A method of mapping sets of uninterpreted record or relation schemes into entity-relationship diagrams is described and then applied to two conceptual design problems. First, the method is applied to the design of relational databases. It is shown that the method can be interpreted as a normalization procedure that maps a given relational schema into a new schema that represents an entity-relationship diagram. That is, the original schema has an interpretation in terms of higher-order concepts, which helps in understanding the semantics of the database it describes. The second design problem is related to the conversion of conventional file systems to the database approach. The method is used in this context to obtain a database conceptual schema from the description of the conventional system, which is one of the fundamental steps of the conversion process.