What is default reasoning good for? Applications revisited

This paper describes the basic advantages of using defaults in the representation of information and in reasoning with information and presents applications of default reasoning in various areas, such as software engineering, information retrieval, law, graphics design etc.

Abstract

Default reasoning comprises methods of reasoning with uncertain and incomplete information which share the idea of using default rules to represent plausible conclusions. Even though the applicational scope of default reasoning was clear right from the beginning, research focussed heavily on theoretical aspects and neglected for a long time pragmatic and applicational questions. This, coupled with a presentation that is usually too technical for potential applicants of this technology has led to the impression that default reasoning (and other forms of nonmonotonic reasoning) have very little relevance to practice. This paper seeks to counter this view. It describes the basic advantages of using defaults in the representation of information and in reasoning with information. And then it presents applications of default reasoning in various areas, such as software engineering, information retrieval, law, graphics design etc.