Summary of video calls

[certik]

I will use this issue to document our progress.

October 22:

• Summary: https://fortran-lang.discourse.group/t/generics-proposal-for-202y-video-call/353/7

[certik]

Summary of November 5 call (present: @tclune, @gklimowicz, @rouson, @everythingfunctional, @FortranFan, @certik, @mleair):

• Work on use cases, some are performance critical in hot loops, some are not (only happens once in a simulation)
• We need to figure out a set of "values" or criteria to judge the requirements / specs / syntax proposals and convince the committee and the community that those are good values
• Reading material: #23
• We will be meeting bi-weekly
• @everythingfunctional provided a Haskell example in comparison.md (#18), semantically Haskell behaves very similarly to Rust

[certik]

Summary of November 19 call (present: @tclune, @gklimowicz, @FortranFan, @mleair, @certik):

• @tclune created an initial document for use cases: #26
• We had a lot of discussion about use cases, going over and improving #26
• We clarified what our primary goal should be: Fortran being the best language in its domain of numerical scientific array oriented computing. Our goal with generics is to help that primary goal. We don't mind expanding the domain, but that would be a secondary goal.
• We discussed the use case of the SciPy library, which still has Fortran as the most used language (besides Python), but it is shrinking relatively speaking towards C++, which will overtake it soon. The stdlib effort is to create a SciPy like library in Fortran. Generics can greatly help, currently stdlib uses fypp, but it would be nice to use generics instead. The use cases from this application are type, and especially kind and rank agnostic functions, which we included in #26.

TODO for next time:

• The same as in https://github.com/j3-fortran/generics/issues/24#issuecomment-722620836, so mainly the "values" document, and more use cases