> Date: Sat, 26 Aug 2000 19:34:53 +0010
> From: [log in to unmask]
> >Furthermore, it only applies to default precision complex, not to
> >non-default [read double precision] complex -- which stands to reason,
> >as for compatability with earlier versions of Fortran, only
> >default-precision complex need be supported in this regard.
>
> Since we seem to have determined how standard complex looks to the programmer
> and the compiler, does any vendor who uses double precision complex not use a
> similar layout of the parts?
If you want to use double precision complex et al, you have to use
the F90 constructs. EQUIVALENCE is not available. The designers
apparently don't want to perpetuate an error-prone facility.
> >The practice of using EQUIVALENCE in lieu of more
> >modern forms apparently is being discouraged.
>
> But aren't we also talking about code that we wrote back in the sixties? If it
> ain't broken ...
No. The code offered ws clearly Fortran 90, if you recall --
> > temp(1:N,1:N) = real(mat(1:N,1:N)
> > call filter(filtercoefs, temp)
> > erg(1:N,1:N) = temp(1:N,1:N)
> >
> > temp(1:N,1:N) = real(mat(1:N,1:N)
> > call filter(filtercoefs, temp)
> > erg(1:N,1:N) = erg(1:N,1:N) + cmplx(0.0, temp(1:N,1:N))
> I am not going to rewrite routines that work just because some constructs are
> discouraged or obsolescent.
If you need to change code using EQUIVALENCE to double precision
complex, you will have to rewrite it.
> I know that most vendors (definitely, it seems,
> mine) will always support FORTRAN.
>
> I do not believe that a standards committee should dictate style, once a
> construct has become part of the language. (cf, the discussion on GOTO about 5
> months ago.)
Some archaic forms thankfully were deleted (e.g. Hollerith constants).
One would hope that folks would use the Block IF statement
in lieu of the logical IF and GOTO and arithmetic IF in all new work . . .
and of course there are better ways than lots of GO TO statements.
> Regards, Paddy
> Paddy O'Brien,
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|