KNOWLEDGE-BASED COMMUNICATION AND MANAGEMENT SUPPORT IN A SYSTEM DEVELOPMENT ENVIRONMENT (ARTIFICIAL INTELLIGENCE, SOFTWARE ENGINEERING, EXPERT SYSTEMS, SPEECH ACT THEORY).

摘要

Artificial intelligence techniques can help software development environments support the variety of activity involved in building complex, evolving software systems and managing their projects. The main contribution of this dissertation is the identification of project management and communication support as an important domain of an effective software development environment, and the application of speech act theory to that domain. A framework, or paradigm, was designed for such an environment using a knowledge-based, program synthesis approach from artificial intelligence. A pilot communication and management support environment (CMS) was implemented. CMS supported an existing project to build a complex software system and helped its users, who could be system designers, managers, maintainers, or end-users, by providing capabilities for project management, documentation/help, critiquing and intelligent communication.;Initial, informal studies of system/user interaction helped form the theory that people, while interacting with a computer system, perform communication or management acts, such as questioning, informing, requesting, critiquing, or planning. An act taxonomy was created and the structure and relationships among some of these acts were formally represented. CMS incorporates a theory of acts and uses a knowledge-based, synthesis approach to process acts. The processing of acts uses assertion, inference, and synthesis to transform acts to other related acts, perform actions, such as sending messages or answering questions, annotate a project model, or collect statistics about system activity. The project model includes structures and properties for project concepts, such as components and tasks. These are used to configure the system and to organize project information and act data. A representation of time as hierarchical periods was also created. Protocols were developed for project activity and formally represented as rules that refer to the project model and acts. Constraints on CMS activity and on the project model were also represented as rules. The project model, the acts, and the rules are all part of the environment's knowledge base. The power of the environment comes from using this knowledge to reason about the project, rather than just to retrieve facts.