Sergio Luján Mora

I. Introduction

Open Educational Resources (OER) are digital contents with an educational purpose available through websites that can be freely reused by teachers and students. Nowadays, there is a growing trend in using OER to improve quality in teaching and learning processes encouraged by important institutions and governmental initiatives, for example, UNESCO in the 2012 World Open Educational Resource Congress [1], the European Commission, in its recent “Communication on Rethinking Education” [2], and the U.S. Government in the “Second Open Government National Action Plan” [3].

OER offer a growing number of Learning Objects (LO), digital components prepared for educational purpose. LO are defined as “any entity, digital or non-digital that may be used and re-used for learning, education or training” [4]. In agreement with this concept, the OER can be conceptualized as LO. The LO are collected in Learning Object Repositories that store both LO and their metadata. The metadata aims to describe LO in a uniform and stable way, through some significant descriptors.

In this paper we review the application of Linked Data to OER context. Linked Data describes a method of publishing data so they can be machine readable and that enable data from different sources to be connected and queried through the use of technologies associated such as RDF [5]. The 5 Star deployment scheme describes, in a gradual way, how the information can be made public and suitable to be used with Linked Data. This paper describes how OER websites could be able to adopt Linked Data approach through the 5 Star deployment scheme in order to enhance their localization, retrieve, reuse and remix into an enriched context configured through Linked Data technology.

We structure this paper as follows: Section II describes the current status of OER, the categorization of OER websites, metadata standards and OER discoverability and search process; Section III describes Linked Data, Linked Open Data, and 5 Star deployment scheme; Section IV addresses issues related to Linked Open Data and OER to explain the adoption of Linked Data through 5 Star deployment scheme, and finally, Section V presents the discussion and future work.

II. Current status of OER

The definition more often cited about OER by William and Flora Hewlett Foundation [6] indicates: OER are teaching, learning, and research resources that reside in the public domain or have been released under an intellectual property license that permits their free use and re-purposing by others.

The providers of OER are universities, academic and research institutions, governmental initiatives, and education communities, which publish educational resources under an open license on public domain through websites. The resources can have distinct granularity, for example, full courses of universities degrees or only course materials, textbooks, streaming videos, assessments, tests, software or any other learning materials. The OER initiatives have grown steadily in recent years. Unfortunately, there is not a directory or unique reference about them. However some directories can be consulted online [7, 8, 9].

A. Categories of OER

In literature we found many approaches to categorize OER in accordance to different parameters, e.g. content, scope, provider, type of repository. McGreal [10] categorizes OER websites based on the locality of the resources and metadata; Sicilia & Ochoa [11], introduces the concept of referatories containing descriptions for resources, aggregated collections, repositories hosting, meta-aggregators, and Open Courseware; Creative Commons [12] proposes a categorization based on the scope and topics in the website such as: unstructured OER that focuses on a single topic or idea, OER website moderately structured (resources by subject area); and fully structured OER, (complete courses); Hylén [13] suggests a typology of different OER repositories according to their activities, large scale operations and small scale operations; and their providers; Butcher by UNESCO [14] recognizes Open Courseware (OCW), University OCW Initiatives, Content Creation Initiatives, Subject-Specific OCW OER, Open Schooling Initiatives, and OCW OER Search.

We adopt two categorization criteria: type of resources offered by OER websites, and locality of repositories where resources are stored. These categorizations are important in order to understand issues related to metadata and search process.

1) Categorization for type of resources

We propose an adaptation of the classification defined by Neil Butcher for UNESCO [14].

a) Open Courseware (OCW)

This initiative promoted by universities focuses on freely available resources organized as courses. These materials mostly correspond to formal courses of academic programs. They include planning materials, thematic content (e.g. lecture notes, text books, samples, simulations) and evaluation tools (e.g. assignments, quizzes, tests). In this initiative we have two types:

• The associations of OCW sites from many universities. For example, Open Education Consortium (www.oeconsortium.org) with over 200 universities and other organizations globally, currently this site offers more than 30,000 courses; OCW Universia (www.ocwuniversia.net) groups more than 150 universities including some from Latin American.
• OCW projects that belong to a single university, such as: MIT OCW (ocw.mit.edu), it is the precursor of this movement and perhaps the most known; OCW Universidad Politécnica de Madrid (ocw.upm.es); OLI Carnegie Mellon University (oli.cmu.edu).

