[RFC] [Ping] Comments on gfortran manual

[Tim Prince]

Steve Kargl wrote:
> On Thu, Dec 11, 2008 at 10:38:54AM +0100, Tobias Burnus wrote:
>> The kind value matches the storage size in bytes. It is recommended to
>> use the SELECT_*_KIND functions, but as backward compatibility the
>> Fortran standard allows to use the type name followed by an asterisk and
>> then by the kind value, e.g. @code{REAL*8}, with the exception that the
>> number has to be doubled for COMPLEX, e.g. @code{COMPLEX*8} is the same
>> as @complex{COMPLEX(kind=4)} (default kind complex)."
>>
> 
> The Fortran standard has never allowed the asterisk form.
> This became a de facto standard in that years ago some
> vendor(s) (DEC?, IBM?, CDC?) used it, so everyone elser
> needed to support this form.
> 
CDC didn't support it.  It didn't become de facto almost portable until
the world settled on 32- and 64-bit platforms.  It's certainly wrong to
suggest there was a standard supporting REAL*8, even though it goes back
to f66 on IBM, and to many brands of f77.  I doubt REAL(8) was originally
conceived as equivalent to REAL*8.  It's good to point out the discrepancy
in numbers for COMPLEX.  In my experience, prior to f90, DOUBLE COMPLEX
was more widely supported than COMPLEX*16, and default COMPLEX was never
written with a size indicator.