A Framework for the Automatic Analysis and Interactive Exploration of Document Aesthetics. Technical Report April 21, 2016

Modern word processing software and typesetting systems such as TeX enable the quick creation of documents of various kinds. Although the quality of the software packages varies, all can produce aesthetically pleasing documents in terms of layout and type setting. Problems typically originate from the large number of parameters which are exposed to the user. These range from simple settings like typeface, font size and column width to more elaborate ones, such as kerning and leading. Most often default values are modified without grasping the consequences for readability and aesthetic appeal of the resulting document.In this paper, we present a system for interactive visualization and exploration of quantifiable aspects of document aesthetics such as alignment, spacing, gray values, but also of image color harmony. This system also allows for comparative analysis of multiple documents and document versions side-by-side. The documents are rated using an extensible and parameterizable plug-in system allowing the user to define a task-specific processing pipeline interactively. The rating is hierarchically organized such that the user can drill down into the different aspects that influence the final score. Our system takes standard document formats such as Adobe PDF or Microsoft XPS as input. Our system serves as a platform for further research on document aesthetics as well as a utility to sensibilize authors for these often underestimated aspects of scientific publishing