Validating Metrics for OCL Expressions Expressed within UML/OCL models

Measuring quality is the key to developing high-quality software, and it is widely acknowledged that quality assurance of software products must be guaranteed from the early stages of development, assessing through metrics the quality of early models such as UML diagrams. There exists several proposals of metrics to UML diagrams, such as class diagrams, use case diagrams, etc. But, even though the incorporation of OCL to UML diagrams improves software quality and software correctness, there are no metrics for OCL expressions. In a previous work we have defined and theoretically validated a set of metrics that can be applied to OCL expressions expressed within UML/OCL combined models. The main goal of this paper is to show how we carried out a controlled experiment to ascertain the empirical validity of the proposed metrics as early indicators of OCL expressions understandability and modifiability.