Metamodel-based comparison of data models

A data model specifies the building blocks of databases, the rules how to assemble these blocks and operations that can be performed based on the built-up structures. We have to create increasingly complex applications. Properties of the underlying data model of a Database System (DBMS) determine how easy is to create an application that uses a database. There are many different data models. We have to choose a DBMS, the underlying data model of which best fulfils the needs of an application. Existing comparisons of data models are mostly based on the experiences of using one or another DBMS. This paper explains bow to perform non-empirical comparison of data models by using the metamodels, which describe abstract syntax of these data models. We also present some results of the comparison of the underlying data model of SQL:2003 and the data model that is proposed in The Third Manifesto.