hard vs. soft real time is a matter of latency tolerance. human-facing systems (like telecom) need to be low latency from a human perspective, which is a higher

hard vs. soft real time is a matter of latency tolerance. human-facing systems (like telecom) need to be low latency from a human perspective, which is a higher threshold than what you could tolerate in a mechanical context where 100ms could mean damaging some expensive hardware.