Academic
Publications
The expert system life cycle: what have we learned from software engineering?

The expert system life cycle: what have we learned from software engineering?,10.1145/97709.97711,Anita J. La Salle,Larry R. Medsker

The expert system life cycle: what have we learned from software engineering?   (Citations: 3)
BibTex | RIS | RefWorks Download
The assimilation of expert systems into the business environment has been rapid. A number of well-publicized expert system success stories have helped convince management that expert systems can play important roles in information systems. However, the unconditional acceptance of a new discipline without recognizing some painfully-learned lessons of the past is inherently dangerous. The speed of expert system technology transfer from research laboratories to business systems, while exciting, nonetheless has the potential to contribute further to the “software crisis. ” Some aspects of expert system development are just like traditional software development; however, some phases have no parallels. We need to articulate new standards that address the unique features of expert systems development and incorporate the lessons from software engineering. A. Introduction Expert systems are different from conventional software systems in a number of ways. For example, -expert system development is highly evolutionary -- prototypes are built; users provide feedback; managers analyze effectiveness; prototypes are expanded. This evolutionary aspect has strong implications with respect to program integrity, testing and maintenance. -depending on the type of application, an expert system may be near the limits of AI technology and therefore require exploratory development. -the process of knowledge acquisition has no parallel in traditional software development. The issues of knowledge base verification and conflict resolution within a knowledge base likewise have no parallels. -traditional software development has not used heavy prototyping. Reliable rules and procedures simply do not exist. Expert system practitioners have been casual about documenting their experiences with heavy prototyping. -change or expansion of a knowledge base also has little parallel in the maintenance phase of the traditional software development life cycle. Rapid prototyping may lull managers into a false sense of security about the complexity of developing expert systems. At the same time, unused prototypes are expensive and can complicate the software development process by creating a throw-away attitude. Two major issues in the expert system life cycle are (1) how traditional software engineering life cycle phases can be applied to expert system development and (2) new phases that are needed to acknowledge the unique non-traditional aspects of expert system development,
Cumulative Annual
View Publication
The following links allow you to view full publications. These links are maintained by other sources not affiliated with Microsoft Academic Search.
Sort by: