Software can be considered a live entity, as it undergoes many alterations throughout its lifecycle. Therefore, code can become rather complex and difficult to understand. More specifically in object-oriented systems, classes may become very large and less cohesive. In order to identify such problematic cases, existing approaches have proposed the use of cohesion metrics. While metrics can identify classes with low cohesion, they usually cannot identify new or independent concepts. In this work, we propose a class decomposition method using an clustering algorithm based on the Jaccard distance between class members. The methodology is able to identify new concepts and rank the solutions according to their impact on the design quality of the system. The methodology was evaluated in terms of assessment by designers, expert assessment and metrics. The evaluation showed the ability of the method to identify new recognizable concepts and improve the design quality of the underlying system.

Software measurement has been of interest to software engineers for almost as long as software has been developed. While the evolution of systematic processes of software development has seen a trend away from reliance on the expertise of individual software developers alone to ensure software quality, systematic processes of software measure development have not evolved to a similar extent. The problem with defining software measures according to an informal process is that the quality of measures can be highly dependent on the expertise of the individual measure developers. If a systematic process of software measure development were defined, that promoted transparency and objectivity in measure development, then this systematic process could support the development of high quality measures by less expert users. In this thesis, a systematic process of software descriptive measure development is described and demonstrated. The approach taken to defining this systematic process is to investigate the various processes by which currently available software descriptive measures have been developed. These processes are then amalgamated with an established systematic method of measure development used in the field of social science. Applying the stages of measure development thus identified to the task of developing measures to describe C++ class and object modularity tests the feasibility of this measure development process. Insights gained through this testing provide feedback to further refine the process. In this way, a systematic process of descriptive software measure development is defined alongside the definition of a set of measures that provide a detailed description of the complex software characteristic of modularity. The products of each stage of this measure development process assist a user to validate the measures with respect to an intended application, and to analyse and interpret the measurement data obtained by applying the measures to a software system. This is demonstrated in a case study that also provides an indirect indication of the quality of the process by which the measures were developed. The major contribution of this work is the systematic process of descriptive software measure development, as it has a wide application and can be used to develop measures to describe many software characteristics of interest. A second important contribution is made by the set of measures of C++ class and object modularity developed to demonstrate this systematic descriptive measure development process.

This research studies the impact of Object Oriented Programming Systems (OOPS) and their underlying concepts on Land and Geographic Information Systems (LIS/GIS) in Australasia. This research considers GIS software development and conceptual data modelling aspects, and the strong relationship with proposed spatial data transfer standards. / Conventional programming techniques appear to have reached their limit in coping with complex and diversified applications. "Something better' is envisaged for future software developments and data models in LIS/GIS. The relatively new object-oriented design method is reviewed and a much improved object-oriented software module for the daily maintenance operations in a Digital Cadastral Data Base (DCDB) is presented as an example. This development allows a significant improvement of the spatial accuracy of DCDB systems and its graphical user interface (GUI) represents a much better data quality visualisation tool. / The choice of the right conceptual data model for GIS has a strong impact on proposed spatial data transfer standards and the way in which future Australian GIS communities will 'view' or model their real world. It has therefore been important to critically review these proposals in the Australian context.

Thesis (MSc.) - Swinburne University of Technology, Faculty of Information & Communication Technologies, 2002. / Dissertation for the degree of Master of Science (Research), Faculty of Information and Communication Technologies, Swinburne University of Technology, 2002. Typescript. Bibliography p. 142-146.

Thesis (Ph. D.)--Virginia Polytechnic Institute and State University, 1993. / Vita. Abstract. Includes bibliographical references (leaves 193-204). Also available via the Internet.

Thesis (M.S. in Computer Science) Naval Postgraduate School, March 1992. / Thesis Advisor: Wu, C. Thomas. "March 1992." Includes bibliographical references (p. 83-85). Also available in print.

Thesis (Ph. D.)--University of Texas at Austin, 2006. / Vita. Includes bibliographical references.

Thesis (M.S.)--Virginia Polytechnic Institute and State University, 1993. / Vita. Abstract. Includes bibliographical references (leaves 71-77). Also available via the Internet.

Thesis (Ph. D.)--University of Hong Kong, 1997. / Includes bibliographical references (leaf 220-248).

Thesis (M. Phil.)--University of Hong Kong, 1996. / Includes bibliographical references (leaf 58-60).