Hello fellow Jovyans,
This week (March 10th-15th) many of the core team members are going to meet in Washington D.C. at George Washington University in the Engineering Building (Thanks Lorena Barba for hosting us) for our (bi)-annual team meeting. We will try to keep you up-to-date with what is being discussed at the meeting.
For the first day we have between 15 and 20 team members in the room, with the most recent addition to the team being Letitia, who you might have heard of with her recent post on the jupyter blog about the native authenticator. The first day was quite slow with many of us lacking sleep from travel.
We used the morning to roughy flesh out some of the topics that we want to (try to) cover from the week, and came up with a short list of topics to tackle in the next few days, and dispatched small team to prepare the work.
- Mission / Vision / Value (Tim G., Lindsey, Paul, Fernando)
- Steering Council Refactor (Darian, Damian, Jason, Peter)
- Teams / Working Groups (Brian, Tim H)
- Distinguished Jupyter Contributors (Matthias, Darian)
- Maintaining Projects (Leticia, Erik)
- Communications (Zach, Carol, Saul)
- Diversity and Inclusion (Ana, Tim G, Brian)
- Jupyter Enhancement Proposal, aka JEP (Vidar, Brian)
- JupyterCon (Grant, Jason)
There are also several “appetizers” that can be discussed in smaller, more
After breaking for lunch, and reconvened in groups from 2pm to 3:30pm to start working on each of this point, then gather from 3:30 to 5ish to discuss the Future of JupyterCon. It’s already late in the day so we are going to break out for dinner and keep discussing less formally.
We of course welcome any comments on the above topics, or just which one of these you are interested in.
Keep in tune for next day summary. We’ll also try to update this message to be clearer if you have any questions.
Morning - Mission / Vision / Values
This morning we had a group discussion around Jupyter’s mission/vision/values. As expected, there was a lot of overlap as well as some divergence in how people thought of the Jupyter project, what its mission was, and where it should be going. The goal of this discussion wasn’t to choose a single thing that we all agree on, but instead to highlight the different perspectives of Jupyter team members.
(note that this involved a lot of raw discussion and summaries of discussion, we may have missed something, or potentially mis-characterized something). More details to come about what we discussed!
Some of the work notes (that might be hard to read w/o context of being in the room can be found there .
Some of the key ideas is that wee have a relatively large scope (say for example bigger than numpy), but not an infinite one. Most of our work revolve around interactive computation when one of your interlocutor is the computer.
The beginning of the afternoon was spend getting an introduction to binder (past, present, future) about mybinder.org. We talked about cost, scalability, abuses, number of users.
Second tier of the afternoon is spent on discussion of the current governance structure. The current Jupyter Steering council has ~17 members, and sometime ill-defined responsibility and process. As it was created when the team was relatively small there is also what we believe a conflation between being part of the governance body and recognition of contribution to Jupyter. This issue has been long standing (more than a year), and we are only making small progress towards it. We appear to slowly converge but this is a complicated process, wee do not want to throw away all the governance structure that are already in place, but still want to do significant changes that do not give a relatively good balance of power. This was an exhausting debate where choices of words and misunderstanding where common. We are thus still carefully working on the wording an process we want to apply to both empower the community, and make the decision process efficient and transparent.
Afternoon - governance and a steering council refactor
The second half of the day was spent discussing a really important, and potentially
sensitive, conversation: governance. For many years Jupyter has had the same
fundamental governance model - a steering council of particularly-active Jovians.
In recent years it has become apparent that this model isn’t serving the Jupyter
community well, in particular because the community is much more complicated, with
many more stakeholders, than was once the case.
Here’s the starting point, and some notes for this conversation:
In the end, we decided that this was a more complex topic than would fit into one half
day of conversation, and agree to take up the issue again the following morning.
Morning - Steering Council Refactor part 2
A few Jovians spent the previous evening putting together a proposed set of options
for how we would undertake a refactoring of the governance in Jupyter. This session
was spent discussing these ideas, pinpointing strengths/weaknesses/potential failure points, and eventually converging on a plan moving forward. Here are the notes from
tl;dr: in the coming months, a team of Jovians will conduct an open research effort to gather input from the community, and then funnel this into a specific governance model proposal. This will be led by leads on the Jupyter project, who will have editorial roles in the process…their job will be to lead cycles of input and synthesis, but will not do the writing themselves.
Afternoon - appetizers!
The morning session tired everybody out, so we decided not to tackle any more major discussion points for the rest of the day. The team split up into groups to work on appetizers
Morning - Diversity and Inclusion
We spent the morning talking about ways that our community can be more inclusive
of others, particularly of those that are not already represented in the Jupyter community. We began by reading an excellent set of slides on why inclusion is always a good idea for successful organizations, and discussed ways in which we will individually take more steps to create a more inclusive community with Jupyter
Here are some notes, thoughts, and great links from this session:
Afternoon - Maintaining projects and roles in those projects
In the afternoon, we discussed challenges around invisible work and tasks to maintain open projects. Jupyter has a lot of repositories, and each of these requires some degree of care in order to grow a community around them and ensure that good work (technical or otherwise) is being done. However, many kinds of work isn’t obviously visible or quantifiable. This was an attempt at teasing apart some of these questions in the hopes that we can recognize and reward some of this work, and in the hopes that we can create more obvious ways for people to intersect with the community:
Morning - Communications
On the final day, we began by discussing Communications channels across the project. Jupyter has several ways of communicating (both internally and externally), ranging from google groups, to gitter, to github, and now a public Discourse page. We discussed what we want out of communication in the Jupyter community, how we can make communication channels a bit more focused and less-confusing, and whether we should make any changes to our current communications structure.
Here’s a set of notes from the conversation:
Morning pt. 2: the meta-JEP process
We also had a short discussion around a recent push to improve the Jupyter Enhancement Proposal (JEP) process. Some ongoing work has been done to
describe better ways that we can collect input and synthesize a plan for various
large and cross-cutting decisions in the community. This intersected with the
other morning session because we also don’t have great channels of communication
in order to solicit feedback and raise awareness of the JEP process. We agreed that the
community should find ways to make JEPs more discoverable and inclusive to others
of the community.
Here are some notes from this discussion:
Afternoon - Appetizers and goodbyes
In the afternoon, the team was a bit exhausted from a week of intense conversation, and also had many flights to catch. We broke for discussion around appetizers, and didn’t cover any major topics.