Path: utzoo!mnetor!uunet!mcvax!cernvax!ethz!marti
From: marti@ethz.UUCP (Robert Marti)
Newsgroups: comp.lang.c++
Subject: Re: Operator overloading considered harmful
Message-ID: <281@bernina.UUCP>
Date: 13 Jan 88 10:09:08 GMT
References: <240@vsi1.UUCP> <4538@watdragon.waterloo.edu>
Organization: ETH Zuerich, Switzerland
Lines: 33
Summary: *function* overloading considered just as harmful ...

In article <4538@watdragon.waterloo.edu>, gvcormack@watdragon.waterloo.edu (Gordon V. Cormack) writes:
> In article <240@vsi1.UUCP>, steve@vsi1.UUCP (Steve Maurer) writes:
>    (as an indictment of operator overloading).
>
> >     foo(int i)  { i #= i ^^ i @& i; }
> >
> > what does it do??    You will never know unless you start digging
> > through miles of include/lib files.
>
> Would it be clearer as
>
>     foo(int i)  { glorp_assign(i, fred(xyz_arrows(i,i),nerf(i))); }

Disregarding the fact that Maurer's example is not legal in C++, both
examples suffer from the same problems if (operator or function)
overloading is used:
1. As mentioned in Maurer's original posting, you may have to search
   your source files for the appropriate operator or function definitions
   for quite a while.
2. Overloading may unintentionally defeat type checking.

Personally, I think the (more restricted) form of overloading you get
by subclassing and inheritance is all you really need.  The argument
for operator overloading -- namely that it is nice to use an operator
such as + for adding complex numbers, vectors, etc -- is just not
good enough for me.  After all, this is just syntactic sugaring!


-- 
Robert Marti                    Phone:       +41 1 256 52 36
Institut fur Informatik
ETH Zentrum/SOT                 CSNET/ARPA:  marti%ifi.ethz.ch@relay.cs.net
CH-8092 Zurich, Switzerland     UUCP:        ...uunet!mcvax!ethz!marti