Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1 6/24/83; site unc.UUCP
Path: utzoo!linus!decvax!duke!mcnc!tim@unc.UUCP (Tim Maroney)
From: tim@unc.UUCP
Newsgroups: net.lang
Subject: Re: Strong Typing and Ignorance
Message-ID: <6421@unc.UUCP>
Date: Fri, 16-Dec-83 11:49:55 EST
Article-I.D.: unc.6421
Posted: Fri Dec 16 11:49:55 1983
Date-Received: Tue, 27-Dec-83 03:12:10 EST
References: <2892@utcsrgv.UUCP>, <137@csd1.UUCP> <718@cbosgd.UUCP>, <217@kobold.UUCP>
Organization: CS Dept., U. of N. Carolina at Chapel Hill
Lines: 30

I was taught that the assertion "language X is strongly typed" is equivalent
to the assertion "if E is a legal expression in language X, then the type of
the value of E can be determined at compile time".  This is the standard
meaning of "strong typing" -- if someone knows of a well-defined contender
for the phrase, by all means post it to the net, but until that time I will
assume that this is the only valid meaning.

Contrary to the assertions of some recent articles, Pascal is not strongly
typed.  There are two exceptions.  The first concerns variant records, which
have a type which must be determined at run time by examination of the tag
field.  The second involves integer range types; it is legal to assign a
normal integer variable to an integer range variable, but it cannot be
determined at compile time whether the integer variable is of the integer
range type or the integer type: if the latter, the assignment is illegal.

C is not strongly typed, but probably not for the reasons you'd expect
(argument mismatch and pointer typing escapes do not make the expression
type indeterminate at compile time, they just complicate the typing
semantics).  The "?:" operator, which is a distributed conditional, can have
differently typed alternative values.  For instance, in the expression
"i>0 ? 1 : 4.2", the type of the value of the expression cannot be
determined until the value of i is known -- it might be int, or double.  I
don't know of any other exceptions to strong typing in C.

I hope this definition will help to counteract some of the wildly-inaccurate
articles on the subject that have appeared lately, and satisfy the curiosity
of those who wanted to know.
--
Tim Maroney, University of North Carolina at Chapel Hill
duke!unc!tim (USENET), tim.unc@csnet-relay (ARPA)