... and I will have to steal your additional and concise points from the software development perspective ;) Thinking about it, maybe to ensure some latency aw

... and I will have to steal your additional and concise points from the software development perspective ;) Thinking about it, maybe to ensure some latency awareness we should adapt the "timeout-exception" concept in areas it is not used today. Given some runtime-lib support call sites in a project could be given a timing threshold to complete their work... if the threshold is exceeded an exception could break the code-flow in a specially prepared build of the project. In very modern dynamic environments ;) such thing could be implemented by transparently wrapping the call sites (via attributes, annotations, monads etc), in more traditional env. the compiler could generate specially adapted code measuring the latency using the high precision timers most CPUs provide...