Diagrammatic Representations and Dependability
Diagrammatic Representation, Knowledge Representation, Structures for Dependability
The use of diagrammatic representations is common throughout engineering and design practice. Diagrams are popular, as many people find them more readily accessible than other forms of representation. Diagrams are also effective at presenting the big picture; that is, diagrams can typically contain far more visible structure than any text-based representation and this structure can be used to reflect the structure of whatever it is that the diagram represents. Diagrams are thus particularly popular and effective in design, where they are typically most effective at presenting high level overviews of entire systems, in which the relationships and interactions between components is highly visible, and thus more readily accessible.
An illustrative example of the significance of a well chosen representation in facilitating communication across technical boundaries for highly dependable systems is the HAZOPS (HAZard and OPerability Studies) hazard analysis technique. HAZOPS is a technique that originated in the chemical industries which involves engineers and experts from a broad range of technical disciplines holding a series of structured brainstorming sessions to identify and assess the potential hazards of a proposed design. A typical HAZOPS is oriented around a diagram of the proposed design. For chemical plants, schematics of the physical plant layout (piping and instrumentation diagrams) are used.
The HAZOPS team examine in turn each component depicted on the diagram and consider the hazards and likelihood of failures or deviations from its intended function. Typically each team member will have access to that information on the proposed design which is relevant to their field of expertise. Thus the team is able to bring a great breadth of experience and data to the analysis yet, by coordinating the analysis around a common focus (the diagram), individual team members need not be concerned with information beyond their own area of expertise. Furthermore, the diagram used in a HAZOPS typically represents the proposed design at a general enough level to be clearly understood by all team members, regardless of technical discipline and expertise, while still being sufficiently detailed to make an analysis based upon it worthwhile. The diagram thus plays the role of a communication artifact, an entity which guides and supports communication concerning the system under analysis.
An effective diagram is typically taken to be one that is well matched to what it represents. This is to say, that the logical and spatio-visual properties of structures inherent to the diagram are chosen so as to have some very direct correspondence with the structures that they represent in the semantic domain; and in particular that they are chosen so as to support desired reasoning tasks by making certain inferences immediate and obvious. Our work investigates the potential impact of proposed technological systems.
Using suitable (diagrammatic) representations enhances the visibility and accessibility of the assessed impact to a broad range of stakeholders. Diagrammatic languages and notations capture the implications of our social analyses in accessible forms. This facilitates the communication of knowledge during the design and deployment of complex, highly dependable computer-based systems across technical and non-technical domain boundaries. Our work highlights guidelines for both the design of effective diagrammatic languages, and the design of specific diagrams within such languages. These guidelines draw upon results from visual language theory, cognitive science, empirical psychology and graphic design. Integrating results from such diverse fields is a non-trivial task, which is here approached through a decomposition of the study of issues of effectiveness in diagrammatic languages according to analogous understandings of (written and spoken) natural languages.
 Corin Gurr, Gillian Hardstone. Implementing Configurable Information Systems: A Combined Social Science and Cognitive Science Approach. In M. Beynon, C.L. Nehaniv, K. Dautenhahn (Eds.), Proceedings of 4th International Conference on Cognitive Technology: Instruments of Mind, CT 2001, Springer LNAI 2117, Warwick, UK, August 6-9, 2001, pp. 391-404.
 Corin Gurr. Computational diagrammatics: diagrams and structure. In Denis Besnard, Cristina Gacek and Cliff B. Jones (Eds.), Structure for Dependability: Computer-based Systems from an Interdisciplinary Perspective, Springer (to appear).
Massimo Felici (Edinburgh)
|Page Maintainer: email@example.com||Credits||Project Members only||Last Modified: 10 August, 2005|