Increasingly, scholars have access to large, heterogeneous collections of information such as those provided by the Digital Public Library of America, Europeana, or Canadiana as well as more targeted thematic research collections like the Whitman Archive. In additions to purpose-built humanities projects, scholars commonly make use of general-purpose resources such as Flickr, Google Search, HathiTrust Digital Library and many other services.
Traditional search interfaces allow scholars to rapidly search for specific items and to explore a collection via facetted search interfaces and other techniques. Search, however, is only one part of a complex ecosystem of behaviors associated with information seeking (Toms and O’Brien, 2008; Buchanan et al., 2005). In practice, scholarly information seeking is not characterized by a single search event or interaction with a single system. Rather, it is a process that takes place over an extended period of time and involves searching different sources for potentially relevant material. As material is found, scholars organize, annotate and make notes about the retrieved material; activities that lead to more questions and more seeking.
These tasks serve a dual purpose of cataloging material for re-discovery, use and interpretation as well as engaging the participant in a process of sensemaking and internalization. The fundamental purpose of this work is not merely the discovery, consumption or indexing of information or even the production of a concrete research output such as a paper. Instead, scholarly information seeking serves to hel
p the seeker develop an internalized, systematic understanding of a body of knowledge and cultivate a distinctive interpretive voice (Audenaert and Furuta, 2010).
We are developing the Visual Workspace for Information Seeking and Exploration (vWise 1) to support these tasks. This paper introduces the key features and capabilities of the vWise platform and discusses how those features and capabilities reflect and build on a theoretical framework for designing systems to support open-ended information seeking and exploration.
vWise is an extensible Web-based application framework that allows scholars to search for content from different data providers and bring the resulting information together into a single display. Once resources have been added to the workspace, they can then organize and reframe that information, for example by culling unwanted results or juxtaposing digital surrogates for newly discovered relationships.
Panels are the primary way people interact with content in the vWise interface. They are used to display different types of information drawn from different sources in a single workspace. Users can organize this information by rearranging panels within the workspace and add visual annotations by setting properties such as background color, border style, font properties and drop shadows. The panels themselves support user interactions tailored to the types of content they display. For instance, users could read a book from HathiTrust, watch a video from YouTube, explore metadata from DPLA or manipulate manuscript page images to improve readability.
Users have several options for adding content panels to a workspace. A basic search service allows them to execute searches using pluggable content providers. Alternatively, custom panel implementations may provide advanced searching capabilities or allow users to drag composite content such as pages from a manuscript onto the workspace to create new panels.
Workspaces provide a site for both individual and collaborative work. To enable collaboration, a user can share a workspace with others. Modifications to this shared workspace are reflected on all users’ display simultaneously. This allows both synchronous and asynchronous collaboration.
As a framework, vWise is designed to be customized and configured by an application developer prior to being deployed. This customization allows the basic application to be extended in four major ways:
Integration with external data sources. vWise provides a pluggable architecture for defining interfaces to external data sources such as a digital library or search service. For instance, we provide data sources that allow users to load videos from YouTube videos, search Wikipedia and display web-pages.
Custom panels for working with data objects. Individual units of content are displayed using panels that provide basic support for moving, resizing and styling. Extensions support rich interactions tailored to specific types of content. For example, an image annotation tool could allow the creation and storage of annotations.
Ad hoc panels for interacting with underlying search services. The core use-case for panels is to display and organize individual content elements. Panels can also be used to implement other application services. For example, we envision implementing a search service panel that will display basic and advanced search options, display facets from the currently active search, maintain a history of recent search and display search results as panels in the main workspace.
Integration with server-side workspace persistence. vWise runs in a browser and stores workspaces and workspace configuration information via a well-defined RESTful interface. Application developers can use the default implementation that comes with vWise or provide a custom implementation. For instance, we developed a custom data storage implementation to connect the vWise interface with a proprietary system used to train emergency responders.
Information seeking and exploration in scholarly research is an intensive, creative activity. Supporting these activities require tools that go beyond merely helping scholars find resources to provide environments that reflect and facilitate the creative process. vWise is based on ideas that have emerged from the literature in hypermedia systems, digital libraries and creativity research, including the following core concepts:
Information Triage: The process of information seeking requires people to rapidly assess the utility of various information resources, discard those that aren’t relevant and prioritize those that show promise (Marshall and Shipman, 1997). vWise goes beyond traditional search interfaces by allowing users to remove individual results, prioritize the remaining objects and combine the results from multiple searches in a single workspace.
Incremental Formalization: One aspect of open-ended information seeking tasks is sense-making—the process of gaining a broad understanding of the structure of ideas within a domain. During sense-making, the emerging organizational framework is partial, provisional and implicit (Shipman and Marshall, 1999). vWise facilitates incremental formalization of knowledge by allowing users to express an emerging organization framework implicitly via the spatial arrangement and visual properties of the panels displayed on the workspace. Within this environment, user can rapidly form and reform organizational structures by manipulating the visual characteristics of the display.
Representational Talkback: The visually expressed knowledge structure serves as an externalized representation of a user’s internal, evolving mental models. Creating and interacting with this content supports representational talkback (Schön, 1983; Nakakoji et al., 2000). Representational talkback occurs when material externalizations of an internal mental model initiate a “reflective conversation” in which they talk back to their creator to inform subsequent stages of the sense making process.
Heterogeneous Sources and Material: Complex information seeking tasks rarely involve one-stop shopping. vWise allows people to integrate, analyze and synthesize information from different sources, by bringing content into a unified information organization space. Manipulation of spatial and visual properties (border color, drop-shadows, backgrounds, etc.) provides a lightweight and open-ended interface manipulating this content. Domain-specific panel implementations allow users to engage in content-specific interactions.
vWise provides support for the complex information seeking needs of scholars and other professionals that goes beyond the capability of traditional search systems. It allows people to gather information from multiple sources and to work either independently or collaboratively to organize and analyze that material using a workspace metaphor.
To date, we have demonstrated prototype implementations of vWise in two different settings. While initially conceived and designed to support the needs of scholars in the humanities, we have integrated a version of vWise into the Emergency Management*Exercise System (EM*ES). EM*ES is a simulation tool used by Texas A&M Engineering Extension Service (TEEX) to train incident managers, supervisors, and jurisdiction officials in the management of a large-scale crisis (TEEX, 2016). We have added vWise as a component to support the training of communication and coordination between emergency managers and cyber response teams. vWise is currently in use in a series of training exercises starting in January of 2016 and running through September 2017. During this period we will investigate options for more widespread use within the EM*ES application.
Our second deployment of vWise was a demonstration prototype for use in wall-sized display of the Humanities Visualization Space (HVS) at Texas A&M University’s Initiative for Digital Humanities, Media and Culture (IDHMC). While vWise is envisioned primarily for desktop-oriented use, when working on knowledge organization tasks, bigger is often better. The HVS provides just such a space and opens intriguing possibilities for scholars to work with this class of interfaces both as individuals and in collaborative settings.
Figure 1: vWise demonstration at the IDHMC Humanities Visualization Space
Moving forward, we continue to develop the core implementation of vWise framework and to add additional content provider connections and panel implementations for displays. We are specifically interested in developing integrations with major content providers for humanities and cultural heritage scholars and are working to pursue such collaborations.
vWise builds on key concepts and findings of a previous project, CritSpace (Audenaert et al., 2010) and provides a new implementation and enhanced capabilities.