Ontologies provide a promising method of organizing and representing information and knowledge. Since the presentation of the Semantic Web vision in 2001, which is based on ontologies acting as ?knowledge stores?, significant amount of resarch work has been done in this field.
This thesis deals with ontologies by focusing on the Topic Maps and OWL standards, which have become the most important formats for representing knowledge in ontologies.
Despite the relatively broad adoption of ontologies in the research community during the last years, creating and using ontologies is still no straightforward task.
This is essentially due to the lack of proper tools that are easy enough to use in practice, but also due to some inherent limitations that apply to ontologies. Most issues related to ontologies pertain to one of these two groups of problems, which could be described as ?implementation challenges? and ?conceptual issues?, respectively.
The thesis first introduces ontologies and related concepts like the Semantic Web and also presents an in-depth view of the Topic Maps and OWL standards.
Three major topics are identified as being especially challenging when implementing ontologies in real world applications: providing persistent, scalable storage, performing complex queries on ontologies, and deducing implicit information by using special inference engines.
Apart from these challenges on the implementation level, a number of crucial issues at the conceptual level of ontology representation formats are also discussed, such as the absence of contextual constraints, the inability to properly represent inconsistencies and the lack of methods for expressing uncertainty in knowledge bases.
