Date of Award

2012

Degree Type

Thesis

Degree Name

Doctor of Philosophy

School

School of Computer & Security Science

Faculty

Computing, Health and Science.

First Advisor

Associate Professor Chiou Peng Lam

Second Advisor

Dr. Jocelyn Armarego

Abstract

It is becoming common to build applications as component-intensive systems - a mixture of fresh code and existing components. For application developers the selection of components to incorporate is key to overall system quality - so they want the `best'. For each selection task, the application developer will de ne requirements for the ideal component and use them to select the most suitable one. While many software selection processes exist there is a lack of repeatable, usable, exible, automated processes with tool support. This investigation has focussed on nding and implementing strategies to enhance the selection of software components. The study was built around four research elements, targeting characterisation, process, strategies and evaluation. A Post-positivist methodology was used with the Spiral Development Model structuring the investigation. Data for the study is generated using a range of qualitative and quantitative methods including a survey approach, a range of case studies and quasiexperiments to focus on the speci c tuning of tools and techniques. Evaluation and review are integral to the SDM: a Goal-Question-Metric (GQM)-based approach was applied to every Spiral.

Share

 
COinS