Information Technology Reference
In-Depth Information
be passed down to exploit a broader range of resources, or an authorized
replacement can be located for a failed actor.
These are areas where the current models are just the first steps towards
progressively higher-level descriptions of business practices, social norms, or
even aspects of the legal system. It is not inconceivable that, in the future,
legislation may be supported by suitable formal models of the community
concerned. There are already a number of examples of the formal modelling
of quite complex laws and regulations in the scientific literature [57,77,87,91].
14.6 Expressing Semantics
So far we have concentrated on issues about expressing behaviour, but
there are also many equally important issues in the field of information mod-
elling. Here again, the main issue is how to establish a robust link between a
formal model of knowledge and the real world with all its complexity and am-
biguity. The models used are generally based on the construction of a large
collection of terms, labelled with attributes and linked by relations. These
knowledge networks are often called ontologies. There is a great deal of judge-
ment and skill involved in constructing a good ontology because of the need
to balance coverage and precision.
As usual, there is a need to limit the resulting complexity, and this is
normally done by defining ontologies to cover specific subject domains. These
then need to be related, usually via an overarching upper ontology (sometimes
called a foundation ontology). This involves the specifier in facing problems
very similar to those occurring when creating a federation (see chapter 11).
The problem here is, again, that combining different domains can easily lead
to inconsistency and ambiguity.
Consider, for example, the interaction between a technical ontology for
the mobile phone industry and a general commercial ontology covering sale
of products. The mobile phone ontology might cover terms like handset and
SIM, and express the whole-part relationship between them. The commercial
ontology might describe spare-part, ownership and warranty, indicating that
ownership of a product implies ownership of its parts. However, when we
sell a mobile phone it often involves a preinstalled SIM, perhaps on a pay-as-
you-go basis. The simple composition of these ontologies as separate fields of
knowledge would conceal a whole range of issues around the more complex
relationship between phone supplier, carrier and customer.
Solving these problems requires a detailed analysis of what it means to
say a term applies to a real-world entity, and needs much of the same modal
machinery outlined previously for the handling of obligations.
