An Intelligent Computational Argumentation System for Supporting Collaborative Software Development Decision Making

Background

Argumentation based design rationale capture methods and tools have been developed to capture software development rationale for supporting collaborative development decision making. They are built based primarily on a classical model of argumentation developed by philosopher Toulmin (1958). An earlier method gIBIS (graphical IBIS) represents design dialogs as a graph (Conklin & Begeman, 1988). While representing issues, positions, and arguments, gIBIS failed to support representation of goals (requirements) and outcomes. REMAP (REpresentation and MAintenance of Process knowledge) extended gIBIS by providing the representation of goals, decisions, and design artifacts (Ramesh & Dhar, 1992). The REMAP work focused on capturing the process knowledge, i.e., the history about the design decisions in the early stages of the lifecycle of a project. Unfortunately, these systems lack intelligent capabilities to reason about argumentation. As opposed to these systems, Sillence (1997) proposed a more general argumentation model. His model is a logic model where dialogs are represented as recursive graphs. Both rhetoric and logic rules are used to manage the dialog and to determine when the dialog has reached closure. Potts and Burns (1988) outlined a generic model for representing design deliberation and the relation between the deliberation and the generation of method-specific artifacts. In their model, design history is regarded as a network consisting of artifacts and deliberation nodes in which artifacts represent specifications or design documents, and deliberation nodes represent issues, alternatives or justifications. They highlighted the importance of recording of design deliberations. A main drawback of this model is that it does not support decision making. HERMES (Karacapilidis & Papadias, 1998) is a system that aids decision makers to reach a decision, not only by efficiently structuring the discussion rationale but also by providing reasoning mechanisms that constantly update the discourse status in order to recommend the most backed-up alternative. It is an active system which not only captures the informal organizational memory embodied in decision making settings but also helps the user during the decision making process. However, a drawback of this system is that the weighting factor is not effective as it does not relate to the position entered.