I was thinking the other day if would be a good idea to have on fortran-lang.org some project showcases, that are good example of fortran usage and show the impact.
I would think this will need to be a more formal stuff so we do not end up with goto style projects.
5 Likes
I like this idea. I would want to see a proposal about how candidates would be selected and reviewed. Specifically I think we would want some more objective guidelines so it’s not just “in the editor’s opinion”.
A few GitHub contributors who exploit the features of modern Fortran are (in random order)
Additions to the list are welcomed.
R has a huge number of packages, and to navigate them, it has a set of curated Task Views listing the most important packages in various domains. It would be nice if Fortran had something comparable. The Guide to Available Mathematical Software (GAMS) shows its age, typically referring the user to FORTRAN 77 packages and the NAG and IMSL commercial libraries. Ideally it would also cite more recent codes in the public domain.
How is see it is more or less a committee that agrees on rules and publishes them, community and not only makes submissions and these are passed for peer review on scientific merit (this shall be relatively straight fwd since things I have in mind are already published) and programming merit (done by experts in the programming art; modern Fortran, so netlib blas implementation will not qualify). The aim is not to neglect what was good in the past but to showcase what is leading these days.
Also the target shall be real life applications and not proof of concept/co design exercises… maybe a separate category for such things.
my personal preference will be only open source stuff to qualify and goto to be a non qualifying construct.
GOTO can be overused, but using GOTO to go to the end of a program or a procedure where errors are handled is not bad programming style IMO.
2 Likes
We have a somewhat regular influx of new Fortran projects to the package index and the fpm-registry which we could select from. Also, we have been highlighting newly added projects to the package index in the monthly fortran-lang.org newsletter for a while now.
The packaging guideline for the package index is a good starting point for criteria, maybe by making all optional points obligatory for the package of the month.
1 Like
I think I did not express myself properly. what I have in mind is not another list of projects that pass some abstract software engineering set of rules regardless of what the package does or whom is using it.
at the end of the day not all packages are equal in importance or quality…
I know the index… actually I even plan to contribute to it.
I see, I was more thinking along the lines of TWIR’s crate of the week due to the thread title.