A Tool for Reducing Implementation Defects for Requirements based on Large Decision Tables Towards a Semantic Gap Reduction

The growth and complexity of Information Systems demands tools and methodologies aiming to reduce the time for development as well as to improve the quality of such systems. The use of business rules to model the behavior of information systems is an adopted technique in software engineering. Business rules are described using multiple techniques such as Decision Tables. A Decision Table is a tabular representation combining conditions and responses. Empirical industrial experience showed that the existence of defects associated to the bad understanding of large decision tables representing business rules. This paper describes the first set of results in a project that aims to reduce the defects generated during the implementation of business rules based on the hypothesis that the automatic generation of code for decision tables can reduce the semantic gap between what is described and what is implemented. The results include the description of a fully functional prototype that allows the definition of large decision tables as well as the code generation. Additionally, the paper also describes the experiments performed so far to test our hypothesis for defect reduction together with some new paths for research.