A large group of database design methodologies relies solely on graphic representations of the conceptual schema. When this conceptual schema is translated manually to the DDL of a target DBMS errors are frequently introduced. The database design tool presented in this paper consists of a conceptual model definition language with its compiler, a data dictionary in which the metamodel of the database is stored and a mapping interface. The tool is based on the ELKA (Entity, Link, Key, Attribute) conceptual model, a model which normally relies on a graphic notation. In this tool ELKA structures are defined by the Database Administrator through the conceptual model definition language from where they are mapped automatically into database structures. The metamodel is accessible to the user through the DBMS's data manipulation language, while the equivalent database structures are accessible both to the user and the application programs. Finally, the modeling power of this tool is explored and necessary extensions are identified that are needed to handle the notion of molecular aggregation.