The Hegemony of Widgets: Schema.org and the Enclosing of RDF

• In Event: Schemas, Graphs, Ontologies - III

Abstract

RDF (the Resource Description Framework) is a system for making statements about objects and their relations on the Web. In order to connect different groups of statements together, it is necessary to have shared definitions of types, attributes, and relations. These shared definitions are called “vocabularies” or “ontologies.” Conventionally, RDF vocabularies have either been as broad as possible (“title” or “creator” in Dublin Core), or deeply specific to a particular domain (“leaf whorl” or “tuber auxiliary shoot” in Plant Ontology). Those publishing RDF were expected to mix and match vocabularies as appropriate for their domain.

In the past decade, the Schema.org vocabulary has taken a different approach. Developed by search engine companies, Schema.org is hugely variant in scope and specificity, with terms as disjoint as “Concept”, “TennisComplex”, and “DrugStrength.” What these terms share in common is their utility in populating “info box” or “card” widgets in search engines and social media.

I will analyze a specific case where a group has used Schema.org to publish their data on the Web: OCLC, maintainer of the WorldCat union catalog, which migrated to the vocabulary in 2012. Starting from this specific event, I will survey the infrastructural landscape that has supported the popularization of Schema.org. In my analysis, I ask the questions: Why would a library catalog use a vocabulary designed for enriching search results and social media posts? What effect does committing to Schema.org have on the descriptive power at hand for describing information? How might RDF be used differently?

Author

• Patrick Golden, University of North Carolina at Chapel Hill