openjournals
AAS/JOSS software metadata
For papers published in collaboration with AAS, it would be valuable to require software authors to provide (at minimum) citation metadata in a standard format (CFF or CodeMeta).
The metadata file should point to the software's Zenodo DOI and reference the JOSS and ApJ papers (e.g. using the CFF "references" element). If you download the software from Zenodo then you could still link it back to all three records.
It sounds like it would be valuable for all JOSS papers. The JOSS paper DOI is known at the time of submission, so for sure authors could add it to a CITATION.cff file that is part of the software repository and archive.
Not sure if they can add the Zenodo DOI, though, as they make the archive deposit after finishing up all their changes. (Unless they reserve a Zenodo DOI, go edit the citation file, and then make the deposit, but I dare say giving such instructions to authors would be clunky.)
@arfon : Maybe whedon could automatically create a CITATION.cff file upon running the @whedon accept command, and message the author to download that file into their repository? (Also, while we're at it, it just occurs to me that whedon could give the authors the instructions then about next steps like archiving, so we take away that task from editors.)
The CFF project provides a tool to automatically generate the .cff file from a Zenodo archive DOI: https://github.com/citation-file-format/doi2cff
@arfon : Maybe whedon could automatically create a
CITATION.cfffile upon running the@whedon acceptcommand, and message the author to download that file into their repository? (Also, while we're at it, it just occurs to me thatwhedoncould give the authors the instructions then about next steps like archiving, so we take away that task from editors.)
Yes, that's an interesting idea, and definitely something we could do with Whedon. We could generate it at the same time we post the 'goodbye' message at the end of a review (e.g. like this).
The idea of how to store citation metadata is in flux, but codemeta.json is closer to where I think we will end up than citation.cff
