Validating and describing linked data portals using shapes (1701.08924v1)

Abstract: Linked data portals need to be able to advertise and describe the structure of their content. A sufficiently expressive and intuitive schema language will allow portals to communicate these structures. Validation tools will aid in the publication and maintenance of linked data and increase their quality. Two schema language proposals have recently emerged for describing the structures of RDF graphs: Shape Expressions (ShEx) and Shapes Constraint Language (SHACL). In this paper we describe how these formalisms can be used in the development of a linked data portal to describe and validate its contents. As a use case, we specify a data model inspired by the WebIndex data model, a medium size linked data portal, using both ShEx and SHACL, and we propose a benchmark that can generate compliant test data structures of any size. We then perform some preliminary experiments showing performance of one validation engine based on ShEx.