POND: A Knowledge Representation Language which Facilates Requirements Capturing

POND, a KL-ONE based knowledge representation language, provides the underlying structure for GATOR, a system used to gather requirements of telephony features. POND differs from a standard KL-ONE language by dividing knowledge into three levels: a Concept Level, an Instance Level, and a Model Level. The Concept and Instance Levels correspond, respectively, to KL-ONE's TBox and ABox. The Model Level represents specialized ABox knowledge (simulations of combining instances). This paper details the types of knowledge represented within each of these levels and how knowledge defined on the higher levels influences the knowledge structure of the lower levels. These details are provided by specific examples from the GATOR system's knowledge representation. AREA: Technology AI TOPICS: Knowledge Representation, KLONE Languages DOMAIN: Software Engineering -Requirements Capturing LANGUAGE: Lisp STATUS: Under research development EFFORT: 1/2 person-years POND: A Knowledge Representation Language 2