Provenance Vocabulary Core Ontology Specification

14 March 2012

This version:

http://purl.org/net/provenance/ns-20120314

Latest version:

http://purl.org/net/provenance/ns

Previous version:

http://purl.org/net/provenance/ns-20110125

Revision:

0.6

Authors:

Olaf Hartig (Database and Information System Research Group, Department of Computer Science, Humboldt-Universität zu Berlin)

Jun Zhao (Image Bioinformatics Research Group, Department of Zoology, University of Oxford)

Copyright © 2009 - 2012 Olaf Hartig and Jun Zhao.

This work is licensed under a Creative Commons License. This copyright applies to the Provenance Vocabulary Core Ontology Specification and accompanying documentation.

Regarding underlying technology, the Provenance Vocabulary relies heavily on W3C's RDF technology, an open Web standard that can be freely used by anyone.

This visual layout and structure of the specification was adapted from the SIOC Core Ontology Specification edited by Uldis Bojars and John G. Breslin.

Table of contents

• 1. Introduction
• 2. The Provenance Vocabulary Core Ontology at a glance
• 3. Cross-reference for the Provenance Vocabulary Core Ontology classes and properties
  • 3.1. General classes
  • 3.2. General properties
  • 3.3. Data creation classes
  • 3.4. Data creation properties
  • 3.5. Data access classes
  • 3.6. Data access properties
• 4. Design Decisions
• 5. References
• 6. Change log

1. Introduction

The Provenance Vocabulary which is defined as an OWL-DL ontology is partitioned in a core ontology and supplementary modules. To avoid making the core ontology too complex the modules provide less frequently used terms and a broad range of extensions of the core terms. At present the Provenance Vocabulary has three modules: Types, Files, and Integrity Verification.

The vocabulary is designed very closely to the model for Web data provenance as presented in [Har09]. This model comprises two dimensions of Web data provenance: data creation and data access. Accordingly, the Provenance Vocabulary basically consists of three main parts: general terms, terms for data creation, and terms for data access.

Detailed information about using the Provenance Vocabulary and many examples can be found in [HZ09].

The Provenance Vocabulary is a domain specific specialization of the W3C PROV Ontology (PROV-O) developed by the W3C Provenance Working Group; classes and properties provided by the vocabulary are Web data specific extensions of the more general concepts introduced in PROV-O. As a consequence, any Provenance Vocabulary based description of provenance can be easily interpreted and exchanged according to the W3C PROV family of standards.

The XML Namespace URIs that must be used by implementations of this specification are:

• http://purl.org/net/provenance/ns# - Provenance Vocabulary Core Ontology Namespace
• http://purl.org/net/provenance/types# - Provenance Vocabulary Types Module Namespace
• http://purl.org/net/provenance/files# - Provenance Vocabulary Files Module Namespace
• http://purl.org/net/provenance/integrity# - Provenance Vocabulary Integrity Verification Module Namespace

Recommended prefixes for abbreviating these four namespaces are prv:, prvTypes:, prvFiles:, and prvIV:, respectively.

Further prefixes used in this document are:

2. The Provenance Vocabulary Core Ontology at a glance

An alphabetical index of Provenance Vocabulary Core Ontology terms, by class and by property, are given below. All the terms are hyperlinked to their detailed description for quick reference.

Classes: | CreationGuideline | DataAccess | DataCreation | DataItem | DataProvidingService | DataPublisher | File | HumanAgent | Immutable | NonHumanAgent |

Properties: | accessedResource | accessedService | completedAt | containedBy | createdBy | deployedSoftware | operatedBy | performedBy | precededBy | retrievedBy | serializedBy | usedBy | usedData | usedGuideline |

The following figure provides an overview of the classes and properties in the Provenance Vocabulary Core Ontology.

3. Cross-reference for the Provenance Vocabulary Core Ontology classes and properties

The Provenance Vocabulary Core Ontology introduces the following classes and properties.

