Abstract:

One of the main advantages of using digital repositories in the e-learning process is that they provide mechanisms to encourage the discovery, exchange and reuse of learning resources. Digital repositories are already well established throughout many countries in the European Union and in the US. Digital repositories are on their way to become a permanent part of the scholarly communication and documentation research infrastructure.
In this paper we review existing digital repositories and projects and formulate requirements that an advanced repository for open learning resources should meet.

Then we present a model for organization and representation of learning resources. It is based on standards to allow the reuse of learning resources and offer interoperability among heterogeneous e-learning systems.

The model consists of three parts: Structural ontology, Classification taxonomy and Metadata.
The Structural Ontology defines the objects (learning resources, users, comments, tags, ratings, etc.), their properties (for example: width and height of a picture; width, height and duration of a video, etc.) and relationships (between resources, users, comments, ratings, etc.). The ontology is developed in OWL.

The Classification taxonomy is used for the classification of the learning resources. The taxonomy is built using SKOS (Simple Knowledge Organization System) - a formal language designed for representation of thesauri, classification schemes and taxonomies. SKOS is built upon RDF and RDFS. The model allows the use of several distinct classification taxonomies. The taxonomies can also be expressed in OWL.
The Metadata can be presented in one of the standard formats Dublin Core and LOM (Learning Object Metadata). However, we propose the use of Dublin Core for its simplicity and wide acceptance. We have made recommendation for encoding Dublin Core metadata for the different types of learning resources.

The presented model is used in building a learning repository where resources can be search, retrieved and used in heterogeneous e-learning systems.

Then we discuss the architecture of the system and the main components: Fedora repository, containing the resources, metadata, relations between the resources, Web portal and Search engine.
We have chosen Fedora as a basic repository platform. The Fedora repository contains the learning resources (documents, presentations, videos, pictures, etc.) or references to external web resources. It also contains the metadata for the resources.

We use the Dublin Core (DC) metadata standard to express the metadata for learning resources as Fedora repository has full support for DC and automatically creates indexes on DC fields and supports search within DC fields.

The Web portal is the other main component of the system. It is a Web application (we have chosen Drupal 7 as implementation platform) that interacts with the user and implement the business logic of system. It communicates with the repository through the web services exposed by Fedora (Fedora API). The Web portal also communicates with Web 2.0 tools (YouTube, Flickr, FaceBook, Twitter, etc.) using corresponding APIs.
We use Apache Solr search server for the Search engine of the system. Its major features include full-text search, hit highlighting, faceted search, dynamic clustering, etc.

In the presented system we have incorporated our experience in several EC funded project in this domain and have added new features and innovations.