I was surprised that a lot of things for the rest of the day were quite technical: social issues “around communities” and then jump to presenting systems to support them. I’ll dig out the links to demos and sites and post them later.
My taking from this session is about choices to make during ontology building process and some implications (I’d like to make a mind map out of it, but don’t have any software here, but at least it’s something)
- one expert vs. group involved
- expressiveness (detailed and complex ontology) vs. maintainability (simple ontology)
- automatic vs. manual
- automatic: quality is questionable
- when ontology is good? mainly depends on user tasks (also: ontology in corporate settings usually reflect power distribution)
- tools to use and their costs?
- change: too fast if generated on the fly
- costs: people (+learning curve to learn about ontologies), time, iterations
- process can be as valuable as ontology itself: commitment + questioning mental models and reflection
- depends on company’s type/size/culture (diversity)
- how to combine/relate group ontologies into organisational ontology -> ontology mapping
- multiple ontology mappings are not easy (usually some common structure for all ontologies is needed)
- monodirectional mapping (objects from a certain system, e.g. database, are mapped into an ontology, so it can be used to retrieved them
- lessons learnt from db modelling
- unstructured data is difficult to extract/create an ontologies
- ontology ownership
Note: You may be surprised why I’m talking about ontologies. At work I’m involved in a process of ontology-building, so I have a lot of practical questions to be answered. My interests are not so much about technical details, but more in social and organisational issues around it. Where do you start in order to build an ontology? Whom do you involve? What roles to they take? What tools can help? Will be an ontology accepted by others? How do you update and maintain an ontology? Any references that answer these questions are welcome
Tags: I-KNOW, ontologiesArchived version of this entry is available at http://blog.mathemagenic.com/2003/07/03.html#a653; comments are here.
Related posts