b) Content Creation Initiatives

This category references the sites that allow to create resources in collaborative environment. An example is website Curriki (www.curriki.org/), with materials suitable for basic education level, currently it offers more than 56,000 resources.

c) Subject-Specific OCW OER

OER Websites with specific content for a specialized area. For example, Health Education Assets Library from Utah University (library.med.utah.edu/h) with content oriented to health science education; and National Science Digital Library (nsdl.oercommons.org/) with resources for science, technology, engineering and mathematics education.

d) OER Repositories and websites

In this category we found websites that offer educational resources from many providers. The content is presented in multiple formats and scope, but not necessarily related to courses in universities or other educational institutions. Some important websites are: MERLOT II (www.merlot.org/), it is a program of California State University, and OER Commons (www.oercommons.org/). In both cases, the content providers are outstanding universities around the world, research institutes, libraries and institutions related to the educational field.

2) Categorization for type of repository

According to the type of repository where LO are stored, it is possible to recognize three cases of OER websites [10, 11]:

1. Type A. OER websites accessing content stored in local repository.
2. Type B. OER websites providing access to content stored on external repositories. Mainly through heterogeneous metadata.
3. Type C. OER websites providing access to local repository as well as to external repositories. These are hybrid websites.

B. Metadata Standards

The National American Standard Organization, a non-profit association accredited by the American National Standards Institute (ANSI) defines metadata as “structured information that describes, explains or otherwise makes it easier to retrieve, use, or manage an information resource” [15].

In case of OER, metadata are structured information constituted by a set of attribute values for specific descriptors or tags that describe a resource. The discoverability success of OER is associated to standardized descriptive metadata used to tag OER, because it enables descriptor-based searches through web interfaces [16] . Some widely used metadata standards in Learning Object Repositories are described below.

1) IEEE Learning Object Metadata (IEEE LOM)

It is a widespread metadata standard released by IEEE Learning Technology Standards Committee, as IEEE 1484.12.1 – 2002 [17], oriented to educational contents. This standard allows to describe LO by specifying their syntax and semantic. IEEE LOM sets a hierarchy of elements grouped into nine categories, which can contain subcategories to describe aggregate of elements.

2) Dublin Core Metadata Initiative (DCMI), Element Set (DCMES)

It is a vocabulary of fifteen properties for a generic scope, used to describe a wide range of digital resources [18]. This reference version of the fifteen element descriptions have been formally endorsed in the standards, ISO Standard 15836:2009, ANSI/NISO Standard Z39.85-2012, and IETF RFC 5013 of August 2007.

3) Learning Resource Metadata Initiative (LRMI)

The LRMI project started as an extension of Schema.org in 2011. Schema.org is a broader standard for tagging online content that provides a collection of shared vocabularies to create a structured data markup schema supported by major search engines such as: Google, Microsoft, Yandex and Yahoo! [19]. The initial LRMI specification was accepted as an official extension of Schema.org in 2013 [20]. LRMI is concerned with extending and applying Schema.org to the description of relevant properties of educational resources; thereby, educational resources published using LMRI are recognized by major search engines [21].

C. OER discoverability and search process

The increasing online availability of OER online in distinct sources configures a complex scenario for discoverability, search and acquisition of resources. According to a recent study [22], users expect an efficient and effective searching and retrieving content to match their requirements. To achieve this goal, the website should provide faceted search, which is based on multiple filters for searching, one for each different descriptor. This advanced search leads to more precise results and provides intuitive navigation mechanisms for stimulating content exploration. Faceted search and navigation is especially useful in OER websites with extremely large content sets [23]. This type of search is usually applicable to type A repositories, as described in section 2) Categorization for type of repository.

The advanced search in sites type B, according to description in section 2) Categorization for type of repository, requires federated search; that is, search in university repositories or databases with distinct metadata. In consequence the resource descriptors are quite different from each other, so search parameters need to be common to all metadata. The search outcomes must be aggregated to provide a single list of results.

In relation to discoverability, that is, the searching of OER in web through search engines there are still many unsolved problems e.g. interoperability among repositories of distinct type, non-standardized metadata in repositories, and mislabeling of resources. Discoverability of resources that match requirements, regardless where they are hosted, remains a challenge [24].

