System and Context Modeling; Visualizations of Where, When, and How

Modeling of systems and their context is done to support communication with stakeholders, to facilitate reasoning about system requirements and design, to support decision making, and in general to create and maintain understanding, insight, and overview. One of the challenges in modeling is to find appropriate representations that are understandable for relevant stakeholders and that contain sufficient information. In this paper, we discuss how to visualize space, time and process flow of systems, both internal as well as in their context. We argue that visualizations must connect to the stakeholders' world. We observe that architects tend to create abstract diagrams, alienating their stakeholders by abstraction, the use of jargon, and the presence of lots of (irrelevant) details. We will illustrate the visualization techniques with some models made as part of the Darwin project. The Darwin project researches evolvability of systems and architectures, using Magnetic Resonance Imaging scanners at Philips Healthcare as case. Introduction The IEEE 1471 standard [IEEE1471 2000, Hilliard 2001] introduces the notion that architecture descriptions consists of models. Models capture one or more specific views. A viewpoint is the combination of stakeholder and a specific concern. Following this line of reasoning many models are needed to create an architecture description. Standards such as DoDAF [DoDAF 2003] provide many predefined views. Currently the term model is also fashionable as means to streamline the engineering process. Model-Based Design [MathWorks 2009], Model Driven Architecture [OMG 2009], or Model Driven Engineering are approaches where models are used to specify, but also to design and engineer. These approaches are propagated by commercial tool vendors, such as MathWorks and IBM. The tools support the creation and management of models. These models are also used to generate the final realization details, without further human intervention. In this paper we discuss the use of models as a means of communication and discussion between stakeholders and as a means to reach design decisions that fit in the context and life cycle of the system-of-interest. These models follow the idea behind IEEE1471, where models relate to stakeholders and concerns. These models serve mostly to help human stakeholders to understand problem and solution space and to reason about choices. The emphasis of this paper is on three basic visualizations: space (where), time (when), and process flow (how). Cernosek and Naiburg explain in [Cernosek 2004] that the value of 1 This work has been carried out as part of the Darwin project at Philips Healthcare under the responsibility of the Embedded Systems Institute. This project is partially supported by the Netherlands Ministry of Economic Affairs under the BSIK program.