(j3.2006) (SC22WG5.4522) Second WG5 ballot on interpretations

[Van Snyder]

On Fri, 2011-09-23 at 12:22 -0700, Bill Long wrote:
> F08/0038:
> ....
> Also, the COUNT intrinsic [338:31] seems to be missing from the list. 
> Seems like an oversight.

It wasn't an oversight.  Because the KIND argument, after the DIM
argument, is also an integer, COUNT can't be handled in the same way as
the others.  There is no way to resolve the following pair of specific
procedures with the same generic name:

COUNT ( MASK, DIM [, KIND ] )
COUNT ( MASK [, KIND] )

It violates C1215.

In a reference "COUNT ( Logical_Array, kind(0) )" is kind(0) the DIM or
KIND actual argument?

The others don't have this problem.  For example, in FINDLOC, the MASK
argument, immediately after the DIM argument, is logical.  So the pair
of generics

FINDLOC ( ARRRAY, VALUE, DIM [, MASK, KIND, BACK ])
FINDLOC ( ARRRAY, VALUE [, MASK, KIND, BACK ])

can be resolved.