Finding a way in the Model Driven jungle: Invited Keynote Talk

Model-driven concepts have been introduced in software engineering methodologies since many years. They were not really new, as engineers have always been using models, but they have been tailored to engineering software. Research in the area has progressed in many directions: languages, processes, standards, technologies, tools. While they have proved to be effective in some application sectors, such as for embedded systems, it is hard to find documented success stories for real-world systems in many other fields. Indeed, there is some skepticism on their applicability for large-scale and for critical industrial systems, which have high complexity and/or high costs of verification and validation: many companies still consider them risky. Full comprehension of risks, costs and benefits is not easy to achieve. A crucial factor is that their adoption requires changes in consolidated processes, and advanced engineering skills âĂş focus is on modeling, rather than on implementation. These implications are often underestimated.