Mike
>Since recent discussion centres on "improvements" for future Fortran,
>may I suggest the introduction of "^" symbol for exponentiation, to
>eventually replace "**".
>
>This would be consistent with some of the algebraic manipulation
>systems, and
>remove the possibility of certain typos going undetected, if the old
>method were to be
>disallowed (using a special compilation flag) except for legacy code.
Why?????
This will only increase the plethora of syntax that is duplicated. Most
compiler vendors will the support both syntaxes. Fortran 90 and its full subset
support for F77 gave us so much duplication of syntax. I just wish that there
is only one way of doing things (and generally not the F90 second way).
Fortran is not C, which is what most of these "systems" are based on. If we
follow this path, Fortran should change its matrix storage schema, which will
probably break about 90% of application code in existence.
I'm not sure where "typos would go undetected", since if you write a routine
which uses *2 rather than **2, surely you test the algorithm. Aren't all
algorithms tested to boundary conditions and normal application? I doubt that I
am an exception in this respect.
Slightly OT
Even IMPLICIT NONE is not the panacea for rigorous testing. It does catch many
things, but if you mistakenly declared a REAL as an INTEGER, you're "safe", but
your algorithm is not. [I found such an instance in one of our applications,
and obviously I never wrote the code :-)]
Regards, Paddy
Paddy O'Brien,
Transmission Development,
TransGrid,
PO Box A1000, Sydney South,
NSW 2000, Australia
Tel: +61 2 9284-3063
Fax: +61 2 9284-3050
Email: [log in to unmask]
Either "\'" or "\s" (to escape the apostrophe) seems to work for most people,
but that little whizz-bang apostrophe gives me little spam.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|