10 October 2018: JOSS / pyopensci Collaboration#


  • Leah, Kylen, Arfon

Meeting Goals#

To establish a relationship with JOSS so that the review process for packages submitted to pyOpenSci are by default “accepted” by JOSS. JOSS then reviews the paper submitted and determines if the package is in scope.

Joss’ Review process varies from rOpenSci As Follows#

  • no required autotesting autotesting

  • And there are a few other items. Essentially JOSS’ process is a bit less strict then what ours will be.

Joss’s primary goal#

  • to provide career credit for people developing software. Thus the paper component of the review process is emphasized a bit more than the technical component. rOpenSci and pyOpenSci have to consider long term maintenance therefore the technical pieces are a bit more important

Joss has some additional checks including#

  • authorship, the paper about the software which becomes a citable item and finally the scope of the package

JOSS’ submission requirements for what tools are accepted is a bit stricter

  • ropensci accepts some things that joss won’t accept - example:

  • API wrappers – if that is a very small contribution it’s a minor utility category - joss doesn’t publish this - to avoid peopel submititng 50 lines of code to joss

    • When people submit to ropensci they check if they want to publish in JOSS. At some point during the review, they write the ppaper if they check “yes”.

    • If would be interesting to add something to ensure that the submittor verifies the criteria of JOSS to ensure it fits into the scope. we just want to ensure that they KNOW there is no guarantee that it will be accepted by JOSS.

    • Arfon will send a link to their submission criteria – this hasn’t happened yet with ropensci – but it could happen someday. so we should set expectations upfront about it. language in our submission template should confirm this…

  • https://joss.readthedocs.io/en/latest/review_criteria.html

Next Steps#

  • We will send Arfon our review process documentation. He will discuss that with the editorial team who meet in 2ish weeks. We should know in 2-3 weeks what the outcome is.