Authoring and publishing adaptive diagrams
thesisposted on 29.11.2016, 04:33 by McCormack, Cameron
Diagrams are an important part of visual communication and have been used for thousands of years to convey ideas and information. Today, the most widespread means of publishing diagrams and other documents is the World Wide Web. The Web was designed to support documents being presented on a range of target devices and output media with widely varying capabilities. Textual documents can be made to adjust their layout and styling to respond to display size, the readerâ€™s preferred font, colour depth, and more, without requiring the author to create many different versions of the document explicitly. Equivalent functionality is needed for diagrammatic content, however this is not supported by existing tools or standards. This thesis explores adaptive diagrams, which are diagrams that can adjust their presentation to different viewing contexts. The requirements for diagram adaptation are investigated and constraint-based tools and a proposed publication format are developed to help authors create and publish adaptive diagrams. A diagram collection and adaptation exploration exercise is presented, the result of which is a categorisation of the reasons for adapting diagrams and of the types of adaptation that can be performed on them. This is the first known attempt at classifying diagram adaptation requirements, and is used as the basis for the remainder of the work presented. A constraint-based extension to the Webâ€™s vector graphics format, SVG, is developed, which allows adaptive diagrams to be written and published more easily. The formatâ€™s ability to represent all of the diagram adaptation types identified in the exploration exercise is shown. The feasibility of a performant implementation of the publication format is demonstrated via an extension to an existing SVG toolkit which completely implements the proposal. A high level model for adaptive diagrams and an interactive, constraint-based authoring tool for constructing diagrams according to this model is presented. The model supports several different types of adaptive behaviour and the tool allows all of them to be specified in an aspect-oriented manner, helping the author to create many different versions of a diagram without needing to create each combination explicitly.A qualitative user study is conducted to evaluate the usability of the authoring tool and the ability for authors to understand its underyling model. The results show that, with varying degrees of help to overcome difficulties encountered during the authoring process, participants familiar with traditional diagram authoring tools and with Web or multimedia design experience can successfully use the tool to produce adaptive diagrams. Comments from participants on the usefulness of the tool were very positive. Participants also performed a diagram adaptation sketching task, the results of which provide evidence of the comprehensiveness of the diagram adaptation type categorisation. Finally, the adaptive diagram publication format is further evaluated by investigating its suitability as an export format for the authoring tool. A comparison is made with standard, scripted SVG, showing that the proposed format provides significant file size reductions and can represent the adaptive behaviour supported by the tool in a more straightforward manner than the scripted alternative.