Traceability model for test effort estimation to support software change management

Abstract

In the last decade, the management of software projects has become a challenging task. The latest published figures on the status of software projects indicate a large failure rate, which has created a crucial challenge for project managers. In software maintenance, the impact of software changes is an important aspect due to the evolving environment of the software development life cycle. Software change is necessary to keep software product current and reusable. Software evolves over time due to specific changes during software development and maintenance at every point and the management aspect of its modifications can become more complicated and potentially risky. Many of the current traceability approaches and tools are devoted to and restricted to high-level objects such as specifications but fewer capabilities are made available to handle lower-level artefacts such as classes and codes. While test effort estimation has been in place for decades, it remains a major challenge for software project management to make accurate estimates and, ultimately, to successfully complete the software project. The aim of this research is to investigate the need and significant use of the traceability model for test effort estimation to accommodate software changes in maintenance tasks. This research proposed a new traceability model for test effort estimation to support software change management that may help software project managers to make better decision on the software changes. This proposed model can support the change management tools to allows a dynamic change of traceability structure in response to operation changes such as addition, insertion, and deletion with test effort estimation during the changes. The proposed model used a case study of a completed software development project with an entire collection of software documents called on-board automobile (OBA) and evaluated by Nominal Group Technique and feature analysis. The results showed that the key components and elements are located at acceptable level and can be applied whilst the score of more than 70% is achieved. Hence, the evaluation results proved that the proposed model and its prototype are acceptable and significant to support software change management.