|
Rand's basic insightthat knowledge does not consist of isolated facts but is organized into a hierarchy based on classes of entitiesanticipates recent developments in information technology. For instance, the "expert systems" developed by artificial-intelligence researchers incorporate data models that (aside from some technical details) are structurally analogous to human conceptual knowledge. In such a system, the data model is combined with a set of rules and an analysis engine. In the hierarchically organized data model, traits are "inherited" from "parent" objects by "child" objects, just as attributes are inherited by human sub-concepts. For the expert system to run effectively, the data model must be built logically. If "child" objects are grouped under "parents" in a haphazard manner, then the system will not run efficiently and may even reach wrong conclusions from the given rules. Similarly, human thought processes, which expert systems are intended to model, may break down if they depend upon poorly formed pseudo-concepts. Indeed, the relative slowness of human thought processes renders them even more susceptible to failure under such conditions.
Such object-oriented methods of organizing information systems are by no means limited to AI research, but are becoming increasingly accepted throughout the field of information technology.