3.1. General classes

3.2. General properties

3.3. Data creation classes

3.4. Data creation properties

3.5. Data access classes

3.6. Data access properties

4. Design Decisions

Vocabulary design decisions:

1. use owl:disjointWith to explicitly identify disjoint classes
2. don't define inverse properties because they are bad for interoperability (see http://dowhatimean.net/2006/06/an-rdf-design-pattern-inverse-property-labels)
3. rule to decide about the direction of a property (i.e. a link) when both are possible: it should be possible to describe provenance of a data item starting from the data item itself (i.e. the data item becomes subject in triples) and 'working' towards more distant provenance elements. Thus, the links should not be directed towards the data item.

5. References

• [Har09] Olaf Hartig: Provenance Information in the Web of Data. In Proceedings of the Linked Data on the Web (LDOW) Workshop at WWW, Madrid, Spain, April 2009, Download PDF.
• [HZ09] Olaf Hartig and Jun Zhao: Guide to the Provenance Vocabulary. Online at http://purl.org/net/provenance/guide.
• [HZ10] Olaf Hartig and Jun Zhao: Publishing and Consuming Provenance Metadata on the Web of Linked Data. In Proceedings of the 3rd International Provenance and Annotation Workshop (IPAW), Troy, New York, USA, June 2010, Download PDF.

6. Change log

Changes from rev.0.5.1 to rev.0.6

• Made the Provenance Vocabulary a specialization of PROV-O. Deprecated all of the abstract classes (prv:Actor, prv:Execution, prv:Artifact) and abstract properties (prv:yieldedBy, prv:involvedActor, prv:employedArtifact). All classes and properties that used these abstract concepts (as super-class, super-property, domain, or range) now use corresponding PROV-O concepts instead.
• Added the class prv:Immutable.
• Renamed prv:HumanActor to prv:HumanAgent and prv:NonHumanActor to prv:NonHumanAgent. prv:HumanActor and prv:NonHumanActor are deprecated now and will be removed in a later revision.
• Renamed prv:performedAt to prv:completedAt. prv:performedAt is deprecated now and will be removed in a later revision.
• Made prv:completedAt a functional property.
• Made prv:retrievedBy a functional property.
• Made prv:CreationGuideline subclass of prov:Plan (in addition to prv:DataItem).
• Removed the deprecated class prv:Representation.
• Removed the deprecated property prv:encodedBy.

Changes from rev.0.5 to rev.0.5.1

• Removed an inconsistency: the deprecated class prv:Representation was defined to be both, equivalent and disjoint with class prv:DataItem.

Changes from rev.0.4 to rev.0.5

• Removed the deprecated class prv:Document.
• Deprecated the class prv:Representation because the general understanding of prv:DataItem is as a representation of data and the Provenance Vocabulary does not distinguish between representations that are Web representations (what prv:Representation was for) and data representations in general.
• Restricted the range of prv:containedBy to prv:DataItem only; removed prv:Representation.
• Changed the domain of prv:retrievedBy from prv:Representation to prv:Artifact.
• Renamed the property prv:encodedBy to prv:serializedBy because the new name represents more accurately what the property means. For convenience we added a deprecated property prv:encodedBy which is equivalent to prv:serializedBy. This deprecated property will be removed in a later revision.
• Moved the file-related properties prv:usedFile and prv:usedGuidelineFile to the Files module (renamed usedFile to usedDataFile).
• Added the property prv:accessedResource.
• Added cardinality restrictions.

Changes from rev.0.3 to rev.0.4

• Added the property prv:deployedSoftware.

Changes from rev.0.2 to rev.0.3

• Introduced the class prv:File and the properties prv:encodedBy, prv:usedFile and prv:usedGuidelineFile to enable file-based descriptions of data creations; consequently, added prv:File to the domain of prv:createdBy. Furthermore, added OWL2 property chain axioms for prv:createdBy, prv:usedData, and prv:usedGuideline.

Changes from rev.0.1 to rev.0.2

• Renamed the class prv:Document to prv:Representation because the name caused confusion; the name made people assume this class represents information resources which is not the case; in fact, it represents a Web representation of an information resource. For convenience we added a deprecated class prv:Document which is equivalent to prv:Representation. This deprecated class will be removed in a later revision.