Collaborative experience-based factory model for software development process

• Main Author: Hanafiah, Mastura
• Format: Thesis
• Language: English
• Published: 2019
• Subjects: Computer software - Development; Knowledge management; Expert systems (Computer science)
• Online Access: http://psasir.upm.edu.my/id/eprint/83775/1/FSKTM%202019%202%20-%20ir.pdf; http://psasir.upm.edu.my/id/eprint/83775/
• Institution: Universiti Putra Malaysia

A key aspect in software development (SD) is the management of its knowledge and experiences. Since many decades ago, organizations have been valuing the experiences and know-how of their employees. One of the frameworks that enables learning and continuous improvement is the Experience Factory (EF) framework. Yet, previous literatures have reported that EF is hard, costly, and risky, and requires considerable amount of effort to implement. Although there are several evolutions on this framework over the years, however, the works have been declining in the recent years due to the unwillingness of the organizations to invest and due to the unclear benefits to the employees. In addition, knowledge management (KM) issues in SD still persist until today and software organizations are still striving to learn from previous experiences. This dissertation proposes a model for managing SD knowledge and experiences based on the EF approach, namely EBF-SD, to address the limitations of EF as well as to overcome the KM issues for SD process in a collaborative environment. The proposed components are SD Process Knowledge Base, Community of Practice Influences, Knowledge Management Process Enablement, and Technology & Infrastructure Support. In order to implement EF, its goals must be clarified and measurable, thus, the components are evaluated against the EF goals. Qualitative methods such as expert review and pilot study are conducted to verify the initial conceptual model, while quantitative method is used to investigate the relationships between the components and EF goals. Data reliability and construct validity are examined via Rasch Analysis and Factor Analysis, while hypothetical relationships are examined using correlational analysis, multiple linear regression and Partial Least Squares of Structural Equation Modeling (PLS-SEM). Empirical study indicates that the components have positive and significant relationships towards EF goals whereby 6 out of 7 hypotheses are supported. Empirical evidences also reveal that technological support is the main significant factor towards the achievement of EF goals. Based on these findings, a prototype is developed to translate the model into a working system, as a proofof- concept, by implementing the proposed components into appropriate functionalities and relevant technological approaches. Evaluation of the prototype via descriptive statistics and PLS-SEM reveals that the prototype is beneficial and significantly contributes to the achievement of EF goals. Other findings suggest that knowledge quality has higher influence in terms of system usage and user satisfaction as compared to system quality. The overall research findings demonstrate that the proposed model is adequate, significant and accepted by the software practitioners in the context of collaborative software development environment.