Abstract

With the rise of digital publishing and open access it has become increasingly important for publishers to store information about ownership and licensing of published works in a robust way. In a fast moving environment Elsevier, a leading science publisher, recognizes the importance of sound models underlying its data. In this paper, we describe a data model for copyright and licensing used by Elsevier for capturing elements of copyright. We explain some of the rationale behind the model and provide examples of frequently occurring cases in terms of the model.

Introduction

In the past decade with the rise of digital publishing there have been significant changes in the perception of copyright in the publishing industry . For Elsevier as a major academic publisher it has become increasingly important to keep a record of the copyright status and licenses connected to each piece of content. Journal articles form the vast majority of Elsevier’s published content and with the increase of various open access options given to authors this copyright and license landscape has become more and more complex. Licensing in particular plays a role in open access publishing and Elsevier’s policy in adherence to the Creative Commons (CC) licensing approach is that all open access articles must have end-user licenses. The open access options that Elsevier currently provides have not only introduced a variety of licenses for end-users which may allow copying and reuse, but also mean that an article’s copyright and license state may change even after publication. For example, an article published in a subscription-based journal may end up in an issue that is funded by an external third party to become open access, or an author may later opt for his article to become open access after it has been published. As a result, it has become vital for Elsevier to record and update the copyright and license state of every article it publishes.

In this paper we describe the data model that captures the copyright and licensing information of a journal article. This data model is in the process of being implemented at Elsevier and was mainly driven by Elsevier requirements in that existing systems involved in the publishing process must be able to retrieve and supply copyright information automatically even after the article has been published. The model allows us to generate copyright and license notices on the fly so that information presented in products is always accurate. In an increasingly dynamic digital publishing environment we recognize the importance of sound underlying data models that may be extended and updated quickly. The draft model went through internal reviewing process by publishing and legal domain experts. It is important to note that the model was created from the publisher perspective.

The paper is organized as follows: in we provide context for the data model in terms of Elsevier’s policy around the storage and access of its own content. Next, in we describe the copyright model and provide some example cases in . In we describe some of the literature around copyright and compare it to our model. Finally we end with a conclusion.

Context: Elsevier Content Model

The copyright and license data model is a small part of a larger model of Elsevier’s content which is composed of scientific publications, enhancements of publications, as well as data from products in specific domains that support scientists in knowledge discovery such as Reaxys and Embase . In the past this content was stored in dedicated data stores for each product but as part of Elsevier’s current content strategy, our aim is to present these repositories as a virtual whole and describe the content using existing (Linked Data) standards. The content and metadata are made accessible (internally) through APIs with the ability to add, retrieve, modify and delete content and metadata in what we call the Virtual Total Warehouse (VTW) . This content is composed of digital assets of various types such as the XML and PDF versions of journal articles, books, figures, as well as various enrichments of content that form the basis of Elsevier products such as ScienceDirect . The metadata of the content exposes its intrinsic properties necessary for retrieval and filtering, such as the type of content, the title, creator(s), and also various characteristics that may change over time even though the content itself does not change.

The metadata describing the content is stored as JSON-LD a W3C recommendation and all Elsevier content stored in the VTW adheres to Linked Data principles . Each piece of content is identified with a unique resolvable URI and is described using a set of properties and values. The context in each JSON-LD file provides mappings from JSON to RDF. We use external metadata standards such as Dublin Core Terms and PRISM for capturing content characteristics and PROV to describe the provenance of digital assets. The values of properties are often resources from vocabularies or lists of values in SKOS format.

Copyright Model Description

The model described in this paper is part of the larger Elsevier content model. It describes copyright-related properties of two classes of objects: Articles and Journals. The model may be adapted to other classes of published objects such as books by expanding the possible values of properties. Although articles are aggregated in journal volumes and issues, they have the same copyright owner as the journal and are not included in the model. In terms of copyright only the relationship between an article and the journal is important which is modelled using the dct:isPartOf property. We have defined dedicated copyright namespaces within the Elsevier domain:

• The copyright namespace cp: http://vtw.elsevier.com/data/ns/properties/Copyright-1/

• The open access namespace for open access licensing related properties oa: http://vtw.elsevier.com/data/ns/properties/OpenAccess-1/

Next we describe the copyright properties of Journal and Article class objects.

Example Copyright Cases

In this section we describe some frequently occurring copyright scenarios and how they are represented by the model.

{"@context":"http://vtw.elsevier.com/property/context.jsonld",
"@id":"http://vtw.elsevier.com/content/issn/12345678",
"cp:copyrightHolder": ["http://data.elsevier.com/vocabulary/ElsevierOperatingCompany#OCUK", 
"http://data.elsevier.com/vocabulary/ThirdPartyPublishers/Soc-12321"],
"cp:copyrightHolderType": ["http://data.elsevier.com/vocabulary/CopyrightHolderType#Elsevier", 
"http://data.elsevier.com/vocabulary/CopyrightHolderType#OtherPublisher"], 
"cp:operatingCompany": "http://data.elsevier.com/vocabulary/ElsevierOperatingCompany#OCUK" }

{"@context":"http://vtw.elsevier.com/property/context.jsonld",
"@id": "http://vtw.elsevier.com/content/pii/S012345678910",
"cp:copyrightHolder": ["http://data.elsevier.com/vocabulary/ElsevierOperatingCompany#OCUK", 
"http://data.elsevier.com/vocabulary/ThirdPartyPublishers/Soc-12321"],  
"cp:copyrightHolderType": ["http://data.elsevier.com/vocabulary/CopyrightHolderType#Elsevier", 
"http://data.elsevier.com/vocabulary/CopyrightHolderType#OtherPublisher"], 
"cp:copyrightYear" : "2013",
"cp:copyrightTransfer": "http://data.elsevier.com/vocabulary/CopyrightTransferType#FullTransferToPublicationOwner",
"cp:licenseGrantedToPublicationOwner": "http://data.elsevier.com/vocabulary/PublicationLicense#NotApplicable",
"cp:copyrightAgreement": "http://data.elsevier.com/vocabulary/JournalPublishingAgreements/JPAv1.3"}

