Knowledge based invocation of software methods and tools

Software engineering is in transition from a purely manual approach to an automated procedure, where computer-aided handling of models, methods, and tools is required. The divide-and-conquer principle is applied to theory-free areas such as method description, tool selection, quality and productivity modeling, and code instrumentation to allow production rules to capture the knowledge about the particular problem and to define solutions. Rules are also used for the interrelation of problem domains in order to define the implementation of a problem solution by the same mechanism. The plain rules are enhanced by weight factors and parameters in order to define the range of the rules and to allow ranking of alternatives. Objectives and goals are also represented by rules which are mapped onto problem rules by a second set of interrelation rules. A basic AI (artificial intelligence) concept, production rules, and an elementary tool, a rule interpreter, are used recursively to solve selected software engineering problems and the AI problem of how to organize the particular problem solutions.<
>