Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

The future of Fortran

99 views
Skip to first unread message

john.chl...@gmail.com

unread,
Oct 14, 2012, 6:29:39 AM10/14/12
to
Since I'm in the middle of this project to translate some ugly C++ to modern Fortran that brings up a question:

Are the fortunes of Fortran waxing or waning?

Never used Fortran before this effort but it seems like a hell of lot better language for numerical code (engineering and scientific) than C/C++ (too low level - better for systems and embedded work).

BTW, gfortran seems like a quality product too - haven't tried any commercial compilers though.

---John

Dan Nagle

unread,
Oct 14, 2012, 8:20:03 AM10/14/12
to
Hi,

On 2012-10-14 10:29:39 +0000, john.chl...@gmail.com said:

> Are the fortunes of Fortran waxing or waning?

Seeking a data point in what is a matter of opinion, ...

A few meetings ago, an applications-oriented member of J3
asked the vendors collectively whether they were committed
to Fortran for the long haul. The question was based
on the premise that the member's sponsor (a NL who should
remain nameless since I don't speak for them) would devote
more effort to gfortran if not.

The collective answer from the vendors was that compiler sales
were generally increasing, and as far as they could tell,
new projects were being started in Fortran. As long as that
was the case, the vendors were dedicated to Fortran.

Fortran sells hardware. As long as that's the case, there
will be Fortran.

Beyond this lies the mine's-bigger-than-yours debate
over which language's feature set best matches an application,
and whether that even matters to programmers who have favorites.

--
Cheers!

Dan Nagle

Janus Weil

unread,
Oct 14, 2012, 8:30:45 AM10/14/12
to
Hi,

> Since I'm in the middle of this project to translate some ugly C++ to modern Fortran that brings up a question:

ten years ago people would probably have laughed straight in your face for such a sentence, thinking that you probably got it mixed up and actually wanted to translate "ugly Fortran" to "modern C++".

Nowadays, the opposite seems to become more popular again (you're not the first one I have heard this from).


> Are the fortunes of Fortran waxing or waning?

Asking such a question here, it's pretty clear what sort of answers you can expect ;)


> BTW, gfortran seems like a quality product too - haven't tried any commercial compilers though.

In my opinion, having a high-quality open-source compiler publicly available gives a strong boost to language adoption and the development of some sort of "community".

In that respect, C++ was clearly ahead for some time, but Fortran (in the form of gfortran) is starting to catch up, I guess ...

Cheers,
Janus

\"Vladimír Fuka <"name.surnameat

unread,
Oct 14, 2012, 8:51:07 AM10/14/12
to
> BTW, gfortran seems like a quality product too - haven't tried any
> commercial compilers though.

Last week a friend who does supercomputing (fluid flow) with MPI with even
thousands of processors told me, that gfortran produces faster executables
than both Intel Fortran and IBM XL Fortran for his code.

glen herrmannsfeldt

unread,
Oct 14, 2012, 6:03:30 PM10/14/12
to
Dan Nagle <danl...@me.com> wrote:

(snip)

> Fortran sells hardware. As long as that's the case, there
> will be Fortran.

> Beyond this lies the mine's-bigger-than-yours debate
> over which language's feature set best matches an application,
> and whether that even matters to programmers who have favorites.

As with voters, many have already decided. The undecided may
choose the language most appropriate for the project at hand.

Though with C interoperability, it isn't so hard to mix
the two, or any C callable language, in a single project.

-- glen

john.chl...@gmail.com

unread,
Oct 14, 2012, 10:50:33 PM10/14/12
to
It has been my experience that most C++ applications I've look at go from somewhat ugly to unbelievably ugly. To tell you the truth, I'd rather work on existing C code than existing C++ code.

Somehow, people that code in C++ insist on using all its complexity - to do otherwise simply isn't using C++ properly. I've noticed of late, since C++ supports multiple inheritance, the use of abstract C++ classes as Java interfaces. Yet, more no-value-added complexity added to the project.

---John

Wolfgang Kilian

unread,
Oct 15, 2012, 3:44:14 AM10/15/12
to
On 10/14/2012 12:29 PM, john.chl...@gmail.com wrote:
> Since I'm in the middle of this project to translate some ugly C++ to modern Fortran that brings up a question:
>
> Are the fortunes of Fortran waxing or waning?

What is certainly decreasing is the number of people who confuse Fortran
with FORTRAN (77 or 66 or ...). In my personal biased experience, it is
probably still the majority of computer 'experts', though.

-- Wolfgang

--
E-mail: firstnameini...@domain.de
Domain: yahoo
0 new messages