Date of Award
2013
Document Type
Dissertation
Degree Name
Doctor of Philosophy (PhD)
Department
Computer Science
Committee Chair
Wei Li
Committee Member
Ramazan Aygun
Committee Member
Mary Ellen Weisskopf
Committee Member
Williams Wessels
Committee Member
Feng Zhu
Subject(s)
Object oriented programming--Computer science, Computer software--Quality control, State-space methods
Abstract
The syntax-based (structural) software metrics are quantitative measures of programs. They have been used to predict various software quality factors. There are many studies on software structural metrics over the years. However, these metrics lack a formal foundation that explains the nature of measurement. The objective of this research is to propose a Z-based theoretical framework for the extraction of object-oriented (OO) design state space and structural metrics to help guide and improve the design and development of software systems using mathematically rigorous methods. To show the framework works, we selected several representative categories of syntax-based metrics that measure different design characteristics, such as the measurements of classes, data attributes, methods, parameters, inter-class coupling, and inter-module coupling in OO systems, for demonstration. The framework provides a graph-based design representation of object-oriented programs, the Z formalization of the design state space, and the methodology for defining software metrics. To support the framework, we developed a comprehensive toolkit - the Design State Space (DSS) - to support the parsing of source code to generate the design state space relation sets based on the Design State Space Graph (DSSG) modelling and the collection of metrics based on the Z formalization of the structural metrics, which depends on this design state space. Also, in this research, we discuss two metric properties based on the proposed formal framework to help build a better perdition model using metrics as independent variables: Metrics Orthogonality (MO) and Degree of Non-orthogonality (DN). We conduct an empirical study using data collected from open source systems to validate the properties.
Recommended Citation
Alomainy, Raouf, "A Z-based formal framework and comprehensive toolkit for the extraction and analysis of object-oriented design state space and structural metrics" (2013). Dissertations. 10.
https://louis.uah.edu/uah-dissertations/10