Celestine A. Ntuen and Kanwal Hanspal
The Institute for Human-Machine Studies, College of Engineering
North Carolina A&T State University, Greensboro, NC 27411 USA
Objects are traditionally known as artifacts used to describe both mental and physical entities. However, objects are usually described by attribute -- value relationships. Such description makes objects and their technical tools such as Object-Oriented Programming (OOP) limited in application to situations in which knowledge of task are to be shared between and among several agents. Obviously, the single object inheritance metaphor is no more adequate.
The history of OOPs ( Danford & Tomlinson, 1988; Coud & Yourdon, 1991) suggests that it emerged as a knowledge representation metaphor whose primary goals were to support property inheritance in a hierarchy of information space. Recent observations and critics of OOPs ( Trajkovic, Gievska, & Dacev, 1995; Pen & Carrico, 1993) note that OOP paradigms do not cover a global set of object knowledge: for example, while objects in the abstraction hierarchy space can inherit properties of the parent in the upper echelon, they do not have the "intelligence" to share their domain task knowledge with other objects, especially when collaboration and execution of conjunctive tasks are desired. In other words, in agent-based and collaborative problem solving systems and interfaces, O-A-V (object-attribute-value) triplet in OOP representation falls short of its acclaimed performance ( Mahfoudhi, Abed & Augu, 1995).
Evolving software systems developed from agent-based paradigms require, among other things, shared knowledge in an object space. There are recent interests in the HCI community to develop more intelligent user interface (IUI) using object properties in order to achieve this knowledge sharing goal ( Gorlen, 1990; Vlissides and Linton, 1990).
In our approach, we view an intelligent object (10) as an artifact to support human activity. The design constructs are based on representation of task