Cyperphysical systems (CPS) are on the horizon to challenge the way we see and develop computer systems. CPS provide the integration of embedded computing, real-time computing, distributed computing, parallel computing, mobile computing, and server systems, resulting in completely new application domains. On one side the arise of CPS is just the natural consequence of various incremental improvements of enabling technologies. But from the software engineering side it creates big challenges for software engineering. CPS will result in significant rise of system complexities. At the same time there are serious security and safety challenges as it requires to combine subsystems with closed-world assumptions and those with open-world assumptions, requiring new interfacing solutions. In this talk we present the research towards the development of a programming model that is aimed to withstand the CPS challenge. This programming model will provide support for mixed-criticality systems and adaptability. As a fundamental change in software engineering we present a computing model that unifies the non-real time, soft-real time and hard real-time paradigms. An important mechanism included in this computing model are tolerance ranges, which allow to adapt the system when necessary. Such adaptation scenarios include fault-tolerance with reconfiguration and optimisation of resources like processing time or power consumption.