Date of Award

1998

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Computer Science

First Advisor

Doris L. Carver

Abstract

In this research, we investigate whether reusable classes can be characterized by object-oriented (OO) software metrics. Three class-level reuse measures for the OO paradigm are defined: inheritance-based reuse, inter-application reuse by extension, and inter-application reuse as a server. Using data from a software company, we collected metrics on Smalltalk classes. Among the 20 metrics collected are cyclomatic complexity, Lorenz complexity, lines of code, class coupling, reuse ratio, specialization ratio and number of direct subclasses. We used stepwise regression to derive prediction models incorporating the 20 metrics as the independent variables and the reuse measures, applied separately, as the dependent variable. Inheritance-based reuse and inter-application reuse by extension can be predicted using a subset of the 20 metrics. Two prediction models for inheritance-based reuse and inter-application reuse by extension were validated using a new set of 310 Smalltalk and VisualAge applications and subapplications. Validation results show that it is possible to predict whether a class from one application can be reused by extension in another application. We also conducted a t-test to test whether the mean metric values between reusable and non-reusable classes are the same. Results suggest that there exists significant differences in the mean metric values between the reusable and non-reusable classes.

ISBN

9780599213166

Pages

142

Share

COinS