[MPI3 Fortran] Agenda for MPI3 Fortran Working group next week

[Jeff Squyres]

On Jun 4, 2009, at 11:47 AM, Iain Bason wrote:

> I think Alexander's point is that it is desirable for a particular
> implementation to maintain backwards binary compatibility with
> previous versions of that implementation.  If some implementation
> defined a communicator as a derived type that contained all the
> internal state the library needs, then that implementation would be
> unable to change that derived type without breaking backwards
> compatibility.
>

Ah, ok -- fair enough.  To me, that's "backwards compatibility", not  
"ABI".  FWIW: "ABI" means a very different thing to me.

(don't get me wrong, although most people thing I'm against an ABI for  
MPI, I'm not -- but that's a different discussion for a different list/ 
group)

> However, I think implementors are smart enough to realize that, and to
> define a communicator as a pointer to some opaque type.
>

Agreed.

-- 
Jeff Squyres
Cisco Systems