(j3.2006) Asymmetry of REAL, CMPLX

[Malcolm Cohen]

>> >> On Mon, 2008-01-07 at 20:27 -0800, Robert Corbett wrote:
>> >>
>> >>> I am one of those who proposed providing a new intrinsic spelled
>> >>> COMPLEX that would be symmetric.  In what sense does it not solve
>> >>> the problem?

On Wed, 09 Jan 2008 08:43:00 +0900, Van Snyder <Van.Snyder at jpl.nasa.gov> wrote:
> This does nothing for existing codes.

On the contrary, it keeps some existing codes working, viz the ones that
use it deliberately to do what it currently does.  That would be all the
FORTRAN 77 codes that use it.

The only "existing codes" it does anything for are the ones with bugs in them.
Why should we care so much about code that already doesn't work?

> Of those that have double
> precision arguments for CMPLX, and no kind argument, I'd bet at least a
> dollar that 98% of them are expecting a double precision result.  So
> introducing an incompatibility, to correct a blunder in the design of
> Fortran 90, does them a favor.

It is not a blunder in the design of Fortran 90.

CMPLX of double precision returning a single precision result was
introduced in FORTRAN 77.

The only F90 "blunder", if one might call it that, was the near-absolute
insistence on F77 compatibility.

Re now is not the time to add COMPLEX; in my opinion it is 30-fold not
the time to be doing FORTRAN 77 interpretations and overturning unambiguous
specifications therein.

If CMPLX is really that unacceptable and needs fixing, let's change the
name while we do it so that customers don't get a "quiet change" in semantics.

Cheers,
-- 
................Malcolm Cohen (malcolm at nag-j.co.jp)