Content-intensive, interactive information systems have become an integral component of computer infrastructure. Applications become increasingly complex. Meanwhile, the information system technology has evolved to a very large extent. Methodologies for developing large information systems have not evolved in the same speed. We also observe an impedance mismatch between software engineering and database development methodologies.
The complexity of user interaction requires anticipating system behavior on both application and database sides. Varying utilization scenarios require adaptable functionality. Data model modifications pose another challenge, as necessary application code changes often exceed the capabilities of common software parameterization techniques. A generic architecture for such systems has been missing so far.
This work focuses on a method for specifying software functionality that has been developed. It is based on computational linguistics and allows to define generic specifications of system functionality and to provide them to users under varying application contexts, constraints and data models. A formal notion of genericity is introduced for comparing information system behavior.
The method supports stepwise transformation of specifications, starting from a specification close to natural language, and finishing with an executable prototype. The transformation is interactive, supporting design decisions according to validation results at each refinement stage. The proof of concept is given by a prototype generation framework.
