User-level threads are threads that are scheduled directly by the process in which they live rather than by the kernel - the kernel schedules the processes, and any process that has threads can schedule them however it wants within its time slice. Describe briefly how such a scheme would work. Can you think of any problems with such an implementation?