Thanks for the explanation, Richard. I had read 14.1.2.3 carefully and
intentionally used the term "different types", quoting it for emphasis. I
was wondering if different interface blocks for a dummy procedure constituted
different types -- my mistake was lack of recognition of "type" with its
strong definition in Fortran 90. You make the differences between procedures
and variables here very clear. Thanks.
Implicit in my statements is the question "if Fortran doesn't allow you to
pass an actual argument procedure with a different interface than that
specified for the dummy argument, then why can't it allow the dummy argument
interface as a differentiator for generic interfaces?". Clearly the compilers
have the capability of checking this, since they do so for type mismatches.
Is this something that should be considered for a future version of Fortran?
-- Bill
William F. Mitchell | [log in to unmask]
Mathematical and Computational Sciences Division | [log in to unmask]
National Institute of Standards and Technology | Voice: (301) 975-3808
Gaithersburg, MD 20899 | Fax: (301) 990-4127
http://math.nist.gov/~mitchell
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|