# OLS-4 Application (v2)

https://docs.google.com/document/d/1HDhUFCaDUDiQsY3NPhsJ395FFTHaQgZee9Idigzh-_A/edit

## Title

**Current:** *Hub23: An open source community and infrastructure for Turing’s BinderHub*

**New:** *TBC*

## Project

**Current:** *Binderhub is a service that allows users to share reproducible interactive computing environments through public code repositories. The subject of our project, Hub23, is an organisational deployment of Binderhub, designed to allow Turing Researchers to use binder to collaborate on repositories internal to Turing. This is sometimes necessary if the underlying repository can not be shared for some reason, or is not yet ready to publish openly. During the OLS program, we aim to build an open community around Hub23 to help to guide future technical developments, and encourage use and contributions from the wider Turing community. We will host a series of [Zero-to-Binder workshops](http://bit.ly/zero-to-binder-python) aimed at introducing Turing researchers to regular binder, followed by structured discussion of what the ideal features of a collaborative reproducible environment for research would be. Any conclusions and subsequent technical development will be fed upstream to Binderhub, and we also aim to open source the methodologies used to create an internal binderhub deployment, allowing other organisations to do so.*

**New:** *TBC*
- [name=Luke] We discussed the idea of an "Open Source (service area) strategy for REG". This could/would incorporate various ideas that we have had so far, such as:
- Surveying REG & the wider Turing to understand various approaches to "Open Source"
- Defining "Open Source", what it means to REG, our approach to open source, our aims, ...
- Maintaining Hub23, running Zero-to-Binder workshops
- Acting as advocates for open science
- Starting a Turing Interest Group
- **Longer term:** A "strongly opinionated" interpretation of the Turing Way. This is a bit ambitious for a 16 week program, but we could produce the abstract/motivation

## Problem

**Current:** *[MyBinder](https://mybinder.org/) allows researchers to share reproducible computing environments for their research, without requiring detailed knowledge of the underlying technology. However, mybinder is most commonly used for teaching and demonstrations, where setting up a reproducible environment for users is secondary to the goal of teaching a concept (see 19:42 in Sarah Gibson’s 2020 [EuroPython talk](https://www.youtube.com/watch?v=wTVMHallSSA), as opposed to collaborative research. Part of the reason for this is that research often starts in private repositories, only becoming public when a certain stage has been reached. Hub23 is a Binder instance internal to Turing, with the goal of allowing researchers to use binder on private repositories, but there is currently a lack of awareness and engagement with Hub23 amongst the Turing community. Furthermore, by increasing engagement and hosting discussions, this project will help uncover features (e.g. real-time collaboration) that could be added to Hub23 (and upstream to binderhub) to increase its functionality as a collaborative research tool. Finally, by open-sourcing the methodology for deployment of a private BinderHub instance, our project will help to fill the gap in public documentation, and avoid duplication of efforts for other organisations interested in doing the same.*

**New:** *TBC*
- [name=Luke] ="Open Science" is supposed to be important to REG, but we're not entirely sure what this actually means or how we do it. It's unlikely that we are going to produce an exhaustive guide to open science, but we can help to define REG's strategy (and that of the Open Source Service Area)

## Achievements

**Current:** *During this program, we hope to achieve the following:*
*1. Allow Turing employees to use Hub23 to work on private Github repositories.*
*2. Increase awareness and discussion with the Hub23 project (both internal and external to the Turing) by hosting a series zero-to-binder workshops and discussions. By doing so, help foster an open source community and mindset within the Turing (see OLS-3 project on Turing’s Open Source Service Area).*
*3. Encourage use of the Hub23 (& Binderhub) tools within the REG & Turing research communities for reproducible research.*
*4. Investigate/plan for potential improvements to the service, its documentation, and its "openness". Feed these upstream to Binderhub.*
*5. Open source the methodology for deploying a private BinderHub instance*

**New:** *TBC*
- [name=Luke] Some potential deliverables mentioned above, but essentially rather than producing the "strongly opinionated approach" we could "write the intro"