Richard Maine [log in to unmask] wrote:
> I think I favor a position simillar to that expressed by one or two
> others here - that procedureness should be a disambiguator. But
> I'm dubious of the wisdom of trying to disambiguate based on the whole
> interface. I also think that would be a big enough project to fail
> to fit as a "minor technical enhancement". Disambiguating on just
> procedureness might fit there, though. (And then because we already
> disamkbiguate on type - yes, a function returning a real would
> also be distinguishable from a function returning an integer, etc).
A procedure that is known to be a subroutine should be distinguishable
from a function, too. A dummy procedure declared by EXTERNAL but
without an explicit type declaration could be a function or subroutine,
so it's not distinguishable from any other dummy procedure.
I don't see any difficulty of distinguishing using entire characteristics,
other than the bulk of the rules, as in my previous message to comp-
fortran-90. I agree it probably is too big to be called a "minor
technical enhancement," and even if it could be so identified, it
ought to be put into priority order with all the other candidate MTE's
that didn't make the cut -- and we don't have time to do that.
Van
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|