Previous: Case models Up: Domain knowledge

Domain schemata

The domain schemata for C&D give the following structure to the domain ontology:

0pt 0pt

• Symptom, associated event pairs. `The form of each piece of knowledge must be such that touching a symptom returns the set of associated events' (from original text).

• Event, differentiation information pairs. `Touching a candidate explanatory event returns a description of the information that will help to confirm or disconfirm the event'. (from original text) This actually comes in a more detailed form:

  0pt

• sep 0pt

  • Events are classified into testable (symptoms) and non-testable events.
  • Events are associated with required events.
  • Events are associated with events that enable an explanatory connection or an explanatory event.

The following operations are directly supported by the domain model:

1. cover-op(+S:symptom, -El:event*).

   Associates a symptom with a set of events. The cover operation handles one symptom only. To cover a set of symptoms an inference has to be made that collects the events of each of the separate symptoms.

2. testable-op(+E:event, -Testable:bool)

   Returns true if the event is testable (i.e. a symptom) otherwise false.

3. required-events-op(+E:event, -Re:event*) Returns the list of required events of an event.

4. enabling-events-op(+E:event, -Ee:event*)

   Returns the list of enabling events of event .

The order in which observations should be made is not represented. In practice different observations often have different cost. This information appears not to be represented explicitly in the standard C&D model.