Since it is cooperative, a developer can directly implement efficiency by splitting their application up into units of work, yield ing to the dispatcher frequently
— I will also add HRRN which apparently also gives priority to procs that have to wait a long time
— Sounds strange
— I'd just round robin it
— Round robin is pre-emptive, while the kernel is cooperative
— Any process can easily take 100% of all processing power indefinitely, crashing it
— Well, isn't that the case anyway?
— Just happens a tiny bit later using SJF?