Inconsistent documentation

[athas]

The very first paragraph of the README begins with

With Staffeli, we work with local course clones.

The term "course clone" is never used again after the Getting Started section, nor does the term "clone" occur anywhere else in the repository.

[athas]

It is also unclear what We aim to keep these clones compatible with git means.

[kfl]

Is the following more understandable:

With Staffeli, we work with local course clones. These clones are local copies of course
information from Canvas, stored in a format that facilitates collaboration and version control,
such as with `git`. However, using a version control system is not required.

We recommend creating a local directory named `canvas`, `absalon`, or similar, toorganise all
your Canvas-related local course clones.

[athas]

I still don't understand what it means. My understanding of Staffeli is shallow. From what I can see it doing, I don't see it cloning any course information, except if you count submissions. I never have to synchronise Staffeli with what happens on Canvas. Or maybe I do and I just haven't noticed?

My understanding is that Staffeli is able to download submissions for an assignment, given a course ID and a hand-written metadata file containing information about which assignment to download. Whenever Staffeli downloads submissions, it also creates metadata files that allows Staffeli to upload corrections.

I don't understand the "clone" metaphor at all, since it seems like the submissions I download for different assignments (for the same course) are not connected to each other in any way. It really seems like Staffeli does not care about courses, except for using course IDs to look up information. To me, "cloning" implies that I have a complete copy of some remote resource, which must be regularly synchronised. That would be a terrible thing, so I am happy Staffeli does not actually appear to treat Canvas like that.

[athas]

To get an idea of my mental model of how Staffeli works, see here: https://github.com/diku-dk/ap-e2025/blob/main/assignments/staffeli.md

Staffeli works OK when using the workflow described above, modulo some ergonomic issues which I think are related to Staffeli really preferring the use of rubric-based grading.