Component Testing Information Research And Retrieval

With the emergence of component libraries and continuous development of component testing technology, component testing tools are fast emerging. A large amount of information produced in the component testing process is directly relevant to testing information. Such kind of information is valuable to the component provider, people who reuse the component and a third party. These users all hope to query and analyze component testing information in different ways to meet their needs. Therefore, in order to meet those users component testing information retrieval needs, reduce the cost of retrieval and understanding of the component testing information, better support to component reuse, and eventually facilitate component-based software development technology, how to completely gather testing information, reasonably classify and represent, as well as effectively store it, the need for an efficient retrieval mechanism seems particularly important.To meet the above requirement, this paper analyzes current development of component testing, investigates its process and characteristics; collects, analyzes and organizes information relating to it; conducts in-depth research on various information classification representation methods. Base on component testing's characteristics, this paper proposes a component testing information representation pattern which combines four classical classification:faceted classification, enumerated classification, keyword classification and attribute-value classification. Besides, this paper integrates traditional database technology with modern information retrieval technology, as well as combines the characteristics of component testing information with XML structure to make some improvements to vector space model in order to implement a natural language-based,a certain degree of semantic-level supported as well as a practical and powerful component testing information retrieval mechanism. Finally, this paper, based on three evaluation indices---recall rate, precision rate and overall query rate, uses Microsoft Visual Studio.NET 2005 as a development tool to conduct relevant experiments on the proposed component testing information representation pattern and retrieval mechanism under some application scenarios.