What Does Reproducibility Look Like for DH Projects?
Abstract
While frameworks for open science have long advocated for reproducibility in research, reproducibility within the digital humanities continues to be challenging to achieve. This is particularly salient in the case of DH projects whose primary output is a front-facing web resource, of which all component parts—the data, the code, the interface, and the user experience—are all key to the project’s scholarly contribution. While institutional repositories and other digital infrastructure are well equipped to store and preserve a wide variety of outputs, DH projects tend to be stored in fragments and not in ways that allow future users to “spin up” or “reproduce” the form and function of the project as it was initially created. This presentation outlines our approach at SFU Library's Digital Humanities Innovation Lab to creating sustainable and reproducible DH projects that go beyond archiving of source code and/or data. Drawing on the Endings Principles as well the recent Digits report on containerization as scholarly product, this lightning talk will describe the DHIL’s multi-faceted strategy—including data deposits, static websites, web archives, and multiple generated containers—for sustaining, openly sharing, and reproducing research outputs in the digital humanities. In particular, we will discuss how we have automated much of this process to demonstrate the feasibility of this approach while also outlining future directions and recommendations for institutional repositories and other institutions to better enable the open sharing of DH projects.