Linked data practice at different levels of semantic precision: The perspective of libraries, archives and museums

Abstract

EDITOR'S SUMMARY Libraries, archives and museums rely on structured schemas and vocabularies to indicate classes in which a resource may belong. In the context of linked data, key organizational components are the RDF data model, element schemas and value vocabularies, with simple ontologies having minimally defined classes and properties in order to facilitate reuse and interoperability. Simplicity over formal semantics is a tenet of the open‐world assumption underlying ontology languages central to the Semantic Web, but the result is a lack of constraints, data quality checks and validation capacity. Inconsistent use of vocabularies and ontologies that do not follow formal semantics rules and logical concept hierarchies further complicate the use of Semantic Web technologies. The Simple Knowledge Organization System (SKOS) helps make existing value vocabularies available in the linked data environment, but it exchanges precision for simplicity. Incompatibilities between simple organized vocabularies, Resource Description Framework Schemas and OWL ontologies and even basic notions of subjects and concepts prevent smooth translations and challenge the conversion of cultural institutions' unique legacy vocabularies for linked data. Adopting the linked data vision requires accepting loose semantic interpretations. To avoid semantic inconsistencies and illogical results, cultural organizations following the linked data path must be careful to choose the level of semantics that best suits their domain and needs.

The support of constructs in thesaurus tools from a Semantic Web perspective: Framework to assess standard conformance

“…Even though serious efforts have been made to obtain ontologies from thesauri and other KOS [19], there are relevant differences between ontologies and thesauri [20,21]. These differences are enough to justify the existence of a Semantic Web standard, SKOS, specifically designed to represent KOS in the Semantic Web context [3]. One relevant difference, enough to support this argument, is provided by the ISO 25964 standard: ontologies deal with classes, which have instances (individuals), while thesauri deal with concepts 11 .…”

“…Thesauri add to information systems the ability to offer navigation through a thesaurus' categories or to retrieve documents that have been classified under one of its categories. Moreover, the Semantic Web and linked data have brought about a renewed interest in thesauri as conceptual tools that can be used to improve semantic interoperability [1][2][3][4]. Their potential as support for concept-based searches, i.e., searches in which the user looks for concepts instead of terms in document content [1,4,5] is a valuable feature in the Semantic Web context.…”

Thesauri and Semantic Web: Discussion of the Evolution of Thesauri Toward Their Integration With the Semantic Web

“…With the evolution from ISO 2788 to ISO 25964 standard, the thesaurus community embraces the Semantic Web approach in which thesauri are shared and reused [7], [17], [23]. Moreover, work on tools that support the collaborative edition and maintenance of thesauri and other KOS is still progressing.…”

“…These differences are enough to justify the existence of a Semantic Web standard, SKOS, specifically designed to represent KOS in the Semantic Web context [7]. One relevant difference, sufficient to support this argument, is provided by the ISO 25964 standard (Part II, chapter 21) [12]: ontologies deal with classes, which have instances (individuals).…”

“…Elements and structures in ISO 25964 and SKOS are reviewed and compared. The comparison of ISO 25964 and SKOS considers previous studies of this issue [13]- [15] 7 , but here the way SKOS has been applied to some real thesauri is taken into account. Three thesauri are chosen for this aim: AGROVOC, EuroVoc and the UNESCO Thesaurus.…”