From Textual Descriptions to Code: A Filtering Approach for Locating Business Rules

Business rules form the backbone of enterprise applications, capturing organizational policies, legal constraints, and the decision logic that governs business processes. In legacy systems, these rules are often hidden in poorly documented code, scattered across multiple modules, and entangled with technical details. When changes are needed, developers must locate and update the relevant code fragments, a process that is time-consuming and error-prone. In this paper, we propose a fully automated approach designed to assist developers in selecting the portions of code that need to be modified following a change in a business rule. The approach reduces the search space of candidate methods likely to implement the affected rule. Starting from a textual description of the rule, our approach combines natural language processing with code analysis to filter methods and retain those relevant for implementing the necessary modifications. We evaluate our solution on a real-world codebase and demonstrate its usefulness in guiding developers during software evolution tasks.