Damian Rouson

@zbeekman Given that we're adopting FORD, everything in [doc/](https://github.com/sourceryinstitute/opencoarrays/blob/master/doc/) can be deleted _except_ for [doc/dependency-tree/](https://github.com/sourceryinstitute/opencoarrays/tree/master/doc/dependency_tree), which should be retained. The banner that install.sh prints at the beginning of execution contains...

Sure. Feel free to do it or let me know if you'd like for me to do it.

Bravo! :clap:

Also, the `sync` type-bound procedure suffers from an unfortunate ambiguity in its name. I understand the reason for the name, but I'd just call it `broadcast` or something similar because...

@milancurcic perfect. Thanks. I'll submit a PR with the suggested test.

@milancurcic It still might be that test_network_sync is unnecessary. Fortran 2018 provides an intrinsic subroutine: ` random_init(repeatable, image_distinct)`, where both arguments are `logical, intent(in)` and where passing `image_distinct=.true.` ensures that...

Actually yet another way to view this is that the test is useful either way and in fact, the test should give the same result either way. How that result...

I have determined that with `gfortran` 11.2.0 and the head of the main branch of OpenCoarrays, calling `random_init` with `image_distinct=.false.` produces the same random numbers on each image without need...

Most likely, yes. Certainly the current approach should work with GFortran 10. If `random_int` works with 10, I'll submit that approach as a separate PR.

Moving the procedure definition to a submodule also generates a compiler error: ``` % tree . . ├── ford.md └── src ├── foo_m.f90 └── foo_s.f90 1 directory, 3 files ```...