It’s again time to organise our next monthly call which will be in the week of February 14-18 ; please see the following doodle poll to mark your availability:
I suggest that on this call we focus solely on the Google Summer of Code org application, prospective project ideas, prospective mentors, and organizing ourselves toward putting together the application like we did last year. The org application window opens today (Feb. 7) and will close in two weeks (Feb. 21, 18:00 UTC).
It’d be great if anybody interested in helping out can think this week about project ideas, and if you’d like to participate either as a participant, mentor, or co-administrator.
Not a new idea by any means and this may have more to do with inter Community collaborations: can some Summer of Code projects or other initiatives by the Fortran Discourse Community motivate and facilitate more Fortranners joining GCC/gfortran volunteers to help out with gfortran advancements?
Just a thought about possible GSoC subjects: how about modernising all these venerable packages, like the list Beliavsky posted? Besides modernising the code itself and the interfaces, it would also require modernising examples and documentation.
Unfortunately I don’t think they have been very successful in recent years in attracting (or accepting) students to the Fortran projects (last student was in 2014).
We can certainly use our channels to promote these projects to potential GSoC students!
I suggest we try to run GSoC in such a way so that at least some students stay around after the summer is over and keep contributing to Fortran related projects or open source in general. If you agree that is a worthy goal, then we can discuss at the call what are some strategies to achieve that.
One idea that I think works well is to find mentors who are excited about the project and help the student interact with the wider Fortran community and get the student excited enough to stay after the summer.
I’d be interested in mentoring this year. My preferred topic would be either some new component for stdlib, or help out in other fortran-lang repositories (fpm, minpack, documentation).
Below is a short summary of the discussion on GSoC (led by @milancurcic; to all particants at the call: don’t hesitate to correct/add if needed):
Administration: @certik + @milancurcic + others? (If someone wants to help with administration, please mention it)
Preparation:
Must be done: Update of the section “Student instructions”
Project ideas: some ideas were not addressed in 2021, or since then;
FPM Idea: “fpm-ize the Fortran ecosystem”: might not fit in GSoC. Can stay here, but need to find some contributors who would like to provide a project/library for this idea (e…g, providing a project for fpm)
FPM Idea: “Tree shaking”: Needs really semantic analysis (e.g., external procedures, C interface) ==> could be quite difficult. Could be useful to break it in smaller tasks for acceptation. Needs a good Fortran scanner
FPM idea (@awvwgk ): Compilation of C++ code (e.g. to compile LFortran)
FPM idea (@lkedward : Shared library; Maybe too complex/complicated for GSoC (?) because of different OS (Linux, MacOS, Windows). Needs prior knowledge. Maybe focus on Linux only?
Stdlib idea
Stdlib idea: adding support for sparse matrices in stdlib
Other stdlib ideas should be provided in the wiki page
LFortran idea : compiling stdlib, fpm + other Fortran-lang projects
LFortran idea: Fortran in the browser (through Webassembly?). @awvwgk: Two different approaches; Fortran in Jupyter: works, but would run on a server (does not require Webassembly or similar)
Other
Idea: Modernize (improving distributions, test suite, documentation,…) classic Fortran libraries; Good idea, but how to judge if the project is successful? Maybe search for old package codes that are currently redistributed/used and to modernize them. Possible issue (@awvwgk): new programmers will be expose to old legacy codes. However, GSoC is not limited to students.
Idea: Port fortran-lang.org to Sphinx. Well-defined project, we have some experience with Sphinx. Good for people not really interested in Fortran, but would like to lean on website generation
Idea: build or embed online playground for fortran-lang.org. Different than the idea “Fotran in the browser”. Possible idea: Web VNC client embedded in a webpage (as presented by @Arjen), but might be overkilled.
Idea: fortran-lang/minpack into SciPy: @awvwgk: this is most likely almost done ==> not sufficient scope for GSoC.
