An approach to measure understandability of extended UML based on metamodel

Abstract

Since UML does not provide any guidance for users to select a proper extension pattern, users are not able to assure the quality of extended UMLs, such as understandability, when they focus on their expression power. A metric of understandability for extended UMLs is proposed, which bases on measuring the deviation of understandability between the extended UMLs and the standard UML in their metamodel level. Our proposal can be used to compare different extended UMLs with the same expression power on the understandability characteristic. Moreover, the proposal can guide users to select an appropriate extension pattern to achieve their goal. We give the definition of the metric of understandability and the empirical validation of the proposed metric. A case from a real project is used to explain the application of the proposed metric. © 2012 Springer-Verlag.