W3C PROV

W3C PROV
Name	PROV
Developer	World Wide Web Consortium
Released	2013
Latest release	2013 Recommendation
License	W3C

W3C PROV W3C PROV is a family of specifications for modeling provenance information about digital artifacts, processes, and actors, intended to enable reproducibility, accountability, and trust across systems. It provides a formal data model and serializations to represent how entities, activities, and agents relate, connecting with technologies such as XML, RDF, JSON, SPARQL, and HTTP to integrate provenance into web ecosystems. The work was produced by the World Wide Web Consortium and influences practices in domains from genomics and climate science to journalism and digital forensics.

Introduction

PROV defines concepts to describe provenance: entities, activities, and agents, with relations like wasGeneratedBy, used, and wasAssociatedWith, enabling lineage tracking for artifacts such as datasets, images, and software. It interoperates with web standards like Resource Description Framework and Extensible Markup Language, and aligns with governance and audit frameworks used by institutions including the European Commission, United Nations, World Health Organization, National Institutes of Health, and NASA. The model supports integration with identifiers and registries such as Digital Object Identifier, ORCID, Handle System, DNS, and IETF protocols to link provenance records to persistent identifiers.

History and Development

PROV evolved from research in provenance provenance systems and workflow management in projects involving groups at MIT, Stanford University, University of Southampton, Monash University, University of Oxford, and University of California, Berkeley. Early motivations trace to initiatives like the Taverna workflow system and the Kepler project, and to provenance-related work at laboratories including Los Alamos National Laboratory and Lawrence Berkeley National Laboratory. Standardization was driven by the W3C Provenance Working Group with contributions from organizations such as Google, IBM, Microsoft, Oracle Corporation, Digital Science, Elsevier, Nature Publishing Group, and CrossRef. The Recommendation status in 2013 followed review cycles influenced by input from agencies such as National Science Foundation, European Research Council, Wellcome Trust, and scholarly publishers like Springer Nature.

Core Concepts and Data Model

The PROV Data Model (PROV-DM) formalizes core classes: Entity, Activity, Agent, and derivative classes, along with relationships such as wasDerivedFrom, wasAttributedTo, and actedOnBehalfOf. It encodes provenance graphs compatible with RDF Schema and links to web identity systems like OAuth, OpenID Connect, and registries such as CrossRef and DataCite. PROV supports provenance assertions for scientific workflows executed on platforms like Galaxy Project, CWL-based engines, and cloud infrastructures from Amazon Web Services, Google Cloud Platform, and Microsoft Azure. It accommodates temporal annotations, usage of software including Git, Docker, Kubernetes, and citation practices consistent with Citation Style Language and bibliographic tools like Zotero and EndNote.

Standards and Specifications

The PROV family includes PROV-DM, PROV-O (OWL2 ontology), PROV-N (notation), PROV-XML, and PROV-JSON serializations, integrating with query languages such as SPARQL and APIs defined by W3C recommendations. It interacts with metadata standards and initiatives including Schema.org, Dublin Core, OpenAnnotations, IIIF, WARC, and persistent identifier schemes like ARK and Handle System. PROV influenced and was influenced by standards from bodies like the International Organization for Standardization, ISO, Digital Preservation Coalition, and archival practices in institutions such as the Library of Congress and British Library.

Implementations and Tools

Open-source and commercial implementations implement PROV in languages and frameworks including Apache Jena, RDF4J, Neo4j, PostgreSQL, MongoDB, and libraries for Python, Java, JavaScript, and R. Tools and platforms leveraging PROV include RO-Crate, PROV-Toolbox, ProvStore, workflow managers like Airflow, Nextflow, and reproducibility platforms such as Binder, Jupyter, and Zenodo. Integration plugins exist for content management and editorial systems from WordPress, Drupal, and scholarly infrastructures like Open Journal Systems and Hypothesis.

Use Cases and Applications

PROV is used for reproducible research in fields including genomics, astronomy, climate science, neuroscience, epidemiology, materials science, and computational chemistry. It supports provenance in data repositories like Figshare, Dryad, Zenodo, and institutional repositories at universities such as Harvard University, Stanford University, University of Cambridge, Massachusetts Institute of Technology, and University of Tokyo. Industrial applications include supply chain traceability for companies like Walmart, Amazon, and IBM Food Trust, as well as provenance for news verification used by outlets such as BBC, The New York Times, The Guardian, and agencies including Reuters.

Criticisms and Limitations

Critics point to challenges in adopting PROV at scale: interoperability gaps with legacy systems managed by institutions like European Space Agency and US National Archives and Records Administration, complexity of mapping domain-specific provenance to PROV-DM in projects at CERN or Human Brain Project, and performance concerns for high-volume provenance streams in platforms run by companies such as Twitter and Facebook. Others highlight legal and privacy considerations involving General Data Protection Regulation and archival laws at institutions like National Archives (UK) when recording agent identities. Tooling fragmentation across ecosystems including Node.js, .NET Framework, and Rust ecosystems also complicates widespread adoption.

Category:Provenance