Version 1.1 of acs/ac-00113.txt

Unformatted version of acs/ac-00113.txt version 1.1
Other versions for file acs/ac-00113.txt

!standard 6.7(00)          05-10-20 AC95-00113/01
!class Amendment 05-10-20
!status received no action 05-10-20
!status received 05-05-02
!subject Null procedures as completions
!summary
!appendix

!topic Null procedures as completions
!reference AI95-00348
!from Adam Beneschan 05-05-02
!discussion

Is there a reason why a null_procedure_declaration shouldn't be
allowed as the completion of a procedure_specification?  It's not a
big huge deal, since you could always say

    procedure blah_blah_blah is begin null; end;

without much more text than if you just said "is null;".  But
someone's going to try using "is null;" as a procedure body, and find
it's illegal, and wonder why.

****************************************************************

From: Randy Brukardt
Date: Monday, May  2, 2005  9:20 PM

My guess would be one of two reasons (a) people didn't think to do this; or
(b) people did think of this, but it seemed like a lot of implementation
work for a benefit of exactly zero (OK, it would save 11 characters, that's
as close to zero as it is possible to get). Probably some ARG members fall
into each group.

The model behind null procedures is that they are similar to abstract
subprograms, and certainly you wouldn't want one of those to be a
completion. If we make them too different from abstract subprograms, they
become a third kind of thing that has to be handled everywhere -- in which
case I have to wonder if they are worth it at all.

****************************************************************

Questions? Ask the ACAA Technical Agent