Modeling Digital Humanities Collections as Research Objects

Abstract

Advancing digital libraries to increase the sustainability and usefulness of digital scholarship depends on identifying and developing data models capable of representing increasingly complex scholarly products. This paper considers the potential for an emergent model of scientific communication, the research objects data model, to accommodate the complexities of digital humanities collections. Digital humanities collections aggregate and enrich diverse sources of evidence and context, serving simultaneously as "publications" and dynamic, interactive platforms for research. The research objects model is an alternative to traditional formats of publication, facilitating aggregation and description of all of the inputs and outputs of a research process, ranging from datasets to papers to executable code. This model increasingly underpins research infrastructures in some scientific domains, yet its efficacy for representing humanities scholarship, and for undergirding humanities cyberinfrastructure, remains largely untested. This study offers a qualitative content analysis of digital humanities collections relying on a content/context analytical framework for characterizing collection components and their interrelationships. This study then maps those components and relationships into a research objects model to identify the model's strengths and limitations for representing diverse digital humanities scholarship. CCS CONCEPTS • Information systems → Data structures.

Interactivity, Distributed Workflows, and Thick Provenance: A Review of Challenges Confronting Digital Humanities Research Objects

“…This caution resonates with the challenges identified for the adoption of the RO model-or indeed for the importation of any data model, even domainindependent data models-into the humanities. The main challenges to implementing ROs for humanities research also present exciting opportunities for a more sustainable cross-disciplinary infrastructure (Fenlon, 2019), but implementation strategies must be centered in scholarly communities, and grow out from the practices, needs, and epistemologies of specific areas of study in the humanities and cultural institutions.…”

“…Prior empirical work on applying the RO model to digital humanities collections found the main limitation of the model for digital humanities collections to be that functional components, designed for ongoing end-user interaction, are not usefully captured in a basic RO model and instead fall to the implementations built on top of research-object management systems (Fenlon, 2019). ROs can, of course, accommodate as flat code objects that are intended to be interactive; and ROs have been employed for this purpose to support data migration and archiving (e.g., the RO BagIt profile).…”

“…ROs and similar models have substantial potential to underpin systems that support a variety of implementations. Realizing the advantages of ROs for the humanities will depend on implementations that create platforms for experimentation and collaborative development by distributed communities (Fenlon, 2019). Such platforms must accommodate dynamic interface-building, to allow scholarly communities with distinctive interests and needs to mobilize ROs in different ways.…”

“…With their capacity for linked data using domain vocabularies, ROs readily accommodate many of the artifacts of networked digital scholarship in the humanities, along with their interrelationships (Fenlon, 2019). But can ROs accommodate humanities workflows in useful ways?…”

“…This paper reviews existing applications of the ROs model to identify challenges confronting the application of ROs to humanities digital scholarship. This paper builds on Fenlon (2019), which investigated the application of the ROs model to digital humanities collections, and which identified three promising strengths of the model for the realm of digital humanities: (1) ROs readily perform the most essential function of a collection: to aggregate related resources in order to support scholarly objectives; (2) ROs have the capacity for explicit, semantic descriptions of interrelationships among components that are often hidden in digital humanities collections (and therefore vulnerable to dissolution); and (3) the RO model accommodates aggregations of linked data, offering researchers the opportunity to create and annotate virtual, fully referential collections.…”