> On May 31, 10:04 pm, Tim Prince nospamcomputer.org> wrote:
>
>
>

>> Damian wrote:

>>> On Jun 1, 10:34 am, user1 example.net> wrote:

>>>> Jan Gerrit Kootstra wrote:

>

>>>> [snip]

>

>>>>> The size of the usergroup is far under estimated, in engineering and
>>>>> science it is one of the major programming languages.

>>>> Are there solid statistics somewhere that indicate the use of Fortran in
>>>> science and engineering applications exceeds the use of C, C++, Java ?

>

>>> A 2006 survey of United States Department of Defense High Performance
>>> Computing users showed a bit less than 60%% of users write "prototyping
>>> software" (not sure why that was the focus) in Fortran and roughly 85%%
>>> write in C/C+/C#. Obviously, many are engaged in multi-language
>>> development, which I think will become even more common with the new C
>>> interoperability features of Fortran 2003. I can't find the survey
>>> results online anywhere, but I can send a PDF to anyone who wants a
>>> copy.

>

>>> Damian

>

>> Maybe they were looking for verification that there may be a shift away
>> from Fortran, which accounts for a much higher percentage of production
>> HPC usage.
>> Would you buy a car which hasn't been designed with Fortran crash safety
>> analysis? Or, ride in a commercial aircraft without crash or bird strike
>> certification?
>> What is the Fortran usage statistic for weather and climate analysis?

>
> I only have anecdotal evidence from working with collaborators, but it
> appears Fortran is used almost exclusively in the climate/weather
> arenas -- some Fortran 77 but primarily Fortran 90. I think partly
> this is because some of the complications (unstructured grids &
> multiphysics interactions) that led scientific programmers away from
> Fortran a decade ago are less important in climate and weather
> modeling.
>
> Damian