- trim the performance, security and redundancy of a computer system
- ..but the key is to be able to precisely and reliably identify a process
Justifying our interests
We could actually end up going unnecessarily deep. Let's think about this. An operating system is made of 2 parts, roughly: the kernel / and various other parts.
Kernel is conceptually made of some parts, one of which is schedulers. The act of "scheduling" means putting the "competing" processes to be executed on the core, one at a time, in a specific way that essentially makes an illusion of multitasking. Scheduling also regards basically other things than only the CPU time, but we'll justify this simplification for now.
Scheduling is a mathematical treatise whose purpose is to create beautiful statistics, to put it another way. It is one of the more - if not most - important function of a kernel. [Read on: Reliable Linux process identification, pt 2/3: Ideas]