The increase in the power, reliability and proliferation of computer hardware and the decrease in hardware cost contribute significantly to the exponential growth in software development. Automating software development is seen as a key approach to combating the spiraling cost of software. This study takes a two-fold approach to examining automated software development. First, different levels of complexity are considered: high-end, mid-level, and low-end. Second, the different components of the process are considered - selection of a software development life cycle (SDLC) model followed by implementation of the various phases of the selected model. This provides an incremental approach to automating software development. This paper focuses on the implementation of an expert system that addresses SDLC model selection and limited requirements analysis. This expert system is currently being used as an instructional support tool in graduate level courses in software engineering and expert systems. Research is in progress to extend the expert system's usefulness to non-academic settings. Incremental extensions of this "proof of concept" model are discussed to provide a roadmap for more comprehensive work in automating the software development process.
展开▼
