An empirical investigation of software engineers´ ability to classify legal cross-references

Requirements engineers often have to develop software for regulated domains. These regulations often contain cross-references to other laws. Cross-references can introduce exceptions or definitions, constrain existing requirements, or even conflict with other compliance requirements. To develop compliant software, requirements engineers must understand the impact these cross-references have on their software. In this paper, we present an empirical study in which we measure the ability of software practitioners to classify cross-references using our previously developed cross-reference taxonomy. We discover that software practitioners are not well equipped to understand the impact of cross-references on their software.