A MANDALA BROWSER USER STUDY: VISUALIZING XML VERSIONS OF SHAKESPEARE'S PLAYS

Visible Language, 2009 by Gainor, Rhiannon, Sinclair, Stefan, Ruecker, Stan, Patey, Matt, Gabriele, Sandra

ABSTRACT

We report the results of a small user study of a visual XML browsing prototype, called the Mandala browser, where dots representing entire documents or portions of documents are plotted around the periphery of a circle and drawn inward by colored magnets that are assigned values by the user. The result is akin to a Venn diagram that provides a visual representation of the interaction between multiple Boolean queries. In this study, eleven participants were given a pre-study interview, then asked to carry out a series of tasks where the dots represented speeches in plays by Shakespeare and finally were debriefed in a concluding interview.

We gained from this study a range of valuable insights into how details of the Mandala browser design could be improved. Participants mentioned, for instance, that they would like to retain a connection between results and the visualizations that produced them, that they would like to be able to make notes on result sets, and that they would like to be able to save subsets within results. They also asked for tools that support collaborative searching, as well as for federated searching across collections. The user feedback confirmed the potential value of the Mandala interface and provided guidance for the next iteration of development.

(ProQuest: ... denotes obscured text omitted.)

INTRODUCTION

Browsing interfaces are intended for people interested -fn gajmng some understanding of the contents of a collection, or in some cases pftflftse results from an initial search retrieval. Browsing interfaces can therefore" be seen as having a complementary relationship with retrieval interfaces, where the goal is to obtain a retrieval target. A typical approach to information browsing is to provide the user with a list of documents, which may be ordered according to some convention such as alphabetically by author or title, or chronologically by date of publication (sometimes called faceted browsing). A more visually complex approach is to organize documents in nested tiles where relevant information is expressed through the position, dimension and appearance of each tile (Schneiderman, 1992). Yet another approach is the Scatter/Gather browser proposed by Pirolli et al. (1996), who developed a prototype visualization where documents were represented by dots that could be grouped and subsetted dynamically by the user.

The literature includes a wide variety of examples of visually compelling browsing interfaces, including Small (1996) who proposed a 3D prospect view for browsing texts of Shakespeare's plays and Bederson (2001), who described asystemfororganizingthumbnailsof images. More recently, designers working with public APIs (programmatic interfaces) from social networking sites like www.flickr.com have developed a number of browsing interfaces that extend Bederson's ideas through tools suchas the color picker by Bumgardner (2005). Another example is the orbiting globes of information at http://dartdesign.de/, which allow the user to browse a constellation of moving representations that rotate under user control. The various crystal-based displays by Spoerri (2007) provide still further examples, and the 'dust & magnet' project by Yi et al. (2005) shows an inventive and interactive use of the magnet metaphor. In the context of rich-prospect browsing (Ruecker 2003), each of these interfaces includes some degree of meaningful representation, coupled with the provision of tools for manipulatingthe display. Browsing interfaces can also be used for studying parts of individual documents, such as the speeches in a play.

In addition to the discussion of browsing interfaces and their features, the literature also includes discussion of the methods of evaluation. For instance, Plaisant (2004) suggests the need for new and more comprehensive strategies for evaluating the success of browsing interface designs.

While most browsing interfaces provide both some representation of individual items and tools for manipulating the display, not all are concerned with the structural arrangement of the items (word clouds created at Wordle.org provide an example of arbitrary, though visually appealing, structural arrangements). However, in outlining their agenda for a phenomenological approach to interface design, Winograd and Flores (1 986) suggestthe possible valueto the user of making an explrete structural coupling between the arrangement of information on screen and the underlying structure of the digital information in a collection. They suggest that this kind of coupling has the potential to provide the user with a number of insights into the nature of a collection - insights that would not otherwise be possible.

Ruecker (2003) outlines a number of areas in which these possible insights might occur, including: contents, structure, context, features, limitations, connecti trends, anomalies, navigation. In addition, he describes three advantages that accrue from the meaningful representation of content items, including reminders of the existence of items that otherwise mightescabiheuser's attention, cognitive reassurance and reduced helplessness. Reducing the helplessness of users and increasing their sense of cognitive reassurance site important goals of information design.