CapiTainS: challenges for the generalization and adoption of open source software

Thibault Clérice (Sorbonne) et al.

Digital Classicist London seminar 2018

Friday June 8th at 16:30, in room 234, Senate House, Malet Street, London WC1E 7HU

Livecast at Digital Classicist London YouTube channel.

The challenge of code generalization is not specific to digital humanities but affects the digital world in general : determining where abstractions should stop is a common struggle in software development.. The tension between compatibility with potential, unforeseen evolution of projects and customization for the current task is one that a project aiming for reuse faces in every technical decision, including questions of scalability and required computational power.

In this context, we would like to present CapiTainS, an institution-independent, open-source Github organization. Initially aimed at making the Canonical Text Service more usable, it contains a set of guidelines for encoding citable TEI texts and metadata, libraries for querying them, a bootstrap application for building a reading environment, API, and software to run continuous integration tests for texts. Some of the software has already undergone 2 major revisions and a new revision of the guidelines is coming in the next months.

In this talk, we will focus specifically on two problems that CapiTainS and, we believe, most other small, open-source projects face: expansion of the code base to deal with new types of data and ensuring involvement within the original scholarly community. If CapiTainS is going to survive the next three years and overcome the vagaries of short-cycle funding, it must expand its user base beyond the original project. In this self-critical analysis, we will consider the mistakes that we made in these two areas, some ideas for how we could have done better, and how we think we should focus our efforts during the next three years.