Pattern recognition software development methodology (PRSDM) based on design pattern recognition techniques and agile methodologies.

摘要

Literature offers little insight into the effectiveness of the Software Development Methodologies (SDM) at different life-cycle phases of development. In comparison, plan-based methodologies, such as Waterfall, Spiral, and Iterative may offer discipline and standards in development practices. However, plan-based methodologies assume that a full listing of requirements is available prior to development and invite limited interaction with the customers prior to final product delivery. The Agile process, on the other hand, as a process, embraces the "divide and conquer" technique leading to a faster product development turn around and more client involvement. This research introduces a model that extracts and empirically tests three main software development driver constructs: Facilitating Conditions, Affect, and Perceived Consequences. These constructs are derived from a well-known human-behavior model. Our model, although it can be used for evaluating the effectiveness of different software methodologies, will be used to measure the effectiveness of the Agile methodologies at requirement, design, and implementation stages. Hypothesis testing for our model favored most the use of Agile process at implementation stage and least at design. Thus, our research introduces anew Agile methodology to enhance the effect of Agile methodologies during design life-cycle stage. We use one of Agile's methodologies, SCRUM, as a base for our methodology and then inject the Design Pattern Recognition techniques and the RAD's Time-boxed concepts to form a methodology that combines the best bred of the above-mentioned concepts.