PhD Oral Exam – H M Ishrar Hussain

Abstract

The importance of early effort estimation, resource allocation and overall quality control in a software project has led the industry to formulate several functional size measurement (FSM) methods that are based on the knowledge gathered from software requirements documents. The main objective of this research is to develop a comprehensive methodology to facilitate and automate early measurement of a software’s functional size from its requirements document written in unrestricted natural language.

For the purpose of this research, we have chosen to use the FSM method developed by the Common Software Measurement International Consortium (COSMIC) and adopted as an international standard by the International Standardization Organization (ISO). This thesis presents a methodology to measure the COSMIC size objectively from various textual forms of functional requirements and also builds conceptual measurement models to establish traceability links between the output measurements and the input requirements.

Our research investigates the feasibility of automating every major phase of this methodology with natural language processing and machine learning approaches. The thesis provides a step-by-step validation and demonstration of the implementation of this innovative methodology. It describes the details of empirical experiments conducted to validate the methodology with practical samples of textual requirements collected from both the industry and academia. Analysis of the results show that each phase of our methodology can successfully be automated and, in most cases, leads to an accurate measurement of functional size.