Variable-length arrays

James Frankel uunet!Think.COM!frankel
Thu Dec 19 10:48:20 PST 1991


   From: tamafig.cray.com (Tom MacDonald)
   Date: Thu, 19 Dec 91 12:18:37 CST

   >>>>>> Peter S. Shenkin
   >     I agree.  Certainly the whole point is to make life easier for the 
   >     programmer, isn't it?

   >>>>>> Richard Stallman
   > You sure feel free to spend my time, don't you?

   I'm much more sensitive to the needs of customers.  I want the feature to
   feel comfortable and appeal to my Fortran user base.  If it feels kludging
   to the compiler developer but not to the C programmer then my vote goes
   to the end user - always.

I'm also concerned about making language features meet the needs of
customers; however, I also feel that a simple set of identifier binding
and scoping rules is to the customer's/programmer's benefit.  I don't
see the overriding need to be able to declare the array before the array
length in the parameter list.  I'm more than happy to support the syntax
which allows:

	void f(int n, double a[n]);

but does not allow:

	void f(double a[n], int n);

I don't believe that programmers will use unprototyped functions just
because they might allow the parameters in the other order, as follows:

	void f(a, n)	/* I don't believe that this construct will */
	int n;		/*   be used by good programmers */
	double a[n];
	{ /* body */ }

Jamie Frankel



More information about the Numeric-interest mailing list