JOSS submission workflow is based on github, so we could draft the paper using existing github workflows (PR, user permissions, etc) on a separate branch

Thanks! I am thinking about promoting GDL specifically in the biology field. I know many groups use IDL, but they don't seem to have heard about GDL although there are also several conference papers describing GDL. JOSS is a good journal, but I doubt its readership from the biology field.

JOSS policy (https://joss.readthedocs.io/en/latest/submitting.html) explicitly mentions that: "Authors are welcome to submit their papers to a preprint server (arXiv, bioRxiv, SocArXiv, PsyArXiv etc.) at any point before, during, or after the submission and review process." so we can simply promote the JOSS paper by depositing it on bioRxiv as well.

Wow, I didn't know that. That would be awesome! Then I think adding some sentences regarding usage of IDL in biology fields and submit the paper to bioRxiv would gain more readership, since bioRxiv automatically tweets the new submission and promote it.

I vote for a single short JOSS paper highlighting astro, bio, geo and other domains of relevance, and then re-posting it on both arXiv's astro-ph listing as well as on bioRxiv, so that the news will spread through tweets, mailings, RSS, etc
Why? Just to keep things simple and doable in short time horizon.

I like your idea of writing a quick and short paper, but I actually think that a writing separate paper would still be helpful since IDL is mainly being used by benchworkers (i.e. experimental biologists), and JOSS is really far from the journals having a broad readership among them. But rather than writing a full paper, perhaps writing a benchmark paper is also quickly doable. Specifically, we can contact some groups using IDL, and just do some benchmarks of their in-house IDL codes, and put the result in the paper - it won't be super hard. It could be also helpful to improve GDL too.