Object use-cases clustering using MDL

Today the software researchers and the industry specialists are paying more attention towards the component based software development. It is a powerful tool to enhance the productivity and accelerate the development time. There are different areas for which the components can be developed. This is well formulated in the 3C model [13]. There are two concepts in the field of software reusable process. One is development with reuse and other is development for reuse [1]. Development with existing reusable software components is known as the development with reuse. One should follow a software life cycle to develop the software. The artifacts resulting from the software life cycle phases can be reused. Our work deals with the extraction of software components from the software design phase. Many researchers have proposed the extraction of software components from the software design. Some of them are based on Unified Modeling Language (UML). These techniques focused on the various models and their corresponding inter and intra relationships. Our method is also based on UML but have a different perspective. It describes various models and their relationships. The information about the models and their relationships has been extracted not from the model directly but from their corresponding petal/mdl (model description language) files. Petal/mdl is an unstructured representation of the information contained in the various UML diagrams like use case, class, sequence etc in textual format. Extracted information is mapped into database tables to form clusters. The process of clustering is to facilitate the retrieval of components so as to gain high precision and recall.