(j3.2006) Fortran 201x wish: SELECT RANK

[Malcolm Cohen]

>If we're
>allowing the use of assumed-rank arrays in Fortran, we should consider the
>implications on the rest of the standard.  For example, we should consider
>how we want assumed-rank arrays to work with generic interfaces.

We already
(a) added assumed-rank to Fortran
(b) considered the interaction with generics
otherwise TS 29113 would be totally broken from the outset.

The only good reason for not adding SELECT RANK was that it was out of scope for 
the TS.

>  Also, if
>we're going in this direction, we should also think about whether we need
>a static casting construct for assumed-type entities.

That would be called (unchecked) "pointer assignment".

However, assumed-type is a poor cousin to CLASS(*) which has proper facilities 
for supporting access to the dynamic type, in fact TYPE(*) is merely a severely 
restricted subset of CLASS(*), precisely for communicating with C.

Cheers,
-- 
................................Malcolm Cohen, Nihon NAG, Tokyo.