Evaluating Object-Oriented Designs

Colorado State University, Fort Collins CO

Investigators

Abstract

The project develops techniques to quantify design attributes of object-oriented software in terms of architectural structures and patterns, and demonstrates that these design attributes are related to external quality attributes such as maintainability, reusability, testability, and reliability. The focus is on identifying design structures and patterns that will make software easier to adapt and test. Design measurement is based on the structure of interconnected objects including the links between objects, the properties of these links, and the properties of individual object classes. Design patterns are identified through program and design analysis and the use of existing design pattern recognition technology. External design quality evaluations are based on process data from commercial organizations and analytical evaluation of change difficulty. Relationships between design attributes and external quality are identified by examining commercial software engineering data and through analyses of the connection between internal and external attributes. Results from this work demonstrate costs and benefits of alternative object-oriented software designs. The work should lead to improved design methods to produce software that is easier to adapt, extend, and test as it evolves.