Validating System Requirements: A Conceptual Approach
Incomplete, ambiguous and inaccurate system specifications have been cited as a significant determinant of IT project failures. Although a consensus on the causes of the inadequacies in specifications has yet to be reached, contemporary data modelling techniques often lack the expressiveness to more fully define the structure and the behaviour of organisations. In addition, many of these modelling techniques incorporate implementation details unfamiliar to business people. But unless domain experts can actively challenge the perceptions of analysts attempting to capture the organisation’s business rules in these models, how can analysts be certain that they are defining the required system? This paper investigates Object Role Modelling (ORM) as a possible language for expressing and validating business rules in collaboration with domain experts. The Business Rules Group (BRG) have produced a comprehensive taxonomy of business rules enabling analysts to determine accurately, what is and what is not a business rule. A case study conducted by the author has indicated that synthesising the concepts and definitions within this taxonomy with those of ORM, fosters a rigorous and yet transparent approach for expressing the data requirements of information systems. Even complex constraints on data structures such as set comparisons can be modelled in ORM at a conceptual level, facilitating a more complete description of the problem domain than is generally the case. Since business rules can be verbalised automatically in ORM’s constrained natural language, data requirements can also be actively validated by business people. By adopting ORM as the language of business rules, analysts may possess a powerful tool for improving the quality of data requirements that may in turn help to reduce IT project failures.
Keywords: Business Rules, Conceptual Data Modelling, Requirements Definition
Lecturer, Department of Information Systems, Massey University