The paper contributes to the elaboration of a methodological framework for solving the problem of data model heterogeneity in the multidatabase systems.
It presents an approach to rigorous definition of data models and to the handling of them as formal objects provided for the development of data model mappings in the process of design of heterogeneous multidatabase systems. The paper defines strict principles and specific techniques facilitating verifiable design of commutative data model mappings. The methods introduced make it possible to construct generalized data model kernel extensions equivalent to various internal data models and to develop and verify algorithms of the generalized level DML interpreters.
|
