Analysis of requirements incompleteness using metamodel specification

Abstract

Incompleteness of requirements has been treated as a huge challenge in software development projects. Since it is hard to obtain all required information before software design and implementation starts, the software modeling process may start with an incomplete requirements specification. In order to help analyzing the incompleteness in requirements, I propose a metamodel approach for detecting the missing requirements that are needed for constructing conceptual models for a software system, and implement it in MetaEdit+. The detected missing information in a conceptual model is reported in natural language, which is easy to understand. Furthermore, the conceptual modelers can identify the potential problems indicated by the report to analyze and update the model. The contribution of my thesis is twofold, i.e. analyzing the link between business rules and the ER models, and implementing a method to automatically detect and show the incompleteness in ER models.