III. Linked Data

The Semantic Web is a Web of Data [25] that provides a common framework to allow large scales integration of, and reasoning on, data on the Web. However, to achieve the Web of Data is necessary to have a huge amount of data on the Web, related among them, in a standard format, reachable and manageable by Semantic Web tools. The interrelated datasets coming from different sources on the Semantic Web are produced through Linked Data.

According to the seminal paper by Bizer, Heath and Berners-Lee, Linked Data refers to data published on the web that it is machine-readable, its meaning is explicitly defined, it is linked to external data sets, and can in turn be linked to from external data sets [26].

Linked Data is based on Uniform Resource Identifier (URI) to identify any entity (concept or thing) that exists in the world, and Uniform Resource Locator (URL) that addresses document and other entities that can be located on the Web. For locating a resource identified by URI, it is dereferenced over the HTTP protocol [26]. So, all URLs are URIs, but not all URIs are URLs [27]. The objects identified by URIs may be interlinked together through a meaningful connection by means Resource Description Framework (RDF), a technology that is critical to the Web of Data. In a way of analogy to human language, RDF is a grammar for language of data, in which URIs are words (nouns and verbs). RDF statements must be expressed with a three part structure (subject, predicate, object) calling the triple. The subject and object of a triple are both URIs, or a URI and a string literal respectively. The predicate specifies how the subject and the object are related, and it is also represented by URI [28].

Another important concept in Linked Data is the vocabulary. A vocabulary in Linked Data collects definitions of a set of classes and properties, called terms of the vocabulary used to describe specific types of things (general types or specific to a given domain). The use of a vocabulary associates semantic to descriptions and turn them into meaningful data [29].

A.Principles of Linked Data

In order to publish data on the Web as Linked Data Tim Berners-Lee [5] defined four Linked Data principles:

1. Use URIs as names for things.
2. Use HTTP URIs so that people can look up those names. (And also as a key protocol for software agents to resolve URIs).
3. When someone looks up a URI, provide useful information, using the standards (RDF for representing data, and SPARQL for querying data).
4. Include links to other URIs so that they can discover more things. (Embedding URIs within RDF datasets. So, humans and computers can discover additional information).

B.Linked Open Data

Linked Data can be applied in non-open context. OER are essentially Open Data, for this reason, the context for this work is Linked Open Data, it is the application of Linked Data to Open Data.

C. Linked Data 5 Stars deployment scheme

Tim Berners-Lee [5] proposed a 5 Star deployment scheme applied to Linked Open Data that describes how the information can be made public in a progressive scale to aim to the purpose of Linked Data; that is, data be helpful to enrich relationship with other datasets. The 5 Star deployment scheme does not provide technical definition, instead, it is targeted to data owners who publish data on the web to encourage them to align with Linked Data principles. Table I [30] shows the stars and their respective description for compliance level.

At the moment, the rating of datasets in Linked Open Data 5 Star deployment scheme can be shown by badges on a web page [31]. However, as far as we known, there isn’t a metadata to specify a rating on 5 Star deployment scheme. A metadata about 5 Star deployment scheme can be helpful for non-semantic search engines to detect if an educational resource have a rating on 5 Star deployment scheme.

D.Licenses for Open Data

The 5 Star deployment scheme is only for Open Data, so it is important to bear in mind the meaning of Open, something is open if anyone is free to access, use, modify, and share it — subject, at most, to measures that preserve provenance and openness [32].

Resources published on the Web have a license that is a declaration in standardized way to the legal conditions for use under which the resource is made available. So, according to Creative Commons [33], a widely spread license for resources, only type of licenses that guarantee the characteristics of “Open” according definition above, could be considered. That is, to get into Linked Open Data 5 Star deployment scheme, resources should have only one of the licenses showed in Table II.

Creative Commons provides machine readable metadata to place in the page which allows software to understand that you have applied a license to the resource, and it is possible to get back some metadata encoded in RDF. It is recommended to use RFDa because it is a general technology for Linked Data markup in a variety of HTML-like languages (HTML5 including) [34].

IV. Linked Data and OER

Linked Data has been used to manage data in several domains, but as far as we know Linked Data has been used with web datasets related to education [35] , but not with OER websites. One major issue is connectivity of the datasets due to its heterogeneity. It is necessary the alignment of common terms in these vocabularies. Some initiatives have started to expose Open Data with Linked Data approach.

