A document based traceability model for test management

• Main Author: Ami, Azri
• Format: Thesis
• Language: English
• Published: 2017
• Subjects: QA75 Electronic computers. Computer science
• Online Access: http://eprints.utm.my/id/eprint/79255/1/AzriAmiPFC2017.pdf; http://eprints.utm.my/id/eprint/79255/
• Institution: Universiti Teknologi Malaysia

Software testing has became more complicated in the emergence of distributed network, real-time environment, third party software enablers and the need to test system at multiple integration levels. These scenarios have created more concern over the quality of software testing. The quality of software has been deteriorating due to inefficient and ineffective testing activities. One of the main flaws is due to ineffective use of test management to manage software documentations. In documentations, it is difficult to detect and trace bugs in some related documents of which traceability is the major concern. Currently, various studies have been conducted on test management, however very few have focused on document traceability in particular to support the error propagation with respect to documentation. The objective of this thesis is to develop a new traceability model that integrates software engineering documents to support test management. The artefacts refer to requirements, design, source code, test description and test result. The proposed model managed to tackle software traceability in both forward and backward propagations by implementing multi-bidirectional pointer. This platform enabled the test manager to navigate and capture a set of related artefacts to support test management process. A new prototype was developed to facilitate observation of software traceability on all related artefacts across the entire documentation lifecycle. The proposed model was then applied to a case study of a finished software development project with a complete set of software documents called the On-Board Automobile (OBA). The proposed model was evaluated qualitatively and quantitatively using the feature analysis, precision and recall, and expert validation. The evaluation results proved that the proposed model and its prototype were justified and significant to support test management.