When information can be published on the web in a reusable format like XML, one could use it to make conclusions. In the following example I will use some information in RDF to deduce something using an inference engine called Sirli (see http://www.ontoprise.com/ in the free downloads section).

The data is about me, my mother and the mother of my mother. See it here : http://www.lowas.be/cms/public/SemanticWeb/jfd.rdf.xml

Sirli will help me to find my grand mother.

How ?

It will read the information in the RDF and use some rules. The rules are simple :

FORALL O,V Maman(O,V) <- O("http://description.org/schema/Maman"->>V).

FORALL O,X,G GMaman(O,G) <- Maman(O,literal(X)) and Maman(X,literal(G)).

Those rules mean that if there someone is mother of someone if there is a relationship denoted by http://description.org/schema/Maman between two persons. The second rules that the grandmother of someone is the mother of the mother of that person.

Using Sirli, asking the following question :

FORALL O,G <- GMaman(O,G).

The response is the following :

O = "http://www.jfdeclercq.com/#jfdeclercq", G = "http://www.jfdeclercq.com/#augusta"

and

O = "http://www.jfdeclercq.com/#ndeclercq", G = "Who Knows ?"

This means there are 2 grandmothers in my information : Augusta and Who Knows (the grandmother of my mother in fact).

By combining such relationships into rules and queries, the web can become a huge knowledge management system.

As described in the book "The Semantic Web", the challenge will be to identify which source can be trusted. Indeed, my mother does have a grandmother... believe me.

IT Internet