Ontology based access to digital libraries
(Sonia Bergamaschi, DSI - Universita di Modena e Reggio-Emilia, Fausto Rabitti, CNUCE-CNR)
The Internet is making accessible a large, and increasing, number of Digital Libraries, originally intended for specific and specialised groups of users, to a wide range of potential users. The problem of controlling, exchanging and integrating the semantics associated to Digital Libraries (i.e., the associated metadata) is becoming more and more important.
This need has been stressed by several initiatives. For example, the Open Archives initiative (OAi), in US, aims at guaranteeing interoperability among Digital Libraries (e-print archives). It has established a set of relatively simple but potentially quite powerful interoperability specifications that facilitate the development of services implemented by third parties.
Metadata in Digital Libraries, for bibliographic data, are usually expressed according to models like Dublin Core or MARC. However, there is the need to generalise the description of data and metadata made available in a large variety of Digital Libraries. The wide acceptance on the Web of XML can be a decisive factor in this direction.
XML, a standard proposed by W3C, is a mark-up language intended to make the information as self-describing, separating the function of document structural description and document presentation. The document tags can be use to describe the meaning of the document components. Controlling the semantic associated to XML tags will be a decisive task. This will open new perspective in accessing Digital Libraries, since XML is going to become the new interoperability standard for distributed Digital Libraries. We foresee a situation where XML will be used both for exchanging digital (often multimedia) documents and their multi-modal presentations (via XSL), and for defining their metadata, using XML DTD or Schema descriptions, with associated RDF (Resource Description Framework) schema descriptions.
For this purpose, we propose an ontology-based approach which aims to build a Digital Library Ontology representing a global virtual view of distributed Digital Libraries and defining mapping rules between local and global views: this mapping will be obtained by building a "Common Thesaurus" of intensional and extensional intra and inter-schema relationships able to reconcile different representation of similar concepts. The starting point is the MOMIS system (Mediating system environment fOr Multiple Information Sources) (see: sparc20.dsi.unimo.it), whose prototype has been recently presented at the VLDB 2000 conference.
- Presentation slides
- e-mail: sonia.bergamaschi@unimo.it
- URL: http://www.dsi.unimo.it/staff/st12/Bergamaschi.idc