Examining, Documenting, and Modeling the Problem Space of a Variable Domain

摘要

In order to reduce development time and costs associated with producing a series of related applications, it is necessary to stop focusing on each application individually. Instead a system family approach can be used to understand commonalities and differences between individual systems. Domain engineering methods provide guidance on how to achieve this change in focus. Generative programming can help reduce these costs even further by enabling us to move beyond the usual practice of manually searching for and then adapting relevant components. This report describes the topics of domain engineering and generative programming, examines various current methods of domain engineering, and proposes a new domain engineering method that will be used to capture and portray a problem domain. The domain engineering process proposed by this report limits its concern to the problem space of a domain. The process consists often steps that can be followed to effectively understand, document and model the problem space. This process combines important aspects of existing domain engineering processes into one method. It also adds several elements that current processes lack including: the use of updated modeling techniques (the use of UML) and giving early consideration to concerns arising from the distributed heterogeneous programming environment that is becoming more and more common today. The process also specifies a grammar that can be used to describe the contents of an extended feature model. The proposed method is explained step by step and is illustrated through the use of an example problem.