Context : The Unified Modeling Language (UML) has been a foundational component of software engineering education for decades. However, the increasing adoption of agile methodologies and the emphasis on lean documentation have raised questions about the practical relevance and usage of UML in industry. Understanding the gap between academic instruction and professional practice is crucial for improving modeling education and supporting the development of meaningful software design. Goal : This study aims to investigate how UML is currently taught and learned in academic settings and adopted in professional environments, comparing perspectives from Brazil and Portugal. The objective is to identify alignment or misalignment between education and industry needs and provide actionable insights to improve modeling practices. Method : We conducted two structured surveys: one with 80 instructors from higher education institutions and another with 206 software practitioners. Unlike previous studies focused on a single context, this research adopts a cross-country mixed-methods design that integrates quantitative and qualitative analyses to enable a comprehensive comparison of UML education and practice. The methodological triangulation combines descriptive statistics with Grounded Theory coding of open-ended responses, providing both breadth and depth of interpretation. Results : The findings show that UML is primarily taught using a subset of diagrams, namely class, use case, and sequence diagrams, within traditional Software Engineering courses. In contrast, practitioners report limited use of UML in real projects, often constrained by agile workflows, lack of tool support, or perceived complexity. While instructors emphasize the value of UML for abstraction and documentation, professionals tend to favor practical communication tools and simplified modeling practices. Conclusions : There is a significant gap between how UML is taught and learned and how it is used in practice. This study advances the understanding of software engineering education by offering the first cross-country empirical comparison of UML teaching and practice between Brazil and Portugal. The results highlight how contextual factors, such as agile adoption, tool accessibility, and institutional curricula, influence the persistence of this gap. To bridge this divide, modeling education should incorporate agile-compatible strategies, emphasize the practical use of tools, and focus on diagram types most relevant to the industry. These findings provide important implications for modern software engineering education, supporting the design of curricula that better align theoretical instruction with professional practice in agile-driven environments.
OliveiraJr et al. (Thu,) studied this question.