Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!apple!brutus.cs.uiuc.edu!tut.cis.ohio-state.edu!att!dptg!ulysses!andante!alice!shopiro
From: shopiro@alice.UUCP (Jonathan Shopiro)
Newsgroups: comp.lang.c++
Subject: Re: New Begginer Questions.
Summary: Answers
Message-ID: <9874@alice.UUCP>
Date: 10 Sep 89 01:07:28 GMT
References: <4202@ohstpy.mps.ohio-state.edu>
Organization: AT&T Bell Laboratories, Murray Hill NJ
Lines: 43

In article <4202@ohstpy.mps.ohio-state.edu>, SMITHJ@ohstpy.mps.ohio-state.edu writes:
> I'm following this conversation with interest as I am a somewhat experienced
> C programmer looking to learn C++.
> 
> Two questions:
> 1. Is the declaration type* var; common in C++ programs as opposed to type *var
>    and if so why.  I find that it misleads one to beleave that the declaration
> 			type* x, y;
>    would be interpreted as
> 			type *x, *y;
>    rather than the correct
> 			type *x, y;
C++ emphasizes the type system.  In C, as in C++, int* is a type, but we
think it more important to say that the variable pi is an int*, as in

	int*	pi;

than to say that the expression *pi is an int, as in

	int	*pi;

Of course the syntax we inherit from C makes multiple declarations
treacherous, as you note, so we generally only declare one variable
at a time.

> 2. Why the addition of the typecast convention type(var) when (type)var 
>    already exists.
We have extended the notion of casting to allow constructing a new value
from arguments.  For example, you might define a type Complex to represent
complex numbers, and allow construction of a Complex from one or two
floats.  You might then write

	float	p, q;  // :-)
	
	(Complex)p;	// cast using old syntax

but you have to use the new syntax for

	Complex(p, q);
-- 
		Jonathan Shopiro
		AT&T Bell Laboratories, Warren, NJ  07060-0908
		research!shopiro   (201) 580-4229