The field calculus is proposed as a foundational model for collective adaptive systems, capturing in a tiny language essential aspects of distributed interaction, restriction and evolution, as well as providing ground for engineering resiliency properties. In this paper, we investigate the inter- play between interaction and restriction: known as “domain alignment” in field calculus, it is extremely powerful but can cause subtle bugs when not handled properly. We propose a disciplined programming approach based on the interplay between a weak and a strong version of alignment, mixing static and dynamic checks. This is exemplified to design a new reusable component dynamically updating the strategy by which a device can extract information from neighbours, which find applications, for instance, in the on-the-fly evolution of metrics in smart mobility applications.