# Pangeo Community Support Brainstorm

Video: zoom.2i2c.org (p: `interact`)

## Ideas for agenda/discussion

- [Product Management vs. Community Management](https://github.com/2i2c-org/team-compass/issues/446)?
- [Rethinking community representative role](https://github.com/2i2c-org/docs/issues/134)
- [Managed service vs. collaborative service](https://github.com/2i2c-org/infrastructure/issues/1473)

## Questions to answer

- High-level questions we need to answer?
- More specific things to try in the short term?

## Background

- tl;dr: Pangeo had some enthusiastic leaders that bit off more than they could chew organizationally.
- 2018
- Intense experimentation and high productivity
- Put data in the cloud, created a bunch of hubs
- The catch: way easier to prototype and stand up once than to run an ongoing service.
- Free access to JupyterHubs and host all of the climate data in the cloud is a huge undertaking given their resources.
- What to do now?
- Keep that enthusiasm and excitement
- But move in the direction of feasability
- And without alienating existing users

### Current Pangeo approach

- Work with partners for various pieces
- Data hosting
- Google Cloud, AWS
- Computing instances
- Microsoft (Planetary Computer)
- Get funding / grants to resource these partnerships
- Grants
- Moore Foundation - Climate Data Science Lab
- M2LInES - Schmidt Futures
- Earthcube - Pangeo Forge
- LEAP - NSF STC
- Each grant says "We'll have a cloud based jupyterhub"
- Plan is to ship these deliverables/resources over to 2i2c
- Challenge
- Each of these grants have separate sub-communities
- They have their own funding source, their own stakeholders, etc.

### 2i2c and Pangeo now

- We've deployed the hubs for Pangeo / M2Lines / LEAP separately
- What we need: user support
- Need more direct support for users on how to use the hub
- Training and one-on-one conversations on how to use the infrastructure
- Challenges to growing this within the community
- People always leaving: Continuity and institutional knowledge
- How to grow connections across 2i2c's communities of practice
- Grants are finite in time
- "Communities" don't have money to spend
- Funded projects and institutions have money
- HPC-style centers have central funding that is often used to pay vendors

### What are the most important "services" the Pangeo leaders provided their community?

- Infrastructure for Open Science
- Make the infrastructure as generic as possible so it's franchiseable and trivial to bring on any new community of people.
- Be able to support a small team with $10k, and a huge team with $1m
- GitLab vs. GitHub
- GitLab - set it up on your own, but you lose cross-repository connections
- GitHub - mono-service
- Federated JupyterHubs?
- How do you build infrastructure that creates bonds between the hubs
- Problem with passive communities
- We need to engage them with more leadership

### First 6 month iteration

- Curate web-based curricula that are already out there.
- Videos tutorials notebooks etc
- Come up with a process for others to use that material in a structured way.
- Monthly fireside chat slots etc
- Ongoing workshops and tutorials etc
- Mechanisms for user interaction and Q/A
- Right now all requests go to Ryan, not to 2i2c
- Not scalable/sustainable for Ryan
- "What we have now is money not a person"
- How to recreate and scale a help desk
- Could you promote people from within communities?

### What Pangeo sub-communities want

- Do they want a bleeding edge environment?
- Kitchen-sink-style hub with the bleeding edge of packages?