License Linked Data Resources Pattern

Abstract

Linked Data resources can be referenced by rights expressions or access control policies. Based on the common model found in six existing rights expression languages and revolving around the n-ary relation pattern, the License Linked Data Resources pattern is presented as a solution to describe existing licenses and ad-hoc rights expressions alike and valid for open and not open scenarios.

Keywords

• Ontology design patterns
• Linked Data
• rights expressions

Introduction

Linked Data (LD) assets (RDF triples, graphs, datasets, mappings...) can be object of protection by the intellectual property (IP) law, the database law or its access or publication be restricted by other legal reasons (personal data protection, security reasons, etc.) [ 1 ]. Publishing a rights expression along with the digital asset, allows the rightsholder waiving some or all of the IP and database rights (leaving the work in the public domain), permitting some operations if certain conditions are satisfied (like giving attribution to the author) or simply reminding the audience that some rights are reserved. Additionally, LD resources can be conditionally available after the evaluation of access control policies [ 2 ], expressing who can act what actions under which circumstances.

After the comparison in [ 3 ] of six important rights expressions and policy languages (ODRL, MPEG-21 REL, XACML, ccREL, MPEG-21 MVCO and WAC), enough commonalities were found to extract a common underlying model, which could satisfy all of them. Based on that model, this poster paper describes a content ontology design pattern, named License Linked Data Resources (LLDR), to model licensing issues over Linked Data resources. 2.1

Pattern description

Intent and requirements The intent of the content pattern Licence Linked Data Resources is to represent the relation that exists among a rights expression, an action, an agent, a LD ⋆ This research is supported by the Spanish Ministry of Science and Innovation through a Juan de la Cierva fellowship and the project BabelData (TIN2010-17550). resource and a condition. In particular, the core idea of the pattern is to model: a rights expression which allows/prohibits/obliges to make an Action (Right) to an Agent over a LD resource under a condition.

The LLDR pattern is commited to satisfy the following requirements: – To keep the structure present in other Rights Expression Languages – To be able to represent existing known licenses (Creative Commons licenses...) – To support database rights: extraction and re-utilization – To support privacy law (personal data handling) and the right to access – To support IP law rights: reproduction, distribution, and transformation – To support these right declarations: unconditionally waiving rights, restating that some rights are reserved, and licensing rights subject to conditions – To support existing licensing practices for RDF resources – To support these business models: open data business models, non open data business models, and hybrid models.

cc:License

lldr:Contract lldr:RightsExpression

cc:Requirement cc:Permission cc:Prohibition dcterms:license lldr:hasObject lldr:LinkedDataResource lldr:hasSubject lldr:hasRight The most convenient way to represent the information described before is to use the so-called n-ary relation pattern, which addresses these situations: “(a) a binary relationship that really needs a further argument; (b) two binary relationships that always go together and should be represented as one n-ary relation; and (c) a relationship that is really amongst several things” ([ 4 ]). One of the proposed patterns for representing n-ary relations consists of introducing a new class for the relation and links to all the participants in the relation. Indeed, the LLDR content pattern is inspired on the third consideration shown in the description of n-ary relations from the W3C Semantic Web Best Practices Group Prefix

Namespace void http://www.w3.org/TR/void/ cc http://creativecommons.org/ns# foaf http://xmlns.com/foaf/0.1/ acl http://www.w3.org/ns/auth/acl gr http://purl.org/goodrelations/v1 dcterms http://purl.org/dc/terms/

Table 1. Prefixes and namespaces in [ 4 ]. Fig. 1 shows the LLDR pattern, where the core elements of the n-ary relation are grayed, Table 1 shows the namespace of some used vocabularies.

All the relations revolve around the lldr:RightsExpression element, this class being the qualified relation1. This class has the direct relations with the lldr:LinkedDataResource (a superclass declared to embrace the LD information units of rdf:Statement, void:Dataset and void:Linksets), the lldr:LinkedDataRight, the cc:Requirement and foaf:Agent. Prohibitions, permissions and requirements are rights expressions themselves. The LinkedDataRights is a superclass representing the applicable rights to Linked Data resources: IP rights (cc:Reproduction, cc:Distribution and cc:DerivativeWorks), database rights (lldr:Extraction and lldr:Reutilization) and the mere access (acl:Access).

Rights expressions appear naturally in groups and not separately. Common licenses and typical authorizations are actually aggregations of atomic rights expressions. For this reason, cc:License and lldr:Contract are both subclasses and containers of rights expressions. The aggregation relationship can be represented in OWL using a partOf-whole relation pattern , and consequently a partOf object property has been declared. As a final requirement, resources must be directly linkable to licenses, as this is a common practice already in use (through a dcterms:license or a cc:license property), and in which case the rights expression does not need to include a specific resource. 3

Pattern Usage Example

In the following examples, classes are represented with boxes, relations with arrows and individuals with ellipses. Fig. 2 represents how a RDF dataset is attributed a known license. This example is currently in use by the Linked Data community (albeit not massively). Fig. 3 represents how the right of extraction (copying a database) is waived exclusively to myAgent. Fig. 4 represents how the access to an important reificated statement (the forecast for the stock market price of Google) is offered for 100 e, using the GoodRelations vocabulary. 1 http://patterns.dataincubator.org/book/qualified-relation.html void:Dataset myRDFDaataset dcterms:license lldr:RightsExpression

a myDeonticExpression lldr:hasObject lldr:hasRight lldr:hasSubject rdf:Statement

a lldr:hasObject _:myStatement myOffer rdf:subject rdf:predicate GexO:OEsGtimatedStockPrice myRDFDataset lldr:extraction myAgent a a a gr:PriceSpecification

a hasCurrency EUR myPrice hasCurrencyValue100

Summary and Outlook

The content pattern Licence Linked Data Resources (LLDR) provides a mechanism to represent rights expressions to be applied for Linked Data resources. Having been recently published2, the immediate goal is declaring its relationships with the ODRLv2 ontology and studying the compatibility with the key elements of other relevant vocabularies like LiMO, L4LOD or ODRS3. 2 http://oeg-dev.dia.fi.upm.es/licensius/static/lldr.html 3 http://data.opendataday.it/LiMo, http://ns.inria.fr/l4lod/v2/l4lod v2.htm and http://schema.theodi.org/odrs/ respectively

References

1. Rodr ´ıguez- Doncel , V. , Go´mez-P´erez, A., and Mihindukulasooriya , N.: Rights declaration in Linked Data . in Proc. of the 3rd Int. Workshop on Consuming Linked Data . ( 2013 )

2. Costabello , L. , Villata , S. , Rocha , O. R. and Gandon , F. Access Control for HTTP Operations on Linked Data . In The Semantic Web: Semantics and Big Data (pp. 185 - 199 ). Springer Berlin Heidelberg ( 2013 )

3. Rodr ´ıguez- Doncel , V. , Su´arez- Figueroa , M. C. , Go´ mez-P´erez, A., and Poveda , M. : Licensing Patterns for Linked Data . in Proc. of the 4th Int. Workshop on Ontology Patterns (to appear) ( 2013 )

4. Noy , N. , Rector , A. , Hayes , P. , and Welty , C. : Defining n -ary relations on the semantic web . W3C Working Group Note ( 2006 )