An intelligent software design argumentation methodology for capturing and analyzing design rationale from multiple perspectives

Software design is usually a collaborative endeavor involving multiple stakeholders of a system. Design issues are identified and resolved through dialog among these participants. In many cases, since the dialog process is not managed in relation to the system requirements, it is not clear to participants how the various design options satisfy different stakeholder's requirements. Furthermore, since the dialog is not usually recorded in documentation, it is generally not possible to reconstruct the rationale for design decisions at a later stage in the software life cycle. This dissertation presents an intelligent software design argumentation methodology. It is based on an argumentation model that allows software requirements to be related to design dialog. An intelligent design dialog process based upon this argumentation model is introduced. The dialog process provides for dialog management and for design rationale capture. Fuzzy inference is used to identify conflicts, to simplify arguments, and to calculate a measure of the favorability of each position in a design dialog. This methodology provides mechanisms for the explicit representation of perspectives and design options, for the identification of implicit conflicts, and for assessment of the impact of design options on stakeholder's specifications. The model is illustrated by a case study and a detailed example of the argumentation process is presented.