eScience 2018 WSSPE 6.1 Workshop

Here are my notes from WSSPE 6.1 and eScience 2018

Daniel S Katz:

Anyone can use the WSSPE name to organize vents
Created WSSPE 6 years ago

Maria J. Crus: Data stewardship

A lot of people don't think they produce any data, or code
People think GitHub is sufficient for archiving code
People share data but forget to share the software that made it
Research software and data are intrinsically linked, but tension:
- Data must be kept as in
- Software must be maintained
Policy (at least in Europe) is focused on data not software
They've introduced the concept of a Data Steward
- Actively seek out researchers to help
- Not the same as an RSE
Reproducibility crisis:
92% of academic use research software, 56% develop their own software
35% use dedicated software for research software management of which 70% use git
Results from their workshop
- Just carefully documenting process in insufficient for reproducibility
- Researchers need to work with RSEs to learn new methods
- Lack of funding for RSEs is a barrier
- Technical recommendations can be sound but too hard for researchers to use
Reproducibility goal: linked and executable code and data
- Caveat: reproducibility does not guarantee good research
This Netherlands (4TU.Centre for research data) group funded The Carpentries workshops as a practical way to help
Data stewards have a role in helping RSEs and researchers
Partnerships required within and across institutions (informal as well as formal) Culture change requires top down and bottom up efforts

There was an interesting workshop idea: speed blogging. Problem: when collaborating as a group and writing notes on paper those notes run the risk of being lost or losing context/discussion. Instead: the collaborative phase is scribed by multiple people and then afterwards everyone chooses a topic to turn into a paragraph in a shared google docs document. Later on, an editor will help these groups create and publish their draft - with a DOI attached.

Stephen Druskat: Mapping the research software sustainability space

- RSS: a lot of players already
- All actors conduct research and publish within space
    ○ Informs policy, education, hiring practices
- Efforts are uncoordinated (not necessarily bad)
    ○ But also has risks, e.g. Duplication, neglect of areas
- Solution: map all RSS activities and classify them
- Version: 1

- Version 2 & 3:
    ○ People split into specific groups (e.g. RSE, R, E)
    ○ Distinction between research software, supporting software
    ○ Sometimes Researchers and RSEs can be the same person

https://github.com/research-software/resosuma-data

J Cohen: Building a sustainable structure for research software engineering activities

- RSE as a term is recent
- Software still seen as a secondary output
    ○ Yet becoming increasingly important
- RSEs can be academics, researchers, developers, or "RSEs" 
- Unique and valuable skill sets - time consuming to maintain
- RSEngineering is not just about RSEs
- For holistic support of RSE:
- Four pillars that support RSEngineering: (call to community)
    ○ Software Engineering
    ○ Community - networking, sharing ideas, knowledge exchange
    ○ Training - base RSE skills & developing the next gen of RSEs
    ○ Policy
- WIP - seeking discussion
    ○ Want to encapsulate RSE (processes, topics, activities) 
    ○ Gain stronger understanding of diff roles in RSE ecosystem
- Remaining questions:
    ○ Anything fundamental missing from pillars? Is naming appropriate?
    ○ How will different roles identify differently with pillars?
    ○ Get strong specification and write paper

https://github.com/jcohen02/rse-models

A slide from my talk:

https://twitter.com/npch/status/1057229839827193856

They did a live survey of the room using www.mentimeter.com on the last day. They asked who were RSEs:

I thought it was interesting to see at least 20 RSEs, especially so considering the crowd at the conference had shrunk quite a bit by the last day.

rse-aunz / rse-au

eScience 2018 WSSPE 6.1 Workshop #48