10 things I have learned from the Software Sustainability Institute
By Dr Robyn Grant (Institute Fellow 2013-14),
Senior Lecturer in Comparative Physiology & Behaviour, Department of Natural Sciences, Manchester Metropolitan University.
The Software Sustainability Institute turned 10 years old this year. Over this time, it has really gone from strength to strength, with established workshops, fellowship schemes and training initiatives. It has continued to grow in influence and visibility. The ethos of the Institute is to cultivate better, more sustainable, research software to enable world-class research. It’s a noble cause, and one that I got involved in seven years ago, when I became an Institute Fellow. Here are 10 things that I have learned from the Institute.
It has been seven years since the start of my Fellowship. My background, especially at the time, was in animal behaviour and neuroscience. Although I worked in a computational research group, I had never developed any software, nor ever thought about what has to be done to ensure that software could be used beyond the end of individual research projects. This all changed when I became a Fellow. I now have the confidence to run behavioural software projects, I have published in software journals and conduct lots of methods-based work. This has really only been possible because of the support, knowledge and recommendations of the Institute. They have been extremely influential in my research career since my Fellowship. I was also one of the few Fellows in my cohort that was a software user, rather than a developer, so I definitely learned a lot from the Institute. Let me tell you about it…
What I do now
The Institute trains and encourages better practices in software development and use. As well as making software open source and available, it is also about making flexible and modular architectures, so that code can easily be updated and applied to new situations.
Since my Fellowship, I led the development of three animal trackers for my research – the Manual Whisker Annotator, the Automated Rodent Tracker, and the Automated Rodent Tracker version 2. These trackers can all integrate together and build on each other, so I can easily add other competencies (or modules), such as whisker tracking and foot tracking.
As well as making this software open and available on Github for others to use, test and develop, we have also published in the Institute journal (Journal of Open Research Software) as well as in the Journal of Neuroscience Methods. This increases the visibility of our work, and encourages others to use it too. It also means that software developers can have first author papers and software citations too, as they can often get overlooked otherwise.
Of course, these practices are not just limited to software development, but also apply to open science more generally, such as big data sets and specialist hardware systems. Indeed, I try to make all my research more open and publish raw code, data and hardware information alongside my publications. With the Institute’s support, I have learned to:
1. Develop open software
2. Develop flexible software
3. Publish, cite and support new software and software developers
4. Do open science
What I know now
However, I would not say that my learning stops here. The Institute proudly states that the Fellowship is lifelong and I continue to work alongside, and with support from the Institute - especially the branch at Manchester University, which is just down the road from Manchester Metropolitan University, where I am based. The Institute continues to support my events, communicate with me via emails and I am always welcome at their courses, workshops and conferences. I especially recommend the Collaborations Workshop, which occurs every year around springtime, and has a fantastic mix of presentations and discussion groups. Don’t be shy, just get stuck in with the Institute - the more you do, the more you get out of it and the more you learn!
And it is not just the Institute who are like this, I have worked with Open Behavior too, who support open software and hardware solutions for behavioural sciences. The deep learning behavioural pose software group – DeepLabCut are also super supportive with a great online community. I cannot stress enough how good it is to be part of a community developing robust, sustainable methods, practices and open science.
So, the overall ethos of the Institute is really important, and I have learned that:
5. The Fellowship is lifelong
6. The Institute is supportive
7. The Institute is a community
8. The open science and software communities that I have met are all lovely and supportive!
What I now try to influence
It is also really important to give back to this supportive community, by spreading their ethos at conferences, in papers and in your own practice. Lead by example.
I ran the International Measuring Behavior conference in 2018 in Manchester, which I organised alongside the commercial software company Noldus. The Institute ran a discussion session on open software at the conference, which was an amazing achievement at a conference, and within a research field, which is traditionally so commercially-focussed.
I, and many of my co-workers at Manchester Metropolitan University, are also trying to get lots of our teaching software replaced with free or open source alternatives, which means our graduates will have access to this software when they leave. This includes using r for statistics, BORIS for animal behaviour recording and QGIS for mapping analysis.
We are also developing software development practices and research software positions at the university, including some specialist positions in Applied Image Engineering, which has produced great code and software across the Faculty of Science and Engineering.
The Institute advocates for culture change within their communities and institutions. So, if you are really in agreement with what the Institute is doing, do try to have influence in your own research field and within your own institutions:
9. Keep telling people about open science and software
10. Influence your research field and institution
I have thoroughly enjoyed working with the Institute. They have taught me almost everything I know about open science and software. I hope that these pointers will help you to support them more too!