Standardizing production of domain components

Abstract

m Domain analysis is a promising path for identifying standard software architectures. Recent advances in the technology and documentation of objectoriented frameworks have made it possible to link the products of domain analysis to concrete software artifacts. The result is a coherent lifecycle process for the domain engineering of reusable components. or nearly two decades, software engineers attempted to create a software component industry based on a model of a repository of “components” or “parts” that could be accessed by many different kinds of (possibly totally unrelated) applications. It took several years of spectacular failures to make it clear that this approach could not succeed. It became increasingly clear that components could only be produced in the context of a domain. Such a domain might be telecommunications, banking, etc. Most current efforts are aimed at designing what are known as domain architectures—that is, the standard architectures of systems created in particular domains. In the computer hardware area, this has been accomplished with great success: A personal computer has a motherboard, expander slots, keyboard, monitor, etc., conformant to a standard architecture. But in the software area, much less is known. With the identification of a domain architecture, it becomes possible to develop systematically reusable domain components that fit within that domain architecture (via suitable interconnection mechanisms). The discipline that has arisen around standardizing production of domain components is known as domain analysis. The companion discipline of domain engineering—the systematic creation of domain architectures based upon the results of domain analysis—has flourished in recent years with the rise of object-oriented framework technologies and patterns.