There are many uncertainties in the Bern RTOS architecture. The next step will be to start the development of a simple scheduler, i.e. no time-slicing, no MPU, just a static table of tasks. Once the minimal prototype works, more functionality will be gradually added, always validating and adapting the Bern RTOS concept.
The biggest uncertainty is whether and how the MPU can be used to restrict tasks access to stack and heap only. And if so, we need to decide where to draw the line between kernel and thread mode. A small kernel leads to a more stable system, but has more overhead from the many syscalls. In the next project these issues have to tackled as early on as possible by proof of concept implementations.
In the best case the kernel with memory protection and basic IPC is implemented at the end of the next project. But emphasis is on proving and adapting the concept of memory protection and resource permission introduced in this project.