Design for automated enterprise data model

Abstract

In that many traditional CASE tools have been nothing but drawing tools for ERD with only subjective view of data modeler, users who are not data modeler have not been able to generate ERD on their business requirements and data modelers have only drawn ERD without soundness and completeness of design. Although some CASE tools supported conceptual data design, they required for users too much preliminary knowledge to learn how to use and handle them. In addition, in a number of studies on conceptual data design with natural language, they claimed passive participation for users with conforming to messages predefined by CASE tools. As an alternative to these traditional CASE tools, we proposed an ERD formulator for automated data design tool, called ERDF, so that even ordinary users, not necessarily data modeler, are capable of formulating ERD on business requirements by use of ERDF. We, first of all, introduced NSM as the standard methodology. We also designed the structure of ERDF including main controller, input controller, operation controller, regulation controller, schema controller, and output controller. We then defined conceptual domains and basic operations to lay down schema operations as well as sentence rules to handle input sentence in natural language. To be faithful to requirement formulation, we laid out supplement design for dialogue and confirmation of soundness and completion.