Hello,
[log in to unmask] wrote:
<snip>
>
> ..., so I supposed that the traditionalists said "What was
> good enough in the past will be good enough in the future."
And the developers have said "Since Fortran doesn't use []{},
we're free to assign completely new meanings to them."
I am strongly opposed to anything which would, for example,
render Co-array Fortran incompatible with the standard, rather
than being an extension. For more on Co-array Fortran, see
http://www.co-array.org/
In a comparison between Co-array Fortran and ZPL, I vastly
prefer Co-array Fortran. In part, it's due to the syntax.
And doesn't ratfor use {} ? I don't suppose there's all that
much ratfor, and you're supposed to run it thru the preprocessor
first, but two meanings for {} (ratfor and the new [whatever] use)
would certainly conflict.
> >
> > My own feeling is that if [] were to be allowed in Fortran, the best use
> > for them would be around array subscripts, making it much easier to
> > distinguish array references from function calls.
>
> I had to explain the generation of a function call by a typo to a fellow team
> member just this week.
That is a common error, but implicit none and ftnchek etc. exist
for such situations. That is, there are other means of catching
the error. The standard recommends use of external.
--
Cheers!
Dan Nagle [log in to unmask]
Purple Sage Computing Solutions, Inc.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|