Selecting candidate component best matching specification from a component library

Abstract

Since the emergence of software engineering, researchers have struggled to develop techniques that would solve or at least alleviate the software crisis problem. One of the promising solutions is software reuse in general and component based software engineering (CBSE) which relies on mass produced components in particular. Realizing CBSE requires its own set of new technologies of which adaptation (modification) is probably one of the most urgent areas that need attention. In this thesis, based on works on estimating cost of software modification, we tried to propose a framework that would enable developers choose a component from a library of component for their query that requires the least adaptation effort. We improved on a work done earlier on a similar topic focusing on functions and extended it to suit the needs of component matching. We will define what a component specification is formally and use it as a basis for the information that we use during the development of the framework. The work shows how the information contained in specification (which purely describes the behavior independent of implementation) can be used to compare components to see which one is easier to modify.