Wednesday, September 05, 2007

The Need of a Common Ontology

With lots of talks around about Semantic Web, one thing that stands out among all the terminologies used is Ontology. In my earlier post, What is Ontology, I touched upon the basics of Ontology and discussed only the basics of it. Towards the end of the post I proposed to build a common Ontology Server.

While reading Guizzardi and Guarino and after understanding about ontology in bit detail. I went back to read Knowledge Representation (John F Sowa) Chapter 2. But in this process what come out is a belief that if the Semantic Web is to become a reality and ubiquitous then we must have a common way to represent the entities. Which brings back to the need of a Common Ontology Server.

Going via the definition of Ontology given by the philosopher Willard Van Orman Quine the fundamental question of Ontology is "What is There?" and the answer which came back is "Everything". Now if something has to accommodate everything then it has to be be designed to keep every other interpretation of entities in this world.

Then the question arises is: Do we Really Need a Common Ontology?
Before we answer this question, we must ask what we will be loosing if we don't have a common ontology? In absence of a common ontology we might not be able to represent an entity in an uniform way. Uniform way here means by following a common structure and pattern. The rule layer (as part of Semantic Web stack) will not be able to operate properly. This in-turn will make other processing complex and tedious.

In search of solution I came across the paper "Discovering Semantic Web services with and without a Common Ontology Commitment. Jorge Cardoso. Proceedings of the IEEE Services Computing Workshops (SCW'06), 2006". This paper talked about the discovery of semantic web services without having a common ontology to represent them. But then there are few limitations of this methodology as well and it is not suitable for the original vision of Semantic Web.

I will keep working on it and see what I come across next. If you have any thoughts on this drop me a line and I will be happy to hear from you on this.

Until Next Time...:)


Stanley Wong said...

I think the real problem with ontology building is it is a very hard thing to do. Most of it is manually done or semi-automated using artificial intelligence.

Are there some resources on how to build ontologies?

Samir Kumar Mishra said...

Technically I dont see any hurdles building an ontology. We have the basic data structures in place and we also have the right set of technologies available.

The problem however is on the conceptual side. Where the way of looking at any thing differs (most of the time) from person to person.