Modeling multiple domains in software reuse

Abstract

Modeling multiple domains is use@l in two ways. Firstly, most of real sof~arw ~sterns are based on more than one domain. In this case, modeiing dqoenderrcies between &mains facilitates better understanding and clearer specljlcation of mferwtce mquiwments for a jiiily of systems. This leaak to a structured and jlexible generic architecture for that family and, therejorw, improves reuse. Secondly, our method encourages one to divide big a%mains into smaller, easier to unabstand and describe subdomairrs. In big domains, w found this an e~ective ww to identlfi reuse opportunities that may not be easi~ observed otherwise. The method described in the paper builds upon wellkrrown concepts of enterprise information architecture, domain analysis and the Domain-Spect~c Sof~are Architecttm (DSM) approach to reuse. We argue tha~ modeling of relationships between the domains is bene>cial 0s it ensures urri$orrnity of sofiware systems across abmains, wduces the cost of sofmare development through inter-domain reuse and rwduces the cost of future sof~arw maintenance. Modeling of multiple, interrelated domains help in planning so@vare Wstems and in applying rwuse at the enterprise level.