A research about educational Linked Data on 2013 [35] reported over 146 datasets relevant to educational domain with certain level of heterogeneity. Also, the results showed that connectivity was mainly related with the reuse of common vocabulary than to links among data. Besides, the research found that Dublin Core standard was the most used vocabulary for connections.

Some OER initiatives begin to publish data as Linked Data, it means to use the principles of Linked Data to get that data to be machine readable, for example providing meaning to terms by using microdata (Schema.org), RDF and ontologies.

On the other hand, Open Data focuses on allowing users to get access to the data and to query data:

1. Use URIs as names for OER as unique identifiers, e.g. URI for: OER creator, OER knowledge areas, OER audience level, OER primary format, and others descriptors.
2. Using URI, the corresponding OER can be found through URL (by dereference).
3. When someone looks up a URI that identifies OER provides integration with other OER.
4. OER should be enriched by including relations with others OER.

A. Linked Open Data 5 Star scheme and OER

The opportunity that we find for using Linked Open Data 5 Star scheme in OER is based on their openness. Besides, according to definition of OER, a key aspect is repurposing; it means, users can use resources without modifications, extract part of the content, modify content, and remix with distinct content piece. The adoption of Linked Open Data 5 Star deployment scheme to OER implies some challenges that are explained below:

1 Star: Available on the web (whatever format) but with an open license.

OER are published under an open license.

2 Stars: Available as machine-readable structured data (e.g. excel instead of image scan of a table)

OER websites don't have a standard for naming formats, hence the same format can appear with different names on websites. These are screen captures of format selection for resources. Fig. 1 shows formats for OER Commons, in left side, and MERLOT, in right side. Besides, to get an idea about format of resources we review type of resources for each format selection in both OER websites. For OER Commons, we have: HTML 84%, Downloadable docs (mostly pdf) 13%; for MERLOT, most representative formats are: HTML 48%, Video 38%, PDF 6%.

Besides, it is possible to have non-structured data. For example tables that contain data could be inserted in HTML pages as images, and significant information that describes an image could be treated as a whole image, so can't be machine readable. Another example is a presentation saved in pdf format that cannot be extracted or modified.

3 Stars: As above plus: Use non-proprietary format (e.g. CSV instead of excel)

The same explanation for 2 Stars is applied for this level. It is not possible to ensure structured information neither that it could be in non-propietary format. However in our review we found a short number of resources in explicit proprietary format, for example, format ebook, less than 1% in OER Commons; flash less than 1% in MERLOT.

4 Stars: All the above plus: Use open standards from W3C (RDF and SPARQL) to identify things.

In spite of significant number of educational datasets available the adoption of Linked Data principles and vocabularies is not an extended practice in this context.

This is a huge task because OER are distributed in heterogeneous repositories and services that limit interoperability. Regarding metadata, while main repositories have metadata standards, some others repositories have described their resources using their own methods, such as XML based schemas and heterogeneous taxonomies [36]. Consequently, the schema and data transformation to RDF is a pending issue.

5 Stars: All the above, plus: Link your data to other people’s data to provide context.

To reach this level of Linked Open Data 5 Star deployment scheme, it is necessary to enrich metadata, i.e. use structured and formal metadata and linking it with established Linked Data vocabularies and datasets on the Web.

So, the process for adoption of Linked Data needs to consider at least these stages: reviewing of metadata standard used in repository, evaluate and decide a consensual vocabulary (it needs an underlying ontology); construct RDFs statements including disambiguation (in case of distinct data sources and metadata); construct a RDF store and finally offer exploration to users through a web application.

The translation from existing vocabularies from metadata to RDF vocabulary is one of the major challenges. For example, metadata retrieved from IEEE LOM metadata, a standard commonly accepted for Learning Object Repositories such as MERLOT, needs to be automatically translated into RDF and mapped with compatible metadata schemas to be exposed as Linked Data accessible via dereferences URIs. By now, mapping and linking of the IEEE LOM to the Linked Data is a work in course [37].

In case of DCMI metadata terms, the property “contributor” was accepted in FOAF property “creator”. To mapping statements FOAF and DCMI made consensus in order to allow automatic mapping processing. In addition, the properties and classes of DCMES have been defined for compatibility with Linked Data principles [38].

With respect to LRMI, the new vocabularies of Schema.org must be assigned to RDF vocabularies, and it is still a challenge. Websites such as OER Commons have adopted this standard to take advantage of search engines.

