Re: Fortran 77 to Ada

[Robert Dewar <robert_dewar@my-deja.com>]

In article <3839167F.F9E3B57E@maths.unine.ch>,
  gautier.demontmollin@maths.unine.ch wrote:
> So, please do constructive answers when you don't know the
> contents of the subject. Typically you decide it is pointless
> before trying to know what shape the Fortran code has.


Any automatic translation of Fortran code into Ada will result
in just what you expect: Fortran code in Ada, and this is almost
always a bad idea.

The best way to extend such code is to do it in a modular way
that allows new functionality to be added by adding new modules.
There is of course no reason not to use mixed language
(Fortran,Ada) for such a task, especially if you are using a
compiler that fully implements Annex G (including the important
pragma for using column major ordering of arrays, to be Fortran
compatible).

If the original code is not susceptible to being extended in a
modular manner, then it probably should be reengineered in any
case. The ideal thing is to do minimal reengineering in Fortran,
to allow the necessary extension, but if major reengineering is
needed then all the more need to recode instead of settling for
the junk that comes out of automatic translators.

There may be legitimate situations for using such automatic
translators, but all I can say is

a) I have seen dozens of cases in which people have tried to use
such tools where it was inadvisable, and often people were doing
it for the flimsiest of reasons.

b) I have never seen one legitimate case where it was helpful

Of course a lot depends on your quality standards. For me, the
code coming out of automatic translators requires so much hand
massaging to meet my minimal quality standards that the
translators are useless.


Sent via Deja.com http://www.deja.com/
Before you buy.