A well known feature of descriptions of complex systems is that a complete understanding is best organized as a unification of the partial understanding of the system at several different levels or scales. Complex systems are often so disparate at different scales that quite different descriptions are required to capture the full essence. A theory of system behaviour at, say the microscopic level of system calls, need not resemble a theory for the behaviour at a macroscopic scale of larger entities, such as patterns of user behaviour. Both are needed in order to understand the whole hierarchy of things going on.
If one is only interested in high level phenomena, then the details of low level phenomena are seldom directly relevant, to a good approximation. This is the principle of separation of scales. The principle states that, as one moves from microscopic to macroscopic scales, new behaviour can emerge as collective phenomena, which often depends only weakly on the microscopic details of the levels below. This is a simple idea, which is quite intuitive, but which has far reaching consequences. It can easily be appreciated with the help of a couple of examples.
A bridge, for instance, has the property of spanning a distance and carrying weight, regardless of whether it is made from steel or copper or wood. The choice of material and the microscopic arrangement of atoms in the metal or wood, of course, tells us something about the strength of the bridge, but perhaps not as much as the structure of the bridge at the scale of the whole thing. In other words, the construction of the bridge at the scale of the users of the bridge is far more important to its function than the microscopic construction of its pieces under a microscope.
Similarly, to an acceptable approximation, the behaviour and operation of a sales database, at the level of information transactions (their order and type), is more important to an information retrieval system than how those transactions are implemented through system calls (e.g. whether the system runs on Unix or on NT). The ability to retrieve information does not depend on whether the storage medium is, an IDE or a SCSI disk. The same job will be done regardless.
To summarize, a description of system behaviour at a high level is, for many purposes, independent of specific details of the lower levels. Computer systems can be modelled by generic computer systems with certain high level characteristics; similarly users can be modelled as idealized users, also with common characteristics. A theory of system administration will be most successful if it appeals to such generalities, rather than delving into unnecessary specifics.