A Deep Orthographic Modelling Environment

As software systems grow in complexity the need to allow their stakeholders to fulfil their goals using optimal representations of the system has assumed increased importance. This need is ubiquitous, spanning all tasks, all levels of abstraction and all phases of a system's life-cycle and must be supported in the face of continual change to a system's structure, behaviour and knowledge. The potential of view-based modelling environments has been realised for decades, but the underlying approaches still have significant weaknesses. Most still employ a synthetic approach, adopt arbitrary ad-hoc viewpoint frameworks, provide limited support for customising the viewpoint framework and focus their support on the development phase of a system's life-cycle, with little support for other phases.
This thesis presents an approach that overcomes these weaknesses by integrating and extending two emerging modelling technologies -- orthographic software modelling and deep modelling. The former provides support for projection-based views and an extensible, multi-faceted viewpoint framework, the latter provides support for flexibly and seamlessly modelling information over multiple classification levels. The developed approach offers an integrated, extensible and generic solution which can be customised for use with arbitrary view-based modelling methods. The validity of the approach has been evaluated by an Eclipse-based prototype. The prototype was customised for the domain of software development using the KobrA method and enterprise architecture modelling using the MEMO method and was used to model two example systems.