Schema Evolution publication categorizer

In this paper we study the problem of translating schemes between
different data models,
in a formal framework that refers to a wide range of models. We first
introduce a graphtheoretic
formalism that allows us to uniformly represent schemes and models,
to compare
different data models and to describe the behavior of basic
translations. The formalism
is based on a classification of the constructs used in the known data
model into a limited
set of types. Then, we study in this framework formal properties of
scheme translation
between heterogeneous data models, and we develop a method for
deriving translations
that enjoy those properties.
Keywords.