A study of invasive composition for the evolution of a health information system

Abstract

In this paper we show that some of the evolution tasks in OpenMRS, a health information system, may require the invasive modification of interfaces and implementations in order to offer an appropriate modularization. We introduce a new composition framework in Java that supports the definition of expressive pattern-based invasive compositions. Furthermore, we show that the composition framework allows us to concisely define an evolution scenario of OpenMRS that supports the consolidation of patient data from different remote instances.