[1091] in java-interest
Re: Operator underloading (was: overloading of operators)
daemon@ATHENA.MIT.EDU (Kevin Krom)
Fri Aug 18 13:50:25 1995
Date: Fri, 18 Aug 1995 10:23:06 +0500
From: krom@cgi.com (Kevin Krom)
To: java-interest@java.sun.com
Cc: Philip.Gladstone@onsett.com
Philip Gladstone writes:
> I propose that we do some operator-underloading. The + sign is already used
> for integers and floating point. This leads to endless confusion, so I
> propose that + is reserved only for integers, and that a new operator
> ("floatAdd" - if you like smalltalk syntax) or a class member function be
> defined to add floats.
>
> This approach has some useful advantages - the comparison operator <
> would be replaced by "floatLessThan", and we could eliminate the
> floating point <= operator entirely (as floating point numbers are
> very rarely equal to other floating point numbers).
>
> The readability of the language would improve - you wouldn't have to
> worry about whether 10/4 is actually 2 or 2.5
I hope I am not wrong in taking this as tongue-in-cheek...
Why not use and extend the Pascal notation? Pascal already has "div" and
"mod" as integer versions of the "/" and "%" operators.
*** SARCHASM MODE ON ***
But to the original point, if we're going to start "underloading" operators,
then why not just get rid of them altogether? Why make one type have a
different look and feel than the others. Then we can get simple expressions:
int a, b, c, d;
float e, f;
f = ((float) (a.Add(b)).Multiply(e.Divide((float) (c.Add(d))))
I think that the above version is much more readable than this ambiguous and
confusing form:
f = (a + b) * (e / (c + d));
*** SARCHASM MODE OFF ***
Flame away...
--
Kevin M. Krom, Software Engineer Phone: (412) 642-6900 x242
Carnegie Group, Inc. Fax: (412) 642-6906
5 PPG Place, Pittsburgh PA 15222 Email: krom@cgi.com
-
Note to Sun employees: this is an EXTERNAL mailing list!
Info: send 'help' to java-interest-request@java.sun.com
