PAV ontology: provenance, authoring and versioning

Area of Science:

• Digital scientific content management
• Knowledge representation and reasoning

Background:

• Establishing trust in scientific content relies heavily on provenance tracking.
• Existing general-purpose vocabularies like Dublin Core Terms (DC Terms) and W3C Provenance Ontology (PROV-O) lack specific classes for agent roles (author, contributor, curator) in digital artifact manipulation.
• PROV-O offers a basic methodology for web resource authoring and versioning but requires extensions for detailed role identification.

Purpose of the Study:

• To introduce the lightweight Provenance, Authoring and Versioning ontology (PAV) for capturing essential descriptions of web resource provenance, authoring, and versioning.
• To address the need for specific classes and properties to distinguish between various agent roles and resource transformations in scientific digital content.
• To enhance interoperability by providing mappings to existing ontologies like W3C PROV-O.

Main Methods:

• Designed PAV with a focus on being lightweight and pragmatic, incorporating requirements from real-world projects.
• Included only terms demonstrated to be useful in existing applications.
• Recommended terms from existing ontologies where plausible to maintain compactness and encourage reuse.

Main Results:

• The Provenance, Authoring and Versioning ontology (PAV) has been developed to capture essential provenance, authoring, and versioning information for digital scientific content.
• PAV distinguishes between content contributors, authors, curators, and representation creators, alongside the provenance of originating resources.
• Five projects have adopted PAV, demonstrating its practical application and providing concrete examples of its usage.
• Mappings have been developed to show how PAV extends W3C PROV-O, enhancing interoperability.

Conclusions:

• PAV provides a focused and lightweight solution for tracking the provenance, authoring, and versioning of digital scientific content.
• PAV offers a clear distinction between agent roles and resource provenance, addressing limitations in existing general-purpose ontologies.
• Comparisons with related approaches (PRV, DC Terms, BIBFRAME) highlight PAV's strengths in pragmatic applicability and specific role distinctions.
• SKOS mappings facilitate alignment between PAV and DC Terms, further promoting interoperability.