K3: Some Critical Reflections
Idea difficult to understand for „normal users“
In the first instance, the expressiveness wished by the users is somewhat „between lists and trees“ :-(
Who really exploits background knowledge and complex inferences?
Who builds ontologies? How? (future work in text-based construction support planned)
Indexing takes time: automatic text categorization (current work at DFKI)
More theoretically: are index structures really ontologies? (Welty, van der Vet, Schmiedel made the point already)
Also: are ontologies what we need for retrieval and information access (think about „navigation ontologies“ and individual information spaces as projections from the underlying ontologies)