Over the last decade, ontology-based information systems gained increasingly in importance. At the beginning, lot of effort went into the analysis of different languages to express ontologies. With the evolving standards for ontologies especially in the Semantic Web effort, led by the W3C, the issue about language became less important. Thus, it was possible to focus on the process of developing ontology-based systems.

The knOWLer system is an innovative ontology-based system for local or distributed information management. Starting from Parka’s ideas and principles, knOWLer has a new design, removing some of the limitations of Parka-DB while keeping the performance at the same level. By choosing OWL as the representation language, knOWLer makes an important step towards modern ontology languages but is still providing support for traditional frame-based systems. Developed by W3C, the OWL language consists of three overlapped layers. The simplest language among them, OWL Lite, supports classification hierarchy and simple constraint features. It provides a lower entry threshold to the language guarantying the computational completeness and decidability. On top of it, OWL DL is more expressive, but still preserves the type separation. The last and most general layer, OWL Full, includes the complete OWL vocabulary and gives the freedom of interpretation provided by RDF.

In the knOWLer system, we extended the OWL Lite layer by allowing the usage of RDF statements, giving users the possibility to treat them as individuals and to apply properties to them. The resulting language will be referred as Extended OWL Lite. The reasoning capabilities provided by knOWLer System follow the semantics defined by the OWL Lite language. The architecture of our ontology-based information management system consists in four distinct modules: knOWLer kernel, Storage Interface,
Import/Export Interface and User Interface. All modules are linked through the kernel, which is the main component.

OWL Representation of the WordNet Ontology

WordNet.OWL is an OWL-ontology based on WordNet 1.7.1 lexical database. WordNet.OWL contains both the ontology schema and the instances. This OWL representation of the WordNet ontology can be freely used provided that proper references [1] are mentioned. Furthermore, please consider the WordNet licence in case of commercial use of this ontology.

Download: wordnet171.owl.gz (aprox. 8.5Mb).

This project concern a Document Management System, based on an ontological structure of the data integrated in a Content Management System, allowing the integration of three layers (see Achitecture paragraph). This architecture was motivated by the needs we identified in several projects in direct collaboration with institutions working in the health, bioinformatics, security, and linguistics domains. The goal was to find a system architecture that was as open as possible to allow integration of new components as smoothly as possible. From an implementation’s point of view we based our system on accepted open standards that will allow anyone interested in a similar system to replicate or use it without restriction.

Architecture

There are two main reasons for defining an architecture such as the one presented here:openness, in the sense of offering the possibility to integrate any existing tool, and ease of the integration of the system into an existing information management infrastructure. On one side the system is connected to all internal and external document sources such as DMSs, file systems, data bases or other data repositories storing documents. This information is internally restructured using ontologies in order to give the user more rapid and more accurate access to their documents. The access to the system is guaranteed through a web interface as shown in Figure 1. This architecture in three layers facilitates the integration of existing document resources through the different access mechanisms offered by each system. The integration of the system itself into an existing infrastructure is facilitated as the third layer allows a standard HTTP-integration.