Research And Implementation Of Mapping Analysis Of Business Model And Source Code

Organizations dealing with legacy systems often follow various strategies: integrate overlapping custom applications, migrate from a custom application to packaged application, and keep the custom application, migrate its architecture from a monolithic architecture to a loosely coupled, agile architecture. These legacy transformation projects all follow a common sequence of high-level activities:Users of the current system identify an abstract target model or architecture, which captures all strategic requirements from the system; The developers in conjunction with users validate or reconcile the target model with the existing systems; and The validated target model is used to construct (or reconstruct) the new system using the selected strategy. However, the validation is full of challenges. This thesis address the task of mapping a given textual domain model (e.g., an industry-standard reference model) for a given domain(e.g., ERP),with the source code of a given application in the same domain, in order to turn the validation to a more efficiently and automatically process. This thesis analyzes the advantages and disadvantages of two different models in information retrieval and uses the vector-space model to abstractly represent business model elements as well as source-code artifacts, and reduce dimension using latent semantic indexing, and finally evaluates the degree of relation between the business model and source-code. The key novelty in our approach is to use the relationships between source-code artifacts in a principled way and business dictionary to improve the mapping process.The thesis illustrates the theory in mapping analysis, proposes improving approaches, designs and implements the architecture of mapping business model with source-code, and address a set of experiments to evaluate the effectiveness of this architecture.