions. It is neither consistent nor practical to leave out
procedure argument mismatch checking in a strongly typed language.

Most recent textbooks define a strongly typed language to allow complete,
statical type checking, e.g. p90 in Ghezzi's "Prog. Lang. Concepts",
p129 in MacLennan's "Principles of Programming Languages",
p184 in Ledgard's "The Programming Language Landscape".
The ambiguity lies on the COMPLETENESS.  Generally it should exclude
(1) type-ambiguity of expression which can be found in 
    (a) conditional expression in C,
    (b) record variant in Pascal,(In ISO Pascal, subrange is treated
        as range constraint not a new type) 
    (c) COMMON and EQUIVALENT in FORTRAN,
(2) untyped pointers (in PL/I) and dangerous pointer operation (in C),
(3) assignment incompatibility (which includes procedure parameter checking).

C.J.L. , IUPUI
cjl@Indiana@Rand-Relay