The /DH.arc Vocabularies repository provides access to semantic artefacts, such as controlled vocabularies and ontologies, created by the Digital Humanities Advanced Research Centre at the University of Bologna. Semantic artefacts hold a special importance as products of research in that they are both key to achieving FAIR-ness in the data being produced by research and are themselves research objects that should be FAIR (Cox et al., 2021), yet the latter is often less considered than the former (Poveda-Villalón et al., 2020). As such, the primary goal of the repository is to make it easier for semantic artefacts produced by /DH.arc to be FAIR and thus more accessible and visible both within and outside of research communities.

The repository runs on Skosmos, a web based open-source ontology browser developed and released by the team at the National Library of Finland, which uses SKOS as the underlying data model.

For controlled vocabularies, integration with Skosmos is fairly straightforward as SKOS is a standard for developing and publishing machine-readable and interoperable vocabularies.

However, ontologies require some additional work to integrate as they are most often based on OWL and Skosmos requires SKOS in order to correctly display files in its browser. To solve this issue we took inspiration from a 2008 W3C document (which predates the 2009 SKOS recommendation and is connected to Jupp et al.) that indicates four patterns for working with SKOS and OWL, including going from more (OWL) to less (SKOS) formal by either overlaying the two languages or transforming one into the other. We settled on overlaying, whereby triples are merged, as this allowed us to make our ontologies viewable in the Skosmos browser (which requires SKOS classes and properties in order to display the underlying RDF triples) while ensuring that the formal semantics they offer are still understandable to both humans and machines. As noted by the aforementioned W3C document, this approach implies two key elements: the resulting ontologies are necessarily OWL Full (rather than OWL DL) and skos:Concept and owl:Class cannot be disjoint.

Each class and property in the original ontologies are given the additional skos:Concept class and a skos:prefLabel (a duplicate of rdfs:label) to enable their browsing. The definitions of OWL specific properties, such as rdfs:domain and rdfs:range, are added to the file in order for them to also be displayed in the front-end while hierarchical implications, such as the use of rdfs:subClassOf, are duplicated using SKOS semantic relation properties, such as skos:broader and skos:narrower, which makes these more evident than in OWL. Lastly, the ontology is defined as both an owl:Ontology and skos:ConceptScheme while concept scheme-specific properties, such as skos:hasTopConcept and skos:inScheme, are included at the ontology and class levels to align the structure to a more traditional KOS and display it correctly in the browser. All ontologies also include a link to their original OWL version.

All files can be easily browsed (desktop preferred) using a visualisation of the concept hierarchies or alphabetical index as well as downloaded in RDF formats. The search interface enables users to search terms in specific vocabularies or all available. Lastly, there is also a REST API with full documentation to access all the data in /DH.arc Vocabularies.

You can view the latest build of the repository on our GitHub.

Issues can be raised via GitHub and any inquiries and suggestions can be sent to dharc@live.unibo.it.

References:
Cox, S. J. D., Gonzalez-Beltran, A. N., Magagna, B., & Marinescu, M.-C. (2021). Ten simple rules for making a vocabulary FAIR. PLoS Computational Biology, 17(6), e1009041. https://doi.org/10.1371/journal.pcbi.1009041.
Poveda-Villalón, M., Espinoza-Arias, P., Garijo, D., & Corcho, O. (2020). Coming to terms with FAIR ontologies. In Lecture Notes in Computer Science (pp. 255–270). Springer International Publishing.
Jupp, S., Bechhofer, S., & Stevens, R. (2008). SKOS with OWL: Don’t be Full-ish! https://ceur-ws.org/Vol-432/owled2008eu_submission_22.pdf.