In summary, the best option is to use terms from widely deployed vocabularies to represent data (not create new terms) [29].

V. Discussion and future work

The adoption of Linked Data technology to connect and enrich OER environments requires the recognition of both perspectives, technical and end-users. The technical team behind OER initiatives should be able to integrate Linked Data and technology associated such as SPARQL into their common development environment. On the other hand, the exploration with Linked Data needs a suitable browser to interpret the results without burdening the users' work.

In order to obtain an advantage of using Linked Data in OER websites, would be recommendable the addition of Linked Data information into the selection criteria form of advanced search. In this research we propose a way to achieve this goal through the use of metadata for 5 Star deployment scheme similar to metadata for Creative commons, associated to resources enriched with Linked Data. So, users could see these resources with an external browser tool in case that the website doesn't offer its own browser tool.

Moreover, Linked Open Data and Educational domain are the natural context for conjunction between Linked Data and OER, thus initiatives in such context should to consider some advanced work in relation to Linked Open vocabularies and educational context vocabularies to enrich the resources.

Eventually, the OER community could encourage to users, not only students but teachers, to use OER with Linked Data, through specialized search facilities, in such a way that they can experiment improved efficiency and effectiveness in the search for educational resources that meet their requirements.

In future works we plan to propose a metadata for 5-Star deployment scheme based on Linked Open Data vocabularies and educational context vocabularies.

Acknowledgment

This work has been partially supported by the Prometeo Project by SENESCYT, Ecuadorian Government.

References

[1] UNESCO, "2012 World Open Educational Resources (OER) Congress," Paris, France.

[2] European Commission, "New Rethinking Education strategy," 11 2012. [Online]. Available: http://europa.eu/rapid/press-release_IP-12-1233_en.htm. [Accessed 03 02 2015].

[3] U.S. Government, "The Open Government Partnership, "Announcing New Open Government Initiatives as part of the Second Open Government National Action Plan for the United States of America," September 2014. [Online]. Available: http://www.whitehouse.gov/sites/default/files/m," Washington, USA, September 2014.

[4] IEEE, "IEEE Learning Technology Standards Committee," 2001. [Online]. Available: http://ltsc.ieee.org/doc/. [Accessed 17 02 2015].

[5] T. Berners-Lee, "Linked Data," 18 06 2009. [Online]. Available: http://www.w3.org/DesignIssues/LinkedData.html. [Accessed 25 01 2015].

[6] The William and Flora Hewlett Foundation, "Open Educational Resources," [Online]. Available: http://www.hewlett.org/programs/education/open-educational-resources. [Accessed 25 02 2015].

[7] OER Quality Project, "Directory of OER repositories," [Online]. Available: http://oerqualityproject.wordpress.com/2012/10/22/directory-of-oer-repositories/. [Accessed 4 10 2014].

[8] Discover and Open Source World, "A guide to free and open source education," [Online]. Available: http://opensource.com/education/13/4/guide-open-source-education. [Accessed 4 10 2014].

[9] Commonwealth of Learning, "DOER - Directory of Open Educational Resources," [Online]. Available: http://doer.col.org/. [Accessed 25 02 2015].

[10] R. McGreal, "A Typology of Learning Object Repositories," in Handbook on Information Technology for Educationa and Training, Springer, 2007, pp. 5-28.

[11] M.-A. Sicilia, X. Ochoa, G. Stoitsis and J. Klerx, "Learning Objects Analytics for Collections, Repositories & Federations," in Proceedings of the Third International Conference on Learning and Analytics and Knowledge (LAK'13), Leuven, Belgium, 2013.

[12] Creative Commons, "Free to Learn Guide Different Types of OER Meet Different Needs," [Online]. Available: https://wiki.creativecommons.org/Free_to_Learn_Guide/Different_Types_of_OER_Meet_Different_Needs. [Accessed 10 01 2015].

[13] J. Hylén, "OECD - Open Educational Resources: Opportunities and Challenges," [Online]. Available: http://www.oecd.org/edu/ceri/37351085.pdf. [Accessed 10 01 2015].

[14] N. Butcher, A Basic Guide to Open Educational Resources (OER), Commonwealth of Learning & UNESCO, 2011.

