I recently attended the NumFocus 2022 summit in Santa Rosa, California. I was not technically representing the Jupyter Project (that was Ana, Brian, and Jason), but was asked to join in order to give a talk about the Jupyter Book and MyST markdown ecosystem.
Here are a few thoughts and highlights from the meeting as I remember it. I’ll write the stuff I remember and can update things if more comes to mind. Perhaps @jasongrout @ellisonbg or @Ruv7 could share their experiences as well?
The meeting itself was organized around two days, you can find the full schedule here:
It broadly had two tracks: one that was a “talk-style” track, and another that were unconference sessions with topics decided on the fly.
Here are two quick pictures to show the unconference schedule:
I gave a talk about the Jupyter Book and MyST ecosystems, and a bit about where those projects might be going. A short update about this is at the link below:
The Scientific Python folks were in a number of sessions at the conference and were interested in building networks, sharing infrastructure, etc. For those who aren’t familiar, check them out here:
We discussed a bunch of improvements to make in the PyData Sphinx Theme (which we use across a bunch of Jupyter projects) and may consider moving it to the scientific-python organization. They also described their specs process and there generally seemed to be interest from folks in participating in this:
NumFocus announced that they’ve got a partnership with Bitergia to run an analytics dashboard for projects in the NumFocus ecosystem. I believe this will be available to Jupyter projects if we are interested as well:
I also heard that the small development grant series at NumFocus is going to increase in size to something like $500,000 over the year. The maximum grant amount is also growing a little bit to ~$20,000. FYI for any sub-projects that have some focused development or team support needs.
Leah gave a “state of the union” talk about NumFocus’ current situation and where it’s going next. It sounds like one of the challenges that NumFocus has is that it only recovers roughly 30% of its operating budget from fiscal sponsor fees (the 15% that NF takes from grants and donations). It makes up the remaining 70% through donations, pydata conferences, etc. We had some discussion about the pros and cons of this, but there wasn’t a specific “conclusion” to this bit.
There was a nice panel discussion about various ways to grow maintainer communities via sprints and infrastructure for onboarding and training. I’m not sure whether the notes for this are public, but Reshama Shaikh from the scikit-learn project shared a video to watch as a nice prep, that others may be interested in:
There was a conversation about governance and @jasongrout gave an update about where Jupyter’s governing model is heading. I’m not sure if there are notes from that conversation, but somebody shared my favorite document (The Tyranny of Structurelessness) so I’ll post it below