You are cordially invited by the CSE, to attend a PhD Dissertation Proposal, on the above given title, by Mr. Hamdi Ali Al-Jamimi, PhD student in CSE, on Saturday, 11th May, 2013, at 03:30 PM, in Building 22, Room 132.
Abstract: The software development process can be viewed as a sequence of various interrelated phases. Each phase, in this process, produces new models by utilizing the models built during its preceding phase. Analysis and design phases are the most crucial part of the software development life-cycle. While the analysis phase is related to understanding the given problem, the design phase is related to the formation of a solution for the analyzed problem. However, the literature underlines that the transition from the analysis artifacts to the design artifacts is unclear. The difficulty in moving from analysis to design is caused by the fact that the artifacts of analysis phase and design phase are different in nature. While artifacts of the analysis phase are linked to the user activity, artifacts of the design phase are related more to the available implementation technology. Our literature survey revealed that the problem of automatically moving from analysis to design through utilizing previously developed projects has not caught enough researchers’ attention yet. Accordingly, the goal of this dissertation is to provide algorithms and tools necessary to enable automatic transformation of the artifacts from the problem space (i.e., requirement analysis models) into artifacts in the solution space (i.e., architecture). It is well-known that software reuse, especially reusing the artifacts of these early phases, improves the productivity and software quality. Thus, the idea is to utilize the previously developed projects, specifically the analysis-design pairs, to generate a new design for a given new analysis models. Our focus is to build capabilities that would allow learning transformation rules from available analysis-design pairs.
Dr. Moataz Ahmed is the thesis advisor.