[15] NISO, "National Information Standard Organization," 2004. [Online]. Available: http://www.niso.org/publications/press/UnderstandingMetadata.pdf. [Accessed 21 01 2015].

[16] D. Hillman, "Using Dublin Core," Dublin Core Metadata Initiative, [Online]. Available: http://www.dublincore.org/documents/usageguide/. [Accessed 02 04 2014].

[17] IEEE, "WG12: Learning Object Metadata," [Online]. Available: http://ltsc.ieee.org/wg12/index.html. [Accessed 04 05 2014].

[18] D. C. M. Initiative, "Dublin Core Metadata Element Set, Version 1.1," 14 06 2012. [Online]. Available: http://dublincore.org/documents/dces/. [Accessed 12 01 2015].

[19] LRMI, "Learning Resource Metadata Initiative," [Online]. Available: http://www.lrmi.net/. [Accessed 15 01 2015].

[20] Schema.org, "What is schema.org," [Online]. Available: https://schema.org/. [Accessed 03 02 2015].

[21] LRMI, "About the LRMI," [Online]. Available: http://www.lrmi.net/about. [Accessed 11 02 2015].

[22] D. Dicheva and C. Dichev, "Leveraging Domain Specificity to Improve Findability in OER Repositories," in Research and Advanced Technology for Digital Libraries, Springer, 2013, pp. 466-469.

[23] K. Whitenton, "Nielsen Norman Group, Filters vs. Facets: Definitions," [Online]. Available: http://www.nngroup.com/articles/filters-vs-facets/. [Accessed 18 02 2015].

[24] P. G. West and L. Victor, "Background and action paper on OER," The William and Flora Hewett Foundation, 2011.

[25] W3C , "Semantic Web," 2015. [Online]. Available: http://www.w3.org/standards/semanticweb/. [Accessed 12 04 2015].

[26] C. Bizer, T. Heath and T. Berners-Lee, "Linked Data - The Story so far," International Journal on Semantic Web and Information Systems (IJSWIS), vol. 5, pp. 1-22, 2009.

[27] Open Org, "Linked Data Basics for Techies," [Online]. Available: http://openorg.ecs.soton.ac.uk/wiki/Linked_Data_Basics_for_Techies. [Accessed 19 02 2015].

[28] T. Baker, "Libraries, languages of description, and linked data: a Dublin Cores perspective," Library Hi Tech, vol. 30, no. 1, pp. 116-133, 2012.

[29] W3C, "Linking Open Data on the Semantic Web," [Online]. Available: http://www.w3.org/wiki/TaskForces/CommunityProjects/LinkingOpenData/CommonVocabularies. [Accessed 25 02 2015].

[30] Tim Berners Lee, "Linked Data," [Online]. Available: http://www.w3.org/DesignIssues/LinkedData.html. [Accessed 18 01 2015].

[31] LOD, "Linked Open Data star badges," [Online]. Available: http://lab.linkeddata.deri.ie/2010/lod-badges/. [Accessed 25 01 2015].

[32] Open Knowledge, "Open Definition," [Online]. Available: http://opendefinition.org/od/. [Accessed 05 02 2015].

[33] Creative Commons, "About the Licenses," [Online]. Available: http://creativecommons.org/licenses/. [Accessed 18 02 2015].

[34] Creative Commons, "Metadata," [Online]. Available: https://wiki.creativecommons.org/Metadata. [Accessed 12 02 2015].

[35] M. d’Aquin, A. Adamou and S. Dietze, "Assessing the educational linked data landscape," in Proceedings of the 5th Annual ACM Web Science Conference, (WebSci '13), New York, USA, 2013.

[36] S. Dietze, S. Sanchez-Alonso, H. Ebner, H. Qing Yu , D. Giordano, I. Marenzi and B. Pereira Nunes, "Interlinking educational resources and the web of data: A survey of challenges and approaches," Program, vol. 47, no. 1, pp. 60-91, 2013.

[37] E. Rajabi, M.-A. Sicilia and S. Sanchez-Alonso, "Interlinking Educational Resources to Web of Data through IEEE LOM," Computer Science and Information Systems, vol. 12, no. 1, p. 233–255, 2015.

[38] Dublic Core Metadata Initiative, «DCMI Metadata Terms,» [Online]. Available: http://dublincore.org/documents/2012/06/14/dcmi-terms/?v=elements#format. [Accessed 05 05 2014].