Design of knowledge-based integrated software process improvement tools

The design of QUEST (software QUality Evaluation System and Technology) tools, and especially of its repository, is discussed. QUEST tools are composed of a software process assessment tool (QUEST-A), a software process improvement tool (QUEST-I) and a repository for software process information (QUEST-R). We place the emphasis of this paper on the design of QUEST-R, which acts as a process asset library (PAL) and a meta-process support tool. The meta-process knowledge can be accumulated from several cases of process improvement, which can be initiated by software process assessment and process metrics. The results or outputs of QUEST-A, QUEST-I or the other tools act as a driver of knowledge inference. To build QUEST-R, a knowledge-based software process model (KB-SPM), which incorporates meta-process objects and their relationships, is proposed. Several factors, including environmental changes and various process improvement strategies, are explicitly modeled as frames. QUEST-R (1) makes it possible to generate and customize project process models in various application domains; (2) makes it possible to reuse the software process objects; and (3) makes it easy to accumulate and enhance meta-process knowledge. Besides supporting the meta-process phase, to increase the effectiveness of the meta-process, the following facilities are proposed: (1) rules for control knowledge; (2) more affecting factors for the generation of software processes; (3) a dependency structure, for describing improvement items and their interdependence; and (4) a mechanism to automatically update the knowledge.