The Resource‐Event‐Agent (REA) enterprise model is a widely accepted framework for the design of the accountability infrastructure of enterprise information systems. Policy‐level specifications define constraints and guidelines under which an enterprise operates, and they are an extension to the REA enterprise model, adding the “what should, could, or must be” to the “what is.” This paper aims both at comprehensive understanding of policy‐level definitions as part of REA enterprise systems and at understanding of the semantic constructs that enable such definitions. We first explore two distinctive semantic abstractions essential to policy‐level specifications: typification and grouping. The typification abstraction links instances of an object class to concepts for which they are concrete realizations, while the grouping abstraction aggregates objects into collections. We next present a number of patterns for the semantic modeling of policies. Following, we look at policy‐level applications for REA enterprise information systems. We explore type and grouping definitions for the REA primitives (resource, event, agent) and discuss enterprise applications for three different kinds of policy definitions: knowledge‐intensive descriptions, validation rules, and target descriptions. Our discussion of specific enterprise applications includes internal control applications (e.g., limit checks), variance analysis based on standard specifi‐cations (e.g., bills of materials), and budgeting applications.

This content is only available via PDF.
You do not currently have access to this content.