Hello,
I have published this paper on JOSS: http://joss.theoj.org/papers/10.21105/joss.00458
We have worked on a new version that adds a lot of new features. We are very near into publishing this new version 2.0.
I have updated the paper.md to match the new version, specifically updated the Features List and the Contributors sections. My question is, can the _published_ paper be updated as well, once we do our formal release?
No. You cannot change the published paper.
But if the new version has extensive new work, you can submit a new paper.
It sounds to me like this could be a new paper.
:(
I don't think it is enough to publish a new paper, nor do I think it makes much sense; the current one has been cited once, and I'd like to use the same.
I am very surprised by an answer so definitive. Since the journal is all about software papers, and significant software updates happen very often in the huge amount of repositories published here, shouldn't something like this be supported or at least considered?
We are fairly flexible on the new paper for a new major software version, this has already been done a few times. However, since JOSS is trying to fit within the traditional journal ecosystem, we can't really have papers changing over time.
Plus, a JOSS article is not necessarily the best route for people to cite the most recent version of a software package鈥攚e are trying to solve a credit & career issue rather than reproducibility. For the latter, people should be directly citing the specific software version they used in a paper, perhaps in addition to a JOSS article.
So, the typical process (or at least what I would recommend) for software evolution and JOSS would be:
(This may "dilute" the citations to the individual papers, if people only cite the newest one, so if that is a concern then you may prefer to just do steps 1 and 2.)
Thank you @kyleniemeyer . Okay then, I will simply use the automatically generated zenodo, but will still point people to cite the published article and maybe mention the version they use.
Most helpful comment
Thank you @kyleniemeyer . Okay then, I will simply use the automatically generated zenodo, but will still point people to cite the published article and maybe mention the version they use.