Asymmetry in Identification of Multiplicity Errors in Conceptual Models of Business Processes

Abstract

ABSTRACT Business rules can be represented by multiplicities in a Unified Modeling Language (UML) class diagram. Diagrams containing erroneous multiplicities may be implemented as an inefficient/ineffective database. System validators must be able to validate such diagrams, including multiplicities, to prevent the implementation of design errors. Prior research reveals conflicting evidence regarding the expected accuracy in validating minimum multiplicities, indicating a need for additional research to further our understanding. Ontology research claims that multiplicities that depict optional participation are ambiguous and lead to poorer understanding and accuracy compared to multiplicities that depict mandatory participation. However, other research has reported better accuracy validating multiplicities that depict optional participation compared to mandatory participation. We conducted an experiment to help resolve this apparent contradiction, and to explore whether any asymmetry exists in accuracy for maximum multiplicity validation. Results indicate an asymmetry for validation of minimum multiplicities such that accuracy is greatest when the underlying semantics represent mandatory participation. Results also indicate an asymmetry for validation of maximum multiplicities such that accuracy is greatest when the underlying semantics represent flexible participation. Given that many business relationships call for optional minimum participation and that many business relationships call for restrictive maximum participation, these error identification asymmetries are cause for concern.