A critical look at the relationship between AI and software engineering

A number of key issues relating to the application of artificial intelligence (AI) to software engineering (SE) and the reverse are discussed. Issues relating to how AI can help SE are: whether automatic programming is synonymous with AI for SE and whether it represents a major paradigm for SE in the next decade; whether expert systems technology is sufficiently successful and mature enough to provide significant solutions to certain aspects of the SE process; whether AI development and maintenance environments are suitable for direct application to the SE process; how AI methodology can be applied to the software design process; and how the AI rapid prototyping paradigm is useful as a SE paradigm. Issues relating to the common belief that SE is not useful for AI are: SE is synonymous with the waterfall model, and the waterfall model is linear and hence not suitable for AI; expert systems can not be specified and thus SE techniques do not apply; and AI software is easy to maintain and thus SE is not required. An attempt is made to relate AI and Se at the level of objectives and issues