This article introduces the generic Document Towers paradigm, visualization, and software for visualizing the structure of paginated documents, based on the metaphor of documents-as-architecture. The Document Towers visualizations resemble three-dimensional building models and represent the physical boundaries of logical (e.g., titles, images), semantic (e.g., topics, named entities), graphical (e.g., typefaces, colors), and other types of information with spatial extent as a stack of rooms and floors. The software takes as input user-supplied JSON-formatted coordinates and labels of document entities, or extracts them itself from ALTO and InDesign IDML files. The Document Towers paradigm and visualization enable information systems to support information behaviors other than goal-oriented searches. Visualization encourages exploration by generating panoramic overviews and fostering serendipitous insights, while the use of metaphors assists with comprehension of the representations through the application of a familiar cognitive model. Document Towers visualizations also provide access to types of information other than textual content, specifically by means of their physical structure, which corresponds to the material, logical, semantic, and contextual aspects of documents. Visualization renders documents transparent, making the invisible visible and facilitating analysis at a glance and without the need for physical manipulation. Keyword searches and other language-based interactions with documents must be clearly expressed and will return only answers to questions asked; by contrast, visual observation is well suited to fuzzy goals and uncovering unexpected aspects of the data.

This article introduces Structural Information Potential (SIP), a measure of information based on the potential of structures to be informative about their content. An example of this concept is the clustered appearance that typically characterizes the first page of scientific articles, which summarizes the article's contents and provides additional data, yielding potentially the largest and most diverse amount of information from a single page in the shortest time with the least effort. This characteristic makes SIP particularly welladapted to triage tasks (i.e., rapid decision-making under conditions of uncertainty and limited resources), an application illustrated by means of a case study on classifying document images. The SIP method consists in unifying the Shannon entropy, the Fourier transform, the fractal dimension, and the golden ratio into a single equation and several algorithmic components. While the application domain is document images, the concept has generic character. The method results in a mathematically and perceptually coherent pattern space, characterized by continuous transition between uniform, clustered, and regular configurations, and corresponding to a Structural Information Potential with a well-defined maximum. The maximum SIP leads to the identification of shapes and patterns with minimal structural redundancy, termed "fluorescent objects" as a complement to regular graphs and the Platonic solids.

This article argues in favor of representing the spatial distribution of information within and between documents, by surveying a broad variety of potential applications, including the entire document lifecycle, multiple sensory modalities, and a large spectrum of tasks and users. The theoretical explanations of this richness are a further facet of the article, and can be summarized as follows: (1) insights emerge from focusing on information structure, rather than information meaning; (2) spatializing information creates new information; (3) simplification increases the polyvalence of representation models; (4) introducing mystery in communication channels motivates discovery and diversifies insights; (5) approaching information design as a Gesamtkunstwerk multiplies the applications; (6) information is a manifestation of a link between structures and the actions these enable, while information design is the art and science of creating such links. The argument is developed around the concrete example of a document structure visualization, the Document Towers, which uses the metaphor of architectural models to represent documents.

No abstract

Revue bibliographique pour le domaine irano-aryen Volume 24 | 2003Comptes rendus des publications de 2001