On the Tension between Agile Methodologies and Research Software

Over the last decades, software professionals have embraced agile development as the de-facto way of building software, but these practices have not seen widespread adoption in academia. In this talk, we will look at the reasons for this lack of adoption and how agile practices may be better adopted and adapted for the academic context. Practices like pair programming, test-driven development, and daily standups are part of the toolkit of every software engineer in industry. However, the adoption of these practices in research contexts is limited. The research community recognises this issue, and there are multiple community efforts to improve engineering practices within research software, with initiatives like The Turing Way and The Carpentries. Using our personal experiences –-both in industry and academia–- we will explain the unpopularity of agile in academia. We propose there is a tension between agile practices and the academic environment. For example, pair programming is incompatible with doctoral projects where a single student handles most of the coding; and daily standups are challenging with collaborators from different departments or universities. We will discuss: The range of agile practices available. How industry and academic contexts differ. How well different practices work in different contexts. How agile practices might be adapted to various academic contexts. We believe agile practices need to be translated to properly work in an academic context. In this talk, we describe our successful and unsuccessful attempts to develop academic-compatible agile practices, and discuss what we learned. A recording of this session is available on YouTube: https://youtu.be/wwNVdZpUEB8