Implementation Semantic Web on E-Learning

The current WWW is a powerful tool for research and education, but its utility is hampered by the inability of the user to navigate easily the nefarious sources for the information he requires. The Semantic Web is a vision to solve this problem. It is proposed that a new WWW architecture will support not only Web content, but also associated formal semantics [4]. The idea is that the Web content and accompanying semantics (or metadata) will be accessed by Web agents, allowing these agents to reason about the content and produce intelligent answers to users' queries.

The Semantic Web, in practice, comprises a layered framework: an XML layer for expressing the Web content; a Resource Description Framework (RDF) [8] layer for representing the semantics of the content; an ontology layer for describing the vocabulary of the domain; and a logic layer to enable intelligent reasoning with meaningful data [18].

XML was designed as a simple, flexible way of transporting structured documents across the Web. With XML, "tags" or hidden labels may be created - such as or - that annotate Web pages or sections of text within a page. XML is machine-readable, i.e. programs can read and understand it, but the program developer has to know what the page writer uses each tag for. In other words, XML allows users to add arbitrary structure to their documents but says nothing about what the structures mean [5].

The meaning of document content is expressed with RDF that is simply a data model and format that allows the creation of machine-readable data. It comprises a set of triples, i.e. three Universal Resource Identifiers (URI) that may be used to describe any possible relationship existing between the data - subject, object and predicate [7] [16]. Thus, all data stored in the system is easily readable and processable. It is important to note that RDF provides the syntax, but not the actual meaning of the properties we ascribe to the data. For example, it does not define what data properties such as Title or Category or Related-To mean. Properties like these are not standalone; they come in packages called domain vocabularies. A learning object, for example, may include a set of properties such as Course, Sub-Section, Author, Title, Similar-To, Difficulty-Level, Rating, etc. Thus, for every domain there is a need for a specific ontology to describe the vocabularies and to make sure they are compatible.

Ontologies in the context of the Semantic Web are specifications of the conceptualization and corresponding vocabulary used to describe a domain [12]. Any semantic on the web is based on an explicitly specified ontology, so different Semantic Web applications can communicate by exchanging their ontologies. Several representation schemes have been defined for the ontology layer. The most popular one, the Ontology Interchange Language (OIL), combined with the DARPA Agent Markup Language (DAML), DAML+OIL, provides a rich set of language structures with which to create ontologies and to markup information so that it is machine understandable [18].

The logic layer part of the Semantic Web is not fully developed yet. Its implementation will allow the user to state any logical principles and permit the computer to infer new knowledge by applying these principles to the existing data. Since there are many different inference systems on the Web that are not completely interoperable, the vision is to develop a universal logic language for representing proofs - systems will then be able to export these proofs into the Semantic Web [1].

Within an e-learning framework, the Semantic Web provides the technology that allows a learning object to be (i) described with metadata, and this description to be extended indefinitely (by anyone, not just the creator); (ii) to be annotated with personal notes and links by anyone; (iii) to be extended in terms of content, allowing multiple versions to exist; (iv) to be shared by, and communicated to, anyone who has expressed an interest in such content; (v) to be certified, for example, as a quality learning resource; and more [16]. The e-learning application described here embraces this functionality.

There are two types of agents used in the application: StudentAgent and InstructorAgent, both of them implemented as Java classes. Users are served by the appropriate agents, which parse the metadata and tailor the user interface to satisfy the user's