On Preserving Information in Schema Transformations: A Constructive Perspective

When a knowledge base (called here schema) is restructured into another schema via some mappings, it is expected that the transformed one contains at least the same information as the original. This characteristic of the transformation is commonly referred to as information capacity preservation or schema dominance. In this paper, we complement the existing definitions of information capacity by offering a formal and constructive perspective to test schema dominance and schema equivalence, in the restricted case of first-order schemas related by first-order mappings. Satisfying some sort of schema equivalence property is essential in areas such as database conceptual design and database reverse engineering. We review, through the lens of our approach, the current literature concerning database normal forms and decompositions, and the process of reverse engineering a database schema.