Evgeni Burovski

I'd like to understand the lineage and the status of this method. If it's completely new, I'd second Pamphile's opinion: it should mature as a separate package. However, is it...

Also cross-ref gh-1477 and gh-18118.

Thank you @zaikunzhang > It is in Fortran, which I regard as a different language that conveniently benefits from the abundant libraries in FORTRAN without bearing the historical burden. OK,...

Thanks Tyler, Jake! Indeed, the plan is to move parts of this from python to C. I'm doing it piecemeal over at gh-19753, with an intent to sync this one...

I think this is ready from my side. Several cleanups are probably needed for the cpp side. At least - how to best document cpp routines; is the current way...

> I'm going to leave this PR as is, and first take a look at parametric version, splPrep instead. Huh. Implementing splprep turned out to be a great of a...

Am tentatively marking this with a 1.14.0 milestone. It's definitely not critical for the release, so can be bumped if it gets in the way. The `mdspan` discussion is somewhat...

Would be great @Kai-Striega ! Also, https://github.com/ev-br/scipy/pull/13 adapts to using mdspan. I checked that the the mods work with both kokkos mdspan and Irwin's version from scipy/special. The link is...

There certainly is interest, I'd say! Ideally the perf improvement happens in ndimage --- the duplication between scipy.ndimage and scipy.signal filtering routines is mostly historical, and would be nice to...

1 step at a time is good. So the story along the lines of "scipy.signal.medfilt does better in 1D and delegates to ndimage otherwise" sounds reasonable. As a first step...