Scaling Model‑Based Formal Methods: A Thirty‑Year Journey from Software to Digital Twins

This retrospective paper reviews more than three decades of our work on the development and industrial deployment of model‑based formal methods in software and systems engineering. We trace the evolution of our research from early work on the Vienna Development Method and its tooling, through successive attempts to scale rigorous modelling approaches to Cyber‑Physical Systems, Systems‑of‑Systems, and, more recently, Digital Twins. Reflecting across these phases, we identify four recurring principles that have shaped our work and our understanding of the practical application of formal methods: balancing modelling effort against engineering value, embracing multiple modelling paradigms, providing complete technology packages, and recognising the limits of first‑principles modelling. We discuss how these principles emerged from successes and limitations encountered throughout our own research and industrial engagements, and consider their continuing relevance as systems become increasingly heterogeneous, interconnected, uncertain, and interdisciplinary. We conclude by identifying opportunities for formal approaches in the dependable engineering of digital twins.