A conceptual justification for model transformations

We analyze inter-model transformations of database schemas from a conceptual point of view. A central question for us is not just whether the information capacity of the transformed schema is suffcient, but rather its suitability for a given task. For this, we require criteria beyond the resulting degree of normalisation which we subsume under the term "conceptual justification". To illustrate our point, we take a closer look at a class of conceptual requirements that frequently cause practitioners to manually denormalise the logical schema: layered schemas, where the natural layering of the data clashes with the dominant access patterns and negatively impacts performance. We show how such requirements can directly in uence the transformation process and give rise to conceptually jused logical schemas. We include an example which is based on the translation from the higher-order entity-relationship model to the relational model.