domcoltd

Hopefully, the SciPy community can stay open-minded about modern Fortran libraries.

Modern Fortran is quite different from Fortran 77, while being as powerful, if not more.

In addition, there has been a significant community effort on improving and modernising the legacy packages, the ecosystem, and the language itself.

With projects like LFortran (https://lfortran.org/), fpm (https://github.com/fortran-lang/fpm), and stdlib (https://github.com/fortran-lang/stdlib), I believe that Fortran will enjoy prosperity again.

Comment from Reddit:

https://www.reddit.com/r/programming/comments/13k74ro/optimi...

“Finally, a fortran package not written in Fortran77 but one trying to be approachable and maintainable rather than the product of unhappy PhD students spending many months to eke out a little more in an incomprehensible, Oracle-comparable mess.”

Indeed, the author mentions many bugs of the SciPy gradient-free solvers in the readme of the repo of Prima.

The extensive tests made by GitHub actions are simply amazing.

> DFOs tend to have issues past a few dozens variables.

It is interesting to check the results in section 5.1 of the following paper:

https://arxiv.org/pdf/2302.13246.pdf

where these algorithms are compared with finite-difference CG and BFGS included in SciPy.

Agreed. If (approximate) 1st-order information can be obtained in any way (even by *carefully* deployed finite difference), then gradient-based methods should be used. It is wrong to use DFO in such cases.

> DFOs tend to have issues past a few dozens variables.

It seems that the PRIMA git repo https://github.com/libprima/prima shows results on problems of hundreds of variables. Not sure how to interpret the results.

Good coffee is needed to understand Powell’s papers on these algorithms. If you don’t have, section 3 of the following paper is a good place to start with:

https://arxiv.org/pdf/2302.13246.pdf

GitHub repo of the project: https://github.com/libprima/prima