Incorporating software agent in personal software process (PSP) automated tool

Hassan, Hazrina (2011) Incorporating software agent in personal software process (PSP) automated tool. Masters thesis, University of Malaya.

[img] PDF
FullThesis.pdf

Download (7MB)

Abstract

Personal Software Process (PSPSM) initiated by Software Engineering Institute (SEI) and Department of Defence, US, was developed to help software engineers to produce highquality products through improving their personal software development processes. With consistent measurement and analysis activities that the PSP proposes, software engineers can identify process deficiencies and make a reliable estimate on effort and quality. However, due to the high-overhead and context-switching problem of manual data recording, software engineers have difficulties to collect reliable data, which can lead to wrong analysis results. This research study aims to develop PSP automated tool called PSPExpert Visualization Agent (EVA) which is designed for software engineers specifically to automate all the processes and fields involved in PSP as well as to introduce the agentoriented concept in the PSP framework. Currently, some developers of PSP automated tools include the agent’s element in their system that focuses on sensor-based tracking of lines of code. However, none of them fully utilize this element for other purposes or provide the agent with a personal assistant for the software engineer. This sensor-based tracking is limited to Eclipse environments such as Java and is not compatible with the use of web platforms as a PSP automated tool. The PSP-EVA system is required to perform several tasks. There are three agents that exist within the system: InterfaceAgent (IA), TaskAgent (TA), and SearchAgent (SA). By incorporating the multi-software agent in this automation system, PSP-EVA is able to have ready-access to applications and also to provide a visual representation of performance to the user. Additionally, evaluation of software engineers and team performance can be easily derived and evaluated statistically. The implementation of this tool, will then measured in terms of data accuracy and tasks performance. It is expected that, the time to collect and perform data analysis and resources will be minimized up to 75% less with 99.9% of data accuracy provided compared with conventional approach.

Item Type: Thesis (Masters)
Uncontrolled Keywords: Software agent, Personal Software Process, PSP, Software Engineering Institute, SEI, PSPExpert Visualization Agent
Subjects: Z Bibliography. Library Science. Information Resources > Z665 Library Science. Information Science
Depositing User: MS NOOR ZAKIRA ZULRIMI
Date Deposited: 19 Jul 2013 07:56
Last Modified: 19 Jul 2013 07:56
URI: http://repository.um.edu.my/id/eprint/610

Actions (login required)

View Item View Item