Schema Transformations ?

We develop a formal basis of correct schema transformations. Schemas are formalized as abstract data types, and correct schema transformations are formalized as information-preserving signature interpretations. Our formalism captures transformations of all schema components , making it possible to transform uniformly constraints and queries along with structures. In addition, our formalism captures schema transformations between diierent data models as easily as those within the same data model. Compared with Hull's notion of relative information capacity, our notion of information preservation captures more schema transformations that are natural, and fewer schema transformations that are unnatural. Our work lays the foundation of a transformational framework of schema manipulations.