Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10 5/3/83; site houxa.UUCP
Path: utzoo!linus!philabs!cmcl2!floyd!vax135!ariel!houti!hogpc!houxm!houxa!9212osd
From: 9212osd@houxa.UUCP
Newsgroups: net.lang.c
Subject: comparing structures, complex data type
Message-ID: <254@houxa.UUCP>
Date: Mon, 26-Sep-83 13:51:00 EDT
Article-I.D.: houxa.254
Posted: Mon Sep 26 13:51:00 1983
Date-Received: Wed, 28-Sep-83 04:08:50 EDT
Organization: Bell Labs, Holmdel NJ
Lines: 33

A discussion on whether the C compiler should
support the comparison of structures has been going on
in this group recently.  Some said that there were not many
applications where this would pay off.  Let me give you an
example.

Consider the absence of the complex data type in C.
Most programmers will do this:

	typedef struct { double re, im } complex;

To compare two complex numbers defined as above is not
straightforward and comparison of structures in general,
would help.

But, to add, multiply, divide, etc., etc., complex numbers in C
is a pain anyway since you have to define macros (or functions)
to do the operations.*  The resulting code is hard to read,
especially to the scientist or engineer who is used to the
FORTRAN or PL/I language.  Consequently, the C Programming Language
is hard to sell to the hard-core scientist or engineer who just
wants to be reassured that a*b is a*b, no matter whether a or b
are complex or not. 

Do I hear any echos?
_
* Sure you could also write yappfcn. The one I'm aware of isn't very
clean, anyway.
_
-- 
Orlando Sotomayor-Diaz /AT&T Bell Laboratories, Crawfords Corner Road
Room HO-3M-325	201-949-1532	Holmdel, New Jersey, 07733
Path: {{{ucbvax,decvax}!}{ihnp4,harpo}!}houxa!9212osd