# OLS-5 Planning

### Suggested Aim:

What is the best way to use reproducible environments in collaborative research projects?

### Useful Links
- REG Handbook
- JupyterHub Docs
- OLS-5

#### Notes 01/02

- It would be worth talking to Jim to:
a) Get his opinions on reproducible environments, etc.
b) See how this project can fit in with the REG handbook (e.g, possibly write a chapter of the handbook as a project deliverable)
- We still need to speak to Evelina, but last time she suggested REG should have a "strongly opinionated" version of the Turing Way - this sounds similar to goal of REG handbook
- From Sarah:

Hey folks! 

I am not sure if you are intending to resurrect this project with OLS-5 or 
not (no pressure either way!). However, there’s a lot of movement at the minute 
with a couple of new sub-projects:

- https://github.com/yuvipanda/github-app-user-auth 
Facilitate securely pushing back to GitHub from a JupyterHub

- https://github.com/yuvipanda/jupyter-syncthing-proxy 
Use syncthing to sync any folder to JupyterHub, theoretically this could mean a 
user’s home directory is no longer dependent on the cluster/VM that’s hosting 
the JupyterHub

Both of these things make me feel pretty confident that Hub23 the BinderHub should 
be torn down, and Hub23 the JupyterHub should be deployed, with these extensions. 
I think that would make the OLS project’s goal of “getting the Turing community 
to use this kind of infrastructure for their research” a lot easier as these 
are the kinds of things researchers will ask for but aren’t currently provided by 
BinderHub.

I know you’re all pretty busy, but I’d love to have a scoping chat with 
whoever’s interested, maybe sometime next month?

### Meeting Jim 22/02/22

- What are the barriers to open source at REG?
- A template for open science projects? Like a cookiecutter? Template could contain soft advocacy e.g. to binder.
- Chapter in REG handbook? how to bring open science to projects, what makes a good opportunity, what are red flags etc.
- does this link with hub23, if at all?
- even the initial scoping phase is a useful project.
- the reasons why we don't do open source well is because the persons that fund us (research) aren't incentivised for it (this is the same reasons that open science is a challenge). If we cannot circumvent these problems and get caught in the same trap then this is something that we need to look at. Perhaps we could have a more formal way of assessing 'openness' during the project scoping phase.
- hack week. formal top-down allocation is the only way to really get involved in open-source.
- Could we have two aims? Hub23 developement -> figure out how this is used in projects?

## Evelina meeting 25 February

- REG contribution to Open source vs Binder/Hub23 (reproducibility)
- pick one to focus on

- Community outreach + deliverable/proposal for REG handbook would be very useful, achievable, actionable
- e.g., project template

-