A wiki is like a blank sheet of paper: One can use it to write love letters or business reports, print on it or cover it with doodles. While this might be appreciated by some, its lack of structure may be threatening to others. Structures, if dictated by the system, can present an obstacle - but they are invaluable when they support the user in carrying out tasks.

Who would be able to file a tax declaration in a timely manner on just a blank sheet of paper, i.e. without the instructions presented by the fields, their labels and order on the form sheet?

Providing a flexible set of structures that are applicable to different (and yet unnumbered) types of content is the main objective of the concepts and technologies, which together form the Semantic Web. The main principles on which these technologies rely are, firstly, the fact that both structures and their relationships can be made explicit and that, secondly, these structures and relationships can be processed further applying the methods of information technology (e.g. algorithms, or so-called “reasoning”).

Open and closed-world assumptions

The peculiar aspect about Semantic Web technologies – and the point in which they differ from, for instance, traditional database methods – is that they are based on a so called open world assumption. Its opposite, the closed world assumption, is what our thoughts and perceptions of the world rely on most of the time.

A closed world assumption works like a time table: If it says in the timetable that trains are running at 7:16, 7:37 and 7:58, I will ‘naturally’ assume that there will be no train service at 07:25. Databases, for instance, rely on closed-world assumptions.

An open world assumption neither rules out the existence of the mentioned 7:25 train, nor the existence of the legendary white raven  – here, just because something is absent doesn’t mean that it does not exist.

KiWi takes into account both approaches: while relying on a closed world assumption for reasoning purposes (“if the timetable contains no train at 7:25, there is no such train”), the KiWi system is still open to easily add new information and also to integrate with external services.

Semi-structured representations of knowledge

And other than, for instance, a database where an empty field is going to produce an error, the Semantic Web is semi-structured: It does not enforce the use of certain structures, but checks whether the available information is consistent with a given model of knowledge representation, called ontology.

An ontology represents the concepts and the relationships between those concepts within a given domain. To cite just two examples of existing ontologies: Dublin Core is a simple ontology used for documents and publishing; FOAF (Friend of a Friend) is an ontology describing persons, their activities and their relations to other people and objects.

Unlike a database, ontologies are designed in such a way that they remain flexible for future expansion. Ontologies can also be used as a ‘translation layer’ to interpret and open the content of a database – but if a consistency check fails, this does not result in a fatal error: At best, one can work around an inconsistency, at worst, the ontology cannot be used for the given purpose.

Within the context of collaborative media (e.g. Wikis), Semantic Web technologies can be instrumental in helping ward off the ever-imminent threat of chaos by backing up user annotations with ontology-based structures that can be used for more efficient search and navigation, and furthermore help moulding social interaction, e.g. by enhancing relationships between users with Friend of a Friend data. Semantic Web technologies also promise to support the user e.g. by personalizing the appearance of the system based on background and interests of the user, as well as on the context of the content currently displayed.

Next article in this section:

Ultimate Support: The Five Dimensions of User-Centredness  

 Image by Yorky on Wiki Commons