Conceptual and Constraint Similarity in Component Matching

Abstract

Component reuse improves both the quality and productivity of software development. Before Software components can be reused they must be located from component repository. Reuse repository system provides an effective means to locate software components which can be reused to solve the problem in hand. As a result programmer need not to always begin from the scratch. This not only leads to good quality software but also reduces the development and testing time and effort. This thesis has resulted in a fully functional reuse repository system with effective and easy to use search functionality so that needed reusable components can be easily extracted from repository. The constructed system has an easy to use interface, due to Web-browser based front-end. Concept similarity and Constraint compatibility has been used for searching the relevant components. For Concept similarity, an Information Retrieval Vector Space Model with stemming algorithm, stop word removal and thesauruses has been implemented. Constraint compatibility is computed by signature matching. To select the components from repository and compute the relevancy of components, system also considers the user experiences. For this purpose, system allows the user to assign the weight to components so that this knowledge can be used for effectively processing future queries. So more the system interact with user, more experience it will gain and as a result it makes the system more intelligent and increases its capability to return more useful and relevant reusable components.