OntoPortal fact-file

What's the Problem?

The process of locating detailed information using the Web involves detective work (What did the author of this document go on to write? What other documents describe this project? Did this work influence any standards or other software?) because information fragments are disconnected from each other and from the records and reports of the activities which produced them.

The Web and the use of hypertext have eased the situation, but:

• Constructing structured and rich hypertext portals is a labour intensive and error prone task.
• Such portals are a key component to the developments to the Semantic Web.
• Toolkits that semi-automate the task of creating content rich and highly interlinked portals are essential.

The accurate and intuitive presentation of knowledge is an integral part of the AKT project. The CS AKTive portal will use various technologies to present the UK Computer Science research community and structured hypertext may play a role in this.

Towards a Solution

We see the key to providing the solution in the integration of hypertext and formalisations of the concepts in a domain (ontology). The links in ontological hypermedia are defined according to the relationships between real-world objects. For example, an ontology that models the significant objects in a scholars world can be used toward producing a consistently interlinked research literature.

OntoPortal provides a framework that:

• Allows authors to import ontologies for use in creating a portal.
• Provides an intuitive editing environment to populate the imported ontology.
• Allows readers to become involved and impose their own conceptual scheme.
• Automatically creates rich and highly interlinked hypertext Web pages.

The technology

Figure 1 illustrates an overview of the architecture for the OntoPortal system. The knowledge, corresponding to facts in the ontology, is stored in an XML-based knowledge base (3). Requests from client browsers (1) to view information about a particular resource initiate a query mechanism which retrieves the appropriate metadata from the knowledge base (2), combining where necessary (e.g. a request for all literature metadata).

Figure 1: OntoPortal architecture

This mechanism also determines the relationships between the requested resource and the other resources in the metalayer, and adds this information to the XML records as hyperlinks that invoke further queries.

The resulting XML document (4) is either sent directly to the client browser (5) if the browser supports XSLT transformations, or transformed into a presentation format (usually HTML) on the server-side (7). In the former case, the user can use the XSLT stylesheet from the server (6), or provide an alternative stylesheet to meet individual presentation needs.

The knowledge gathering process is facilitated through HTML forms used by editors to enter metadata describing a resource. To integrate the resource with its proper context, valid ontological relationships with other resources are presented to the editor, from which the appropriate ones are selected.

Try a Demonstration

A static demonstration (no authoring capabilities) of OntoPortal populated with "metadata research" is available here. Two videos are available to demonstrate OntoPortal's authoring (6MB) and navigation (8.8MB) features. Otherwise, a sample interaction with screen shots is provided below.

Sample Interaction

In Figure 2 a user, having located a research paper relevant to a task, wishes to discover further papers by the same authors in order to increase understanding of the concepts the research paper discusses. For example, in a Digital Library, realizing this intent involves resorting to a search engine and searching for papers written by these authors. However, the search results are limited by the publishing rights of the Digital Library, and will not return links to relevant research papers in other Digital Libraries. Furthermore, the user has no understanding of how the papers returned by the search engine are related to the original paper.

Figure 2: Literature instance

Figure 3: Researcher instance

Using the OntoPortal system, in investigating the metadata associated with the original research paper, the user is able to find out directly which other research papers the authors of the paper have been involved in writing. The user moves to the 'Researchers that authored this literature' link information, and from here is able to take advantage of familiar bibliographic author/literature relationships in order to explore the research papers produced by each author. For example, following the link to metadata about a particular researcher (Figure 3), the user is able to view a list of publications by that researcher, visiting metadata about those which seem relevant (Figure 4). The user is able to become fully immersed in the research community and context surrounding the research paper.

Figure 4: Literature instance

Figure 5 illustrates an example authoring form for a researcher instance. Using this form, authors can associate other instances with it. For example, to indicate that this researcher runs a particular course, the course name is located in the "Course Relationships" list box.

Figure 5: Authoring a Researcher instance

Further Reading

Key documents:

• Kampa, Dr Simon and Miles-Board, Mr Timothy and Carr, Dr Leslie (2001) Hypertext in the Semantic Web. In Proceedings ACM Conference on Hypertext and Hypermedia 2001, pages 237-238, Aarhus, Denmark.
• Kampa, Dr Simon and Miles-Board, Mr Timothy and Carr, Dr Leslie and Hall, Prof Wendy (2001) Linking with Meaning: Ontological Hypertext for Scholars. Technical Report ECSTR-IAM01-005, Electronics and Computer Science, University of Southampton.

Semantic representation

View in the AKT Triplestore Browser or as RDF.

Also available in DOAP RDF (Description Of A Project)