Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!swrinde!cs.utexas.edu!ut-emx!nowhere!sking
From: sking@nowhere.uucp (Steven King)
Newsgroups: comp.lang.c++
Subject: Re: Matrix Class Libraries
Message-ID: <1991Mar19.161203.14680@nowhere.uucp>
Date: 19 Mar 91 16:12:03 GMT
References: <1991Mar18.025539.18998@athena.mit.edu> <ARI.HUTTUNEN.91Mar18173446@hydroman.hut.fi> <1991Mar18.213545.20966@athena.mit.edu>
Organization: American Anarchist Union
Lines: 60

In article <1991Mar18.213545.20966@athena.mit.edu> bjaspan@athena.mit.edu (Barr3y Jaspan) writes:
>In article <ARI.HUTTUNEN.91Mar18173446@hydroman.hut.fi>, Ari.Huttunen@hut.fi (Ari Juhani Huttunen) writes:
>|> I think this points out one problem in C++. You really should be able to
>|> say:
>|>       inline double& operator[][](int, int) const;
>|> 
>|> The problem isn't very serious, but it just isn't natural to access array
>|> members using notation  array(x,y), but array[x][y].
>
>At the risk of starting a religous war, I'll respond.  :-)
>
>In this particular case, the first method is preferable.  Being able to say
>array[x][y] implies (in the conventional C style) that array[x] is meaningful
>(it always is in C).  For my Matrix class, it isn't meaningful (you could
>argue that I should have derived Matrix from Vector, in which case array[x]
>could be meaningful, but that isn't the point.)
>
>So if you really always wanted to use [] for array indexing and () only for
>function calls, [] would have to accept any number and type of arguments (ie:
>operator[](int, int).)  But then it would be semantically identical to ()
>anyway, so what's the point?
>
   Okay, at the risk of fanning the flames of discussion...

   why not:

   class Matrix {
      double *d ;
      int M_m, M_n ;
   public:
      class aux {
        double *vec ;
        aux ( double *v ) { vec = v ; }
        friend class Matrix ;
      public:
        double& operator [] ( unsigned n ) { return vec [ n ] ; }
      } ;
      aux operator [] ( unsigned n ) { return aux ( &d [ n * M_m ] ) ; }
   } ;

   Obviously, various bounds checking etc., should be included. This then
 does allow for statements of the form:

   Matrix m ;
   m [j][i] = 1.0 ;

   Of course it does create an extra temporary and and a copy assignment
 of the matrix aux, but it does give that syntactical sugaring...


>In any case, I agree that it isn't a serious problem.
>
>-- 
>Barr3y Jaspan, bjaspan@mit.edu
>Watchmaker Computing


-- 
Look Ma! No .sig!
                                        ..!cs.utexas.edu!ut-emx!nowhere!sking