A generative approach for building data-oriented information systems

Abstract

In recent years the development of object frameworks has become state of the practice for building data-oriented information systems. This approach has shown some success in improving productivity in building such systems. Using code generators has become a somewhat forgotten technology. But it is just the domain of data-oriented information systems where most of the processing consists of associative data access, that offers great potential for the use of generators. The paper describes a generative approach for building data-oriented information systems and compares it to the pure object framework approach. It shows why a pure object framework approach fails to achieve a sufficient amount of productivity gain in this domain. The approach takes advantage of patterns which occur in the implementation of data-oriented information systems. The common object framework approach is extended by the massive use of generators which apply these patterns on a model of the application.