variable length arrays ... please do the right thing

Peter Shenkin uunet!still3.chem.columbia.edu!shenkin
Tue Feb 9 15:26:13 PST 1993


> From: uunet.UU.NET!uunet!hal.com!gum (John Gmuender)
> Gentlemen,
> Please do the right thing and ...
> Please do not require C compiler writers to satisfy use before def.

Please don't put the convenience of the compiler writer above that
of the user.

We've all heard of "syntactic sugar";  those who insist that
that an exception should not be made to this rule even in a function 
prototype are introducing "syntactic molasses" for the user.

It's true that if implementers find it too difficult to do it won't
be done, perhaps no matter who mandates it;  however, language USE
is discouraged by features which clutter up the syntax for the sole 
purpose of making the language easier to compile.  (I can live a 
little more easily with constructs that make the language easier to 
optimize.)

As a C user, I'd be happiest with something like the Cray VLA proposal, 
which (1) places no restriction on order of function parameters, and
(2) requires no special syntax within any of:

	the function call
	the function definition
	the function prototype.

If implementers outside of Cray really do find this too onerous, my
second choice would be one of the proposals that still preserves (1),
while introducing some special syntax for the function definition and
prototype, but not the call.

	-P.
************************f*u*cn*rd*ths*u*cn*gt*a*gd*jb************************
Peter S. Shenkin, Box 768 Havemeyer Hall, Dept. of Chemistry, Columbia Univ.,
New York, NY  10027;  shenkinastill3.chem.columbia.edu;  (212) 854-5143
***************** Watch this space for an important message ******************




More information about the Numeric-interest mailing list