CCE Theses and Dissertations

A Front-Loaded Agile Software Process for Web Applications

Date of Award

2004

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Graduate School of Computer and Information Sciences

Advisor

Michael J. Laszlo

Committee Member

Easwar Nyshadham

Committee Member

William L. Hafner

Abstract

The current state of application development for the World Wide Web is characterized by anarchy and ad hoc methodologies. In recent years, various hypermedia methodologies have been proposed to facilitate the deployment of Web applications. However, no standard has emerged to fulfill the need for a systematic and methodological approach to complex and dynamic Web application development. The primary goal of this dissertation was to elaborate a software development process for Web applications that focuses most of the developers' efforts and creativity on the phases that determine software requirements, on analysis and design, but that is, at the same time, flexible and adaptable so developers can respond to changes in requirements without major cost overruns or delays. In order to develop an effective process, the researcher examined the recommended tasks for Web development as provided by the literature. As a result, the new software process defines the following phases: Feasibility Study, Requirements Definition, System Specification, System Design, Program Design and Development, System Test, Implementation and Production, and Maintenance.

The Web software process was developed for a particular initial-level organization. A statistical instrument, the Rating and Evaluation Guide (REG), which utilizes data gathered via questionnaire, was used to measure the perceived effectiveness of the developed software process. The new process was evaluated against a standard development methodology from the literature. The evaluation was performed by randomly selected teams of developers in an initial-level organization, and the evaluation revealed that the new Web application development process was perceived to be more effective than a generic software process provided by the literature.

The scores for all twenty-six evaluation items in the REG were higher for the new software process than the generic methodology. Additionally, the new software process achieved an overall score of 175 points on the REG scale, a high rating score, which is substantially higher than the anticipated minimum score of 132. Furthermore, the new software process scored 41 out of 48 maximum points in the REG Properties category, which exceeds the anticipated minimum score of 40. A high score for the Properties category reveals the inherent attributes of a high quality methodology are included in FLASOFTi, and the high scores for all six distinct properties in this category indicate that the method's overall quality has been recognized by the participants as quite high.

The findings of this study will be of practical value to initial-level organizations in which the corporate culture tends to require planning development projects thoroughly before working on them. Although the researcher recognizes that the "one-size-fits-all" approach is not appropriate in applying a software process to Web projects, that the factors that influence the development of a Web site are complex and vary from organization to organization, most computing environments that require small to mid-size Web application development projects will be able to use the new method with minor revisions to suit both the project and the organizational culture.

This document is currently not available here.

  Link to NovaCat

Share

COinS