Related Work

Several ontologies describe the Intellectual Property (IP) domain which covers copyright. In Zhang et al. present ontologies for IP protection. They are process oriented and concern the lifecycle of an IP work including its usage and focus on modeling aspects of IP violations such as piracy. ALIS IP is an IP ontology based on French copyright law centered on three concepts: work of mind(1) created by the author(s)(2) who are entitled to IP rights(3). The ontology covers various rights aspects, such as the distinction between moral rights (eg: attribution rights) and economic rights (eg: copying- , performance rights) and allows categorizations of the works according to type and authorship (eg: collaborative work). The Copyright Ontology by García et al. is very similar to ALIS IP in the concepts it covers, such as the distinction between a work, authorship and rights and similarly to Zhang et.al. has a strong focus on Digital Rights Management and e-commerce.

These ontologies are designed to cover IP law and processes revolving around IP but did not cover our requirements for a copyright model. First, they have a very broad coverage and include many process oriented concepts we do not need. Second, we would have had to create custom properties in order to model use cases such as governments retaining copyright, and capture specific publisher rights. Instead, we developed a light-weight model that covers all our requirements and will create mappings to external ontologies in the future.

Creative Commons provides a legal framework and expression language for communicating the allowed degree of sharing and reuse of content (on the web) thus capturing the rights of end-users. The CC licenses appear in three formats: legal code describing the license in legal terms, an explanation of the license and in Creative Commons Rights Expression Language (CC Rel) a machine readable representation that search engines may recognize. CC Rel is a set of properties needed to license the work including but not limited to its title, the attribution name, the license URI, permissions and prohibitions pertaining to the work. The language is light weight but does not include elements we need in our copyright model such as the copyright holder (attribution name and URL are weak pointers) and the copyright year. We do use CC licenses to to indicate end-user rights in open access articles.

Conclusion

In this paper we presented the copyright and license model that is being implemented at Elsevier for journal articles. Our model describes the rights of the authors, the publisher (publication owner), and the end-user community. The aim is to always identify the copyright holder, the degree of copyright transfer from the perspective of the publication owner, the publishing license granted to the publication owner if applicable and the end-user license. In addition, the model allows for some exceptional cases such as those where external agents such as governmental bodies keep copyright or do not allow the work to be copyrightable. We also provided a number of examples of the scenarios that occur in publishing.

The ontologies in IP domain described the previous section have too broad a coverage and we would have had to extend them in order to adequately capture copyright information we need. In the future we intend to map our properties to external ontologies.

The current model is centered on journal publishing but could be extended to apply to book publishing by extending the values of properties and capturing editor rights.

We hope that this model will prompt a dialogue with other (science) publishers on the best way of sharing and storing copyright information. Since journals are routinely bought and sold,and open access content may be reused to some degree, simple standard ways of capturing copyright information would benefit the publishing industry making the navigation of a complex legal landscape easier.

Acknowledgements

I would like to thank Rob Schrauwen and Véronique Malaisé for their helpful comments during the writing of this paper. The work on the model could not have been done without the domain experts whether from the implementation or the legal perspective: Jessica Clark, Jan Hanraads, Helen Long, Ralph Lupton, and Natalie Qureshi, the helpful input by Chris Shillum and mostly the long discussions with Rob Schrauwen.

References

1. Allen, B.P. 2011. Linked Data Standards and Infrastructure for Scientific Publishing. W3C Workshop on Linked Data Enterprise Data Patterns (Dec. 2011). http://www.slideshare.net/bpa777/ledp2011-v3

2. Camp. L.J. 2002 DRM: Doesn’t Really Mean Digital Copyright Management. Proceedings of the 9th ACM Conference on Computer and Communications Security 78-87.

3. García, R., Castella, D., and Rosa, G. 2013 Semantic Copyright Management of Media Fragments. In Proc. 2nd International Conference on Data Management Technologies and Applications (DATA).

4. Ryan, B. 2013. 1 - Copyright basics. Optimizing Academic Library Services in the Digital Milieu. Chandos Publishing 3-12.

5. Seeley, M., and Wasoff, L. 2012. 15 - Legal Aspects and Copyright. Academic and Professional Publishing. Chandos Publishing 355-383.

6. Sporny, M., Longley, D., Kellog, G., Lanthaler, M., and Lindström M. 2014. JSON-LD 1.0 A JSON based serialization for Linked Data W3C Recommendation. http://www.w3.org/TR/json-ld/

7. Abelson, H., Adida, B., Linksvayer. M., and Yergler, N. 2012, CC REL: The Creative Commons Rights. Expression Language. The Digital Public Domain. Foundations for an Open Culture

8. Berners Lee, T. July 27 2006. Linked Data - Design Issues. (Retrieved March 7 2015). http://www.w3.org/DesignIssues/LinkedData.html

9. Cevenini, C., Contissa, G., Laukyte, M., and Riveret, R. 2009. Ontologies and Law: A Practical Case of the Creation of Ontology for Copyright Law Domain. Handbook of Research on the Social Dimensions of Semantic Technologies and WebServices. 819–37

10. Zhang, X.M., Liu, Q., and Wang, H.Q. 2012. Ontologies for intellectual property rights protection. Expert Systems with Applications