لطفا منتظر باشید ...
16.3 Evaluation of the Contributions of This Work
The conflict repository described in this chapter makes substantive contributions to previous work in this area. These include greater expressiveness and content coverage, which in turn help make the repository potentially more effective in supporting prototypical uses.Expressiveness Previous efforts to create conflict knowledge repositories (Matta 1996; Castelfranchi 1996; Ramesh and Sengupta 1994; Feldman 1985) all include either a conflict type taxonomy or a conflict handler taxonomy, and even both, with links between conflict types and the potentially applicable conflict handlers. None of these efforts, however, capture the linkage between collaborative design processes and their characteristic conflict types, nor do they capture the important information encoded by the conflict management meta-process described in this chapter. Finally, they don't take advantage of process abstraction and bundle/trade-off concepts to enable quick discovery and comparison of alternative processes for similar needs. It is our preliminary judgment that the schema presented above captures all the significant aspects of the conflict management information we have encountered in the literature we have reviewed to date.Coverage Previous efforts in this area have produced repositories that are quite small in scale. The taxonomy described in Matta et al. (1998), Feldman (1985), Ramesh and Sengupta (1994), and Castelfranchi (1996) includes no more than about thirty conflict types and handler processes. These efforts in addition focus on individual disciplines. Matta's work, for example, focuses on the concurrent engineering literature, Feldman's on the sociological literature, and Castelfranchi's on multi-agent systems. While one can argue that they provide complete coverage at an abstract level, they necessarily leave out descriptions of a large number of specific, potentially useful conflict management techniques.The repository described in this chapter is significantly larger in scope. It includes roughly the same number of conflict types as those described above but a significantly larger number of conflict management processes (about 200 at the time of writing). The contents of the MIT repository have been drawn from several disciplines including distributed artificial intelligence, sociology, and industrial engineering, as represented by roughly fifty publications from such venues as the Journal of Concurrent Engineering Research and Applications, the Journal of Artificial Intelligence in Engineering Design Analysis and Manufacturing, the Sloan Management Review, the International Conference on Artificial Intelligence in Design, and the National Conference on Artificial Intelligence. Our repository continues to grow, with the support of a continuing three-year grant from the National Science Foundation (Grant No. IIS-9803251).
Better Support for Prototypical Uses The MIT conflict repository has been evaluated only on a limited internal basis to date, so it is premature to draw definitive conclusions about its utility for students, researchers, and practitioners. It is clear, however, that the Process Handbook provides a level of enabling technology that has not been exploited in previous conflict repository efforts. Previous work has resulted mainly in textual documents (with the notable exception of Matta et al. who made the repository available over the Web), and does not include the kind of search, navigation, business process redesign, and structured discussion tools available as part of the Handbook. Previous experience with these tools suggests that they can be powerful enablers. The Handbook has been successfully used, for example, to teach classes at the Sloan School of Management as well as at Babson College. The Handbook process redesign methodology has been applied in several domains, most recently (in cooperation with the consulting firm AT Kearney) to redesign the hiring processes in a major financial services firm. The participants in this study felt that the approach was effective in generating a much wider range of novel and promising process alternatives than would have been uncovered by traditional methods (chapter 12 in this volume).
