Re: [tlug] Fortran --> Python (was linux engineer)

[Nicolas Limare <nicolas+tlug@example.com>]

>    As for the question "Why (still!) Fortran?": When I was in graduate
>  school (20 years ago...sigh), we had a collaborator from Germany
>  who constructed state-of-the-art stellar atmospheres. I asked him how 
>  big his (fortran) code was. "When I was in grad. school it was a few 
>  million lines...I haven't counted since then.")  So... Who is going 
>  to re-write a few million lines of Fortran code into Python? (And
>  why?!)

Dijkstra wrote, in EWD498 "How To Tell Truths thet Might Hurt?",

  In the good old days physicists repeated each other's experiments,
  just to be sure. Today they stick to FORTRAN, so that they can share
  each other's programs, bugs included.

Irony aside, rewriting these codes is a colossal work, but
multi-million lines codes written by a researcher can not be read and
verified, and will likely include lots of bugs and differences between
the documentation and implementation of the algorithms. A clean
rewrite with well defined APIs and unit tests would be necessary to
make computational research verifiable and reproducible.

>  It's a good bet that graduate students studying under him did not write
>  their own stellar atmosphere code in perl or python or c

They could write in C or C++, mixing these languages with FORTRAN is
easy. Unless the PhD advisor can't read C or C++, which is frequent.

> I've also heard from some theoriticians - even 'young' ones - that
> Fortran is still the superior language for numerical work.

I think the superiority is in the compilers, better tuned for unusual
HPC architectures than gcc or icc. They also benefit from the
possibility to manipulate the FORTRAN array types better than any
C/C++ programmer would do in loops over scalars.

-- 
Nicolas LIMARE
http://nicolas.limare.net/                         pgp:0xFA423F4F

Attachment: signature.asc
Description: Digital signature