[MPI3 Fortran] Teleconference Wednesday, Sept. 17: Summary

[Craig Rasmussen]

On Sep 17, 2008, at 10:25 AM, Jeff Squyres wrote:

> On Sep 17, 2008, at 12:17 PM, Supalov, Alexander wrote:
>
>> Thanks. I mentioned MPI_STATUS as a candidate for optional argument
>> status at the telecon as well. However, beyond that, one should  
>> probably
>> be rather conservative: it might be possible to imagine realistic
>> situations, in which an accidental erroneous call (with some omitted
>> arguments) might be treated as a valid one for a shorter set of
>> arguments. This may lead to very tough errors.
>
>
> With the strong type checking in F03 and distinct types for MPI  
> handles, I think the chance of this is pretty minimal.  I think we  
> do need to be mindful of this scenario, but I think we can design  
> interfaces such that common error cases like this shouldn't happen.
>

I think Alexander is correct in that the use of multiple optional  
arguments can be tricky.  If we go with your idea of overloading  
(carefully), you can make a middle argument optional without making  
all of the trailing arguments optional too.

Craig