Does this mean that state transition depends on output of previous state ONLY, and that output is produced only once (unicast, rather than broadcast)? Thus, wh

Does this mean that state transition depends on output of previous state ONLY, and that output is produced only once (unicast, rather than broadcast)? Thus, when you measure previous state, you effectively deprive the "next state computation" on those inputs, and hence, block state transition?