> there appears to be considerable time worrying about the technicalities of the scheduler […] whether this really implies linear time bound. IMHO these are not

> there appears to be considerable time worrying about the technicalities of the scheduler […] whether this really implies linear time bound. IMHO these are not worth worrying about But it is not linear time, we shouldn’t let the existence of other algorithms make us confused about proper analysis of this one. The call to `sleep(n)` is not a trivial operation, it inserts our task into a priority queue, so we are piggybacking on the OS’ support for sorting tasks based on when they need to be woken up, and thus hide the actual complexity behind the call to `sleep` (which we call `n` times, so if it’s `lg(n)` then the time complexity of the entire algorithm is `O(n lg n)`).