A Design Pattern Language to Assist the Design of Alarm Visualizations for Operating Control Systems

This section includes the design pattern language proposed. It is composed by a collection of interrelated 29 design patterns organized by different criteria that describe the design space for alarm visualizations.

The Source Material for the Design Pattern Language

The first and perhaps most direct factor contributing to the quality of a design pattern language is the quality of the corpus of source material. This design pattern language for alarm visualization design results from both the feedback provided by designers and an extensive review of design principles, standards and guidelines, controls systems reports, and visualization and interaction techniques from Alarm Management, Human Factors, and Visualization. In particular, the literature reviewed can be classified into two broad categories:

  • Grey literature. Informally published written material (such as reports) that may be difficult to trace via conventional channels such as published journals and monographs because it is not published commercially or is not widely accessible.
  • Peer-reviewed literature. Articles that have been evaluated by several researchers or subject specialist in the academic community prior to accepting it for publication.

Grey literature Standards and guidelines for alarm management Control systems reports Alarm system displays reports Visualization systems reports
Peer-reviewed literature Design principles for human factors Models for visualization design Models for alarm-initiated activities Visualization design principles Visualization techniques Interaction techniques

Design Patterns Categories

Design patterns should attend to the needs of designers for addressing specific design problems. Aiming at achieving this purpose, it is required to classify design patterns according to different criteria, being possible to have the same collection of patterns categorized under a variety of classification schemes. In this work, the classification scheme is divided into two main dimensions: Purpose and Level of abstraction. This classification scheme can be characterized as orthogonal. Orthogonality means that any of the design patterns classified by Purpose may be classified as well by Level of abstraction.

  • Purpose. This classification dimension addresses the requirements of organization, interaction, and visual encoding of data to build an effective alarm visualization. As a consequence, this dimension establishes the following categories:
    • Presentation. The patterns included in this category are related to design decisions about how to structure the interface of an alarm visualization.
    • Representation. The patterns included in this category refer to the assignment of specific visual characteristics to alarm data attributes in order to facilitate visual sense making by human operators.
    • Interaction. The patterns included in this category are associated with those elements that allow the human operator producing a change in a view of a corpus of alarm data, easing the acquisition of insight.
  • Level of abstraction. This classification dimension specifies whether the design pattern reflects a contextualization of a design problem and its solution, thus, the design pattern is very general, or deal with a specialization of the problem the general pattern addresses, having a similar solution structure but more specialised. Accordingly, this dimension is composed by the following categories:
    • Feature-level. The patterns classified under this category describe visual features that an alarm visualization should support. A feature in this work is a grouping of visual capabilities that provides value to the user.
    • Mechanism-oriented. The patterns classified under this category are of a detailed nature depicting how the visual features of an alarm visualization should be addressed. In particular, they describe different visual mechanisms such as visual structures and view transformations that should be applied to support such visual features

Design Patterns Relationships

Like all languages, a design pattern language has syntax and grammar. The language syntax links the pattern into a web of other needed patterns. In particular, this design pattern language displays five types of relationships among patterns:

  • Composition relationship. This relationship indicates that a pattern is composed of one or more other patterns.
  • Combination relationship. This relationship indicates an alliance of patterns to be applied together in order to address a specific design purpose.
  • Alternative relationship. This relationship indicates several design patterns that lead more or less, to the same result.
  • Generalization relationship. This relationship indicates that a pattern has generic features to serve a more universal purpose.
  • Specialization relationship. This relationship indicates when a pattern shares the same functionality than other but possesses more specialized characteristics or features.
Language graphInteractive version of the design pattern language