1 of 15

What is Research Software?

And why is it critical to the research endeavour?

Professor Carole Goble CBE FREng FBCS

Software Sustainability Institute UK

ELIXIR, ELIXIR-UK Head of Node

The University of Manchester, UK

carole.goble@manchester.ac.uk

2nd International Funders Workshop:

The Future of Research Software 18-20 Sept 2023

Montreal, Canada (hybrid)

2 of 15

Acknowledgement: Dan Katz

https://doi.org/10.5281/zenodo.6677821

Dan Katz

3 of 15

Software is critical to every step of the research endeavour just like it is critical to every other endeavour in our lives.

https://research.com/research/definition-of-academic-research

4 of 15

Software is critical to every step of the research endeavour��

often

team science�data driven �hypothesis driven�exploratory�pooled curated knowledge

5 of 15

The research community relies on software

S. Hettrick; https://www.software.ac.uk/blog/2016-09-12-its-impossibleconduct-research-without-software-say-7-out-10-uk-researchers

S.J. Hettrick, et al,; 10.5281/zenodo.14809

U. Nangia and D. S. Katz; 10.6084/m9.figshare.5328442.v1

J.E. Hannay et al., “How Do Scientists Develop and Use Scientific Software?” Proc. ICSE Workshop Software Eng. for Computational Science and Eng., 2009, pp. 1–8.

say scientific software is important for their own research

>90%

say they would not be able to do their research without software

~65%

say they develop software as part of their research

~50%

6 of 15

Software is crucial to the Research Endeavour

  • ~20% of NSF projects over 11 years discuss software in abstracts ($10b)
  • 2/3 main DOE ECP areas are research software (~$4b)
  • $300m of FY2021 NIH projects include “software development”
  • £840m UK Investment in 2013-2014

Collected from http://www.dia2.org in 2017

Collected from https://reporter.nih.gov in 2022

SSI Analysis of data from 49,650 grant titles and abstracts published on Gateway to Research covering 2010-2014

Funding

  • Software intensive projects are a majority of current publications
  • Most-cited papers are methods and software
  • 65% of UK university preprints paper reference software (2017)
  • PLoS 40-50% articles generate code

Nangia and Katz; 10.1109/eScience.2017.78

“Top 100-cited papers of all time,” Nature, 2014

10.1038/514550a, PLoS: https://doi.org/10.6084/m9.figshare.21687686.v3

Publishing

7 of 15

What is Research Software? An intent perspectiveGruenpeter et al., “Defining Research Software: a controversial discussion,” 2021�https://doi.org/10.5281/zenodo.5504016

Research Software

created during the research process or for a research purpose

source code files, algorithms, scripts, computational workflows and executables, applications, tools, packages, modules etc.

Warning: Differentiation may vary between disciplines

Software in Research

used for research but not created during or with research intent

software components including operating systems, libraries, utilities, dependencies, packages, scripts, frameworks etc.

8 of 15

What is Research Software? A role perspectiveRob van Nieuwpoort and Dan Katz, “Defining the roles of research software,” �https://doi.org/10.54900/9akm9y5-5ject5y from 1st Funders Workshop, 2022

assembles or integrates existing components into a working whole

infrastructure or an underlying tool

a component of instruments

including pipelines

is the instrument

including curation

facilitates distinctively research-oriented collaboration

including knowledge resources

analyses research data

presents research results

Roles of research software

9 of 15

What is Research Software? An owner perspective Warrick Ball, University of Birmingham. Astronomers re-invent small programs instead of collaborate to sustain and improve big ones�

it is easier and more rewarding

to create one’s own new software

than to extend existing software

RSECon 2023 https://docs.google.com/presentation/d/1TkuHc08arvO9aTXH8waI6cL6PIrVSsmg8N_qbj0v9g/edit#slide=id.g276814a6923_0_10

10 of 15

What is Research Software? A maturity perspective Tom Honeyman, ARDC

Analysis Code

one-off “me” research,

side effect software

Prototype Tools

research need “professorware”

Research Software Infrastructure

professionalised product

Researchers who code

Researcher Developer

Research Software Engineer

Researchers

transition

11 of 15

What is Research Software? A stack perspectiveDependencies, Roles, Visibility�

https://xkcd.com/2347/

User facing shiny thing

Applications, tools, scripts …

Discipline or project specific

Immediate visibility

Underware

Scientific & non-scientific infrastructure

Some discipline specific, some general

Platforms, libraries, utilities, frameworks…

Cross-domain generic reusability

Overly familiar, invisibility

12 of 15

What is Research Software? A stack perspectiveDependencies, Roles, Visibility�

https://xkcd.com/2347/

User facing Software

Applications, tools, scripts …

Domain or project specific

Immediate visibility

Underpinning Software

Scientific & non-scientific infrastructure

Some discipline specific, some general

Platforms, libraries, utilities, frameworks…

Cross-domain generic reusability

Overly familiar, invisibility

13 of 15

Software needs active nurturing to be sustained�like machines/puppies, software will break/die if not maintained

All levels of software stack - stack collapse!

https://xkcd.com/2347/

Corrective -fixing faults

Preventative - increasing maintainability

Adaptive - adapting to changes in environment, porting, new versions of underlying layers

Perfective - meeting new/different user requirements and achieve new capabilities

Keeping the Show on the Road

Dealing with change

Michael Feathers Working Effectively with Legacy Code, 2004

Supportive – user and contributing developer support

14 of 15

People make, use and resource software�They are at the heart of software sustainability

Domain researchers are often not software experts

Software development, maintenance and reuse is not rewarded and is even penalised

Researchers, funders: Can’t pay, won’t pay, don’t pay

Community Open Source is not Free labour

Underpinning infrastructure & seamless integration is under valued in favour of innovation

All software needs dedicated supported core and can’t rely on volunteers.

15 of 15

Summary – Research Software

Software is critical to the research endeavour

What Research Software is has many perspectives: Intent, Role, Maturity, Owner, Stack …. more!

Sustainability is hard but essential, especially for underpinning software.

Needs resources and mixed, multiple streams: direct (money) and indirect (community, reward, recognition, open source support etc).