This question is open-ended; there are many possible answers. You don't have to type very much, and you can think about these questions even when you are not in front of a computer.

There are many metrics that a scheduler has to balance: turnaround time, response time, throughput, and various definitions of fairness. We will consider the following types of systems in this question:

• A high-performance computer used for large number crunching tasks (genetic sequencing, graphics rendering...)
• A multimedia computer (video games, movies...)
• A smartphone
• A general-purpose computer on the CIMS network

1. For each of the system types above, which metrics do you think are the most important, and why? Some things to consider: How many users does the system have? Are user processes typically interactive? How long is a user willing to wait? What frustrations will users have if certain metrics are ignored? (There are many possible answers to this question.)
2. Taking into consideration your answers above, propose a scheduling policy (FIFO, etc.) for each system above.
3. For each of the systems above, what is the scheduling policy that would most completely undermine the purpose of the system?