How to cite this repository?

[BenjaminRodenberg]

There is an open PR on this topic: precice/fenics-adapter#137

Let's try to finalize this before we release the first version of the fenicsx-adapter.

[uekerman]

This one is indeed tricky. Is this work more a "fork of the FEniCS adapter" (1) or does it more "use" the FEniCS adapter (2)?

Depending on this, it could be either

• Arved + "authors of FEniCS-preCICE paper" (1)
• Arved + whoever contributed here (2)

For either case, we could create an entry on Zenodo. For (2), we could then additionally cite the paper in the README, "this work uses [1]" ...

[BenjaminRodenberg]

I would definitely go for option (2), because this repository is an independent project.

There is, however, still a very strong relationship between this repository and the FEniCS adapter paper:

• We use the same API (-> credits go to the FEniCS paper authors), but we have to modify the implementation and some tests (Arved, Ishaan and I)
• We use the same setup for the tutorials (-> creadits go to the FEniCS paper authors), but there are slight compatibility updates (Mostly Arved)

We should make sure to show this strong relationship here to give credit to the authors of the paper, but also to reduce the effort we have to put into documentation and redesigning the API (I would rather not write another paper about the FEniCS-X adapter, if it is not necessary).

Putting the first release on Zenodo would be great, because we (1) will not have a dedicated paper for the FEniCS-X adapter (see above) and (2) still need a way to cite the FEniCS-X adapter and give credit to contributors of this repository that did not author the FEniCS adapter paper.

[BenjaminRodenberg]

I'm trying to summarize some thoughts from precice/fenics-adapter#137 and apply them to the situation for the fenicsx-adapter. We generally have the following three possibilities to refer to this repository:

1. Write a paper about the fenicsx-adapter and point to it here: Requires a lot of effort; almost always outdated, because we don't write a paper for each version.
2. Create a Zenodo entry for the first release of the fenicsx-adapter and point to it here: Noisy and requires to take care of this for each release. But would be up-to-date and allow to point to each specific version.
3. Use the preCICE distribution where the fenicsx-adapter that contains the release of the fenicsx-adapter: No additional overhead, because the preCICE distribution is released on a regular basis. Some versions of the FEniCS adapter might be missing, but we at least have an update every half year.

I'm undecided between option (2) and (3), which is also how I understand this comment by @uekerman.

[arvedes]

That's tricky, indeed. I'd be okay with any of the options you mentioned. I like the concept of Zenodo a lot, however I'm not sure if it'd be as accepted in the scientific community as a reference as a paper would be. Do you have any experience with that?

I could also imagine to contribute to a new paper about this adapter, however I won't have the capacity to do that as a first author.