search

ME-ICA / tedana

TE-dependent analysis of multi-echo fMRI
https://tedana.readthedocs.io
GNU Lesser General Public License v2.1
161 stars 95 forks source link

CFF #881

Open CPernet opened 2 years ago

CPernet commented 2 years ago

Since there is no direct paper to the toolbox, any of you guys wanna add a citation.cff so we can reference your awsome tool properly ... https://citation-file-format.github.io/ that would be great - thx

cff-version: 1.2.0 message: "If you use this software, please cite it as below." authors:

tsalo commented 2 years ago

This is something we have been discussing in #820, but the problem is that the CFF format doesn't currently support multiple citations very well, which is a problem when we want folks to cite both the JOSS paper and the appropriate version-specific Zenodo DOI. I believe that it's possible to include multiple citations in the CFF file, but GitHub will ignore everything after the first one.

CPernet commented 2 years ago

well yes it cites the github repo but you can add all the references, here is one of mine https://github.com/LIMO-EEG-Toolbox/limo_tools/blob/master/CITATION.cff

tsalo commented 2 years ago

I am still concerned for a few reasons:

  1. Even though your CFF file has multiple references, even selecting the BibTeX option in "Cite this repository" doesn't export those references. In practice, does anyone actually cite those additional papers? It seems like giving folks an easy alternative to the method boilerplates you have in the limo_tools wiki (which do have all of the necessary references) is asking them to ignore the more complete citation.
  2. CFF doesn't seem to integrate with tools like versioneer at all, so we would have to make a commit to update the version number in the CFF file before each release. Our current release procedure is pretty streamlined through the use of versioneer and Actions to update version numbers and publish to PyPi automatically. I would hate to add extra steps to it.
  3. How do folks handle the Zenodo DOI situation? We only get a DOI minted after we make releases, which means we can't include the DOI in the release code (e.g., directly in our boilerplates). This is a problem with our current setup as well, so I can't say that CFF would make things worse, but I would hope that anything we do to improve the management of citations would consider that as well.
  4. Tedana automatically generates pretty detailed boilerplates that are constructed based on user-provided settings, so a static list of references wouldn't be sufficient.

That said, I can't speak for the rest of the maintainers team, so I'd like to get their thoughts. @ME-ICA/tedana-devs WDYT?

The possibilities I can come up with are:

  1. Add the JOSS paper to the CFF file as the preferred citation. We could include the "all versions" Zenodo DOI as well, but I don't think anyone would actually cite it.
  2. Use the CFF file in its default setting (i.e., cite the repository). No Zenodo DOI included.
  3. Use the "all versions" Zenodo DOI.
  4. Add a step to our release process where we upload the code to Zenodo manually before the release is made, to get a version-specific Zenodo DOI, then push the new DOI to our CFF file (and possibly elsewhere in the repo), and then make the release.
handwerkerd commented 2 years ago

I guess I'm the grumpy elder here who thinks that, if you're using software you should at least read enough of the documentation to know what they ask you to cite. Standardized formats can make this easier, but they also risk hurting reproducibility if they automatically generate reference lists without making a researcher think about what each of the references are.nThis is a long way of saying I'm not the intended audience for this so take my opinion with a grain of salt.

If our goal is to make sure everyone who uses tedana cites at least one thing, they should cite the JOSS paper (your option 1). Anyone who actually cares about reproducibility would also includes the zenodo DOI and/or the version of tedana that they ran.

eurunuela commented 10 months ago

From reading the comments, it sounds like we probably want to add the JOSS paper to the CFF file, but not the other citations. Is that right?

Just trying to keep the conversation going to see if we can take some action on this.