Research Engineer Jurriaan Spaaks on how the eScience Center’s Software Directory presents another important step towards Open Science.
The road to Open Science faces many obstacles, one of which is to successfully generate and develop software that is reusable. To ease the findability, scientific referencing and reproducibility of research software, the Netherlands eScience Center developed the Research Software Directory (RSD). The RSD is an open-source content management system that makes it easier for researchers to promote and share their software with other parties.
In this interview, Jurriaan Spaaks, eScience Research Engineer and co-developer of the Research Software Directory, discusses the value of the RSD and how it promotes Open Science.
The RSD is a content management system that has been specifically developed to present research software. It is set up in such a way as to avoid duplicating information that is already stored in other repositories such as GitHub, Zenodo and Zotero. Instead, we simply harvest information from these sources through their respective APIs, which means less work for whoever is filling the content management system with data.
Open Science has three pillars. First, public access to research papers is needed so that the description of a study and its outcomes are accessible to everyone. Second, the data underpinning the study should be available, thereby allowing others to assess whether the analysis outlined in the paper works as described. Software constitutes the third pillar, since that’s usually how published results are generated from data. All three pillars are needed to ensure the ideas, analyses and conclusions of a given study are exposed to critical review, and also to enable others to expand on previous work.
For the first two pillars, we have reasonably mature mechanisms in place. For example, we’ve always taken great care of collecting, storing and indexing research papers, and in recent years that care has been extended to also include data. However, compared to data and papers, it is more difficult to collect, store and index software – it requires constant maintenance.
The Research Software Directory promotes Open Science by bringing together the data, the software and the resulting papers in one place. Additionally, it nudges its users toward better software development practices by including software citation information and the use of persistent identifiers such as DOIs to uniquely identify various versions of the research software.
What many people don’t realize about adopting someone else’s software is that it is actually preceded by a few steps, and that missing even one of these will prevent adoption from happening. Search engines are crucial for the findability of software, so we had to design the RSD in such a way that it uses Search Engine Optimization (SEO) techniques.
But even if you do your SEO perfectly, adoption will not happen if visitors to your web page can’t tell if the advertised research software will actually help them address their research questions. In other words, you have to make this clear to visitors. This is why we opted for a simple page layout, where we describe the research software at three levels of detail and present each piece of research software along with its scientific context.
The final and often overlooked step is to provide visitors with a good starting point for using the research software. This can be a tutorial or a video, or even a beginner API supporting only some simple use cases.
We received positive feedback. Most researchers see its value when we show them our own online portfolio, which is powered by the Research Software Directory. Researchers get frustrated when searching for software – the RSD presents relevant information clearly and concisely and as a result helps others to evaluate the usability and contribution of the software for their own research.
Yes, certainly. We went through quite a bit of effort to document the software to such a level that others can run their own copy of the RSD. If you’re a researcher and you want to run your own customized copy of the Research Software Directory, (e.g. to keep track of all the research software that’s written in a given department or research group) we can even help you get started.