Domain Engineering - A Software Engineering discipline in Need of Research

Before software can be developed its requirements must be stated. Before requirements can be expressed the application domain must be understood. In this paper we outline some of the basic facets of domain engineering. Domains seem, it is our experience, far more stable than computing requirements, and these again seem more stable than software designs. Thus, almost like the universal laws of physics, it pays off to first develop theories of domains. But domain engineering, as in fact also requirements engineering, really is in need of thoroughly researched development principles, techniques and tools. The aim of this paper is to advocate: that researchers study these development method components, and that universities focus their education on basing well-nigh any course on the use of formal techniques: Specification and verification, and that software engineers take heed: Start applying formal techniques. A brief example of describing stake-holder perspectives will be given - on the background of which we then proceed to survey the notions of domain intrinsics, domain support technologies, domain management & organisation, domain rules & regulations, domain human behaviour, etc. We show elsewhere how to "derive" requirements from domain descriptions. Domain requirements: by domain projection, instantiation, extension and initialisation; interface requirements: multi-media, dialogue, etc.; and machine requirements: performance, dependability (reliability, availability, accessability, safety, etc.), and maintainability (adaptability, perfectability and correctability). The current paper presents work-in-progress. The text of the paper is therefore very schematic.