Analysis and System Administration
Networks and computer systems are objects to be observed and
understood by a process of scientific inquiry. Even in isolation,
computers are too complicated to predict from purely deterministic
principles of logic and reason; then, once we place them in a
non-deterministic environment of users and networks, the task becomes
impossible. The result is that network or system manager is in the
position of the natural scientist observing unknown phenomena and
learning about behaviour that he or she strives to explain.
Computer Science looks traditionally to logic and mathematics rather
than to natural philosophy to describe computers, examining the
consistency of assumptions rather than the reality of their
implementation. Clearly such studies are valuable, but they do not
represent the practice of computer systems. Scientists have
long since abandoned such fundamental approaches to studying the real
world. Trying to describe the behaviour of the human body in terms of
atoms and their basic interactions would be an intractable problem.
Whether our goal is to reason about the world, or to measure it, we
need scales of measurement. In terms of computer science engineering
dimensions can be thought of essentially as a type theory of
measurement.
System administration is about the planning, deployment and
maintenance of human-computer systems. The human element of the system
here, since it leads to significant uncertainty. Humans make
approximation essential.
|
