We believe that clarifying an ambiguity in the use of instances and classes is important for domain modeling and knowledge representation in general. While supporting the notion of spanning objects and multiple universes of discourse in modeling languages is a necessary part of this clarification, it does give rise to some interesting questions that must still be addressed. The most prominent question is on the relationship between the universes in which a spanning object exists. What relationships can be expressed between objects in different universes? Can inferences made about an spanning object in one universe be used in another universe in which the object exists? If such mapping across universes is allowed, it may affect the completeness and tractability of the inference mechanisms, since the essential nature of spanning objects is second-order.
Generated with CERN WebMaker