Extracting procedural knowledge from software systems using inductive learning in the PM system

The issue of software reuse has been found to be a much harder task than previously thought. Some of the problems are due to the lack of emphasis placed on non-functional requirements during the software development phase, such as maintainability and understandability. Other problems arise from the difficulty of defining precise criteria for considering a software module reusable. They are usually elusive, and vary dramatically from one domain to another. This paper presents PM, a software system the goal of which is the automation of the software reuse process. PM uses an incremental approach in performing analysis and storage of software modules, at different levels of granularity. Its fundamental characteristics are domain independence and flexibility, accomplished applying inductive learning techniques and analyzing reusable and nonreusable code examples.<
>