On Thu, 25 Jun 1998, David Adams wrote:
> My old Fortran 90 notes say that DOUBLE PRECISION *might* be marked
> as "obsolescent" in the next version of Fortran.
>
> Is that happening or likely to? Would you recommend anyone writing
> new code to continue using DOUBLE PRECISION.
>
> And what about double precision constants? Should we still be using
> 1.0D0, or should that now be 1.0_DP, where DP has been defined in a
> statement like
>
> integer, parameter :: DP = selected_real_kind(10,50)
DOUBLE PRECISION is certainly to remain a part of Fortran for a long
while to come, even if "just" for the reason of backward compatibility,
and if not in the standard then still in nearly all compilers.
But I think it is a GOOD IDEA to use the modern REAL(kind=my_fp_kind)
notation throughout. It is easier to port and can be used for COMPLEX
as well. So far there was no official version for "DOUBLE COMPLEX" though
most compilers have a complex*16 (and real*8, etc).
In the subset F you are required to use the modern notation. ELF probably
too, but I have to check. So it is probably an even better idea to stick
with these subsets as much as possible. They seem to be quite a safe bet
for the future.
Cheers,
Werner
-----------------------------------------------------------------------
| Werner W Schulz |
| Dept of Chemistry email: [log in to unmask] |
| University of Cambridge Phone: (+44) (0)1223 336 502 |
| Lensfield Road Secretary: 1223 336 338 |
| Cambridge CB2 1EW Fax: 1223 336 536 |
| United Kingdom WWW: |
-----------------------------------------------------------------------
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|