Project
Visual Understanding Environment 2.0

Start Date
October 2005

Website

http://vue.tufts.edu/

Overview & Goals

The Visual Understanding Environment (VUE) project based at Tufts University is focused on creating flexible tools and processes for integrating digital resources into teaching, learning and research. Designed with the higher education community in mind, VUE provides a compelling and accessible visual environment for structuring, presenting, and sharing digital information in support of scholarship. VUE’s support for establishing and communicating meaningful relationships among ideas and digital resources is unique compared to other educational and research applications. In addition, VUE’s visual approach to representing relationships among digital resources combined with easy-to-use tools for annotating and exploring these semantic structures greatly facilitates thinking and reasoning with digital information. Designed with higher education in mind, the functionality that differentiates VUE from other concept mapping tools is drawn from an understanding of the centrality of digital information and published resources in scholarly activity and the many associated tasks of structuring and managing digital content. VUE is designed from the ground up to support forms of thinking, teaching, learning and researching which involve interaction with rich collections of digital information.

The scope of work for the VUE 2.0 project focuses on the following four areas with the primary goal of enhancing scholarship with digital information.

1. Position VUE as a flexible interface to a greatly expanded collection of digital libraries and repositories.

2. Promote interactive teaching with VUE, by providing map-based presentation tools.
    
3. Advance the analysis and evaluation of conceptual understanding expressed in VUE maps by adding support for semantic mapping techniques based upon ontologies and RDF.

4. Reduce barriers to integrating VUE into workflows involving learning management systems such as Sakai.

This additional functionality builds upon a number of VUE’s existing tools and is made possible by VUE’s open technical architecture.

Project Participants

The VUE 2.0 project based at Tufts University is a collaborative effort involving staff from Tufts University Academic Technology and MIT’s Academic Computing with faculty participation from both institutions. David Kahle, Tufts University Director of Academic Technology, is the Principal Investigator (PI) for the VUE project and Jeff Merriman, MIT Associate Director for Software Development and Strategy, is the Co-PI. In addition to the core project team, faculty representing a number of academic disciplines, including art history, computer science, medicine and public health, continue to contribute directly to VUE’s evolution by providing compelling, real-world use cases involving teaching and research with digital information.

Project Highlights

The Visual Understanding Environment has evolved into a valued tool for both educators and researchers by thoughtfully combining tools for accessing, organizing, analyzing and presenting information and ideas within one application. As a result of recent development efforts, VUE has become one of the most sophisticated applications available to scholars for aggregating digital resources from various sources and for visually structuring and communicating critical relationships among concepts and supporting materials.

During this two-year project sponsored by the Andrew W. Mellon Foundation, the VUE team has focused on improving access to scholarly information and integrating VUE into academic workflows. VUE 2.0 beta, released in November of 2007, provides an enhanced user interface for discovering and connecting to repositories of digital content. Related datasource architectural improvements make it easy to package and integrate OKI compliant repository plug-ins for VUE, thus providing numerous digital library owners with a responsive and flexible visual interface for working with their collections with only a modest investment in local software development. VUE has also advanced as a partner application for leading digital repository and learning management systems. Recent support for information ontologies and RDF enable users of the FEDORA digital repository system to create and publish digital collections through VUE’s rich interface. Likewise, scholars may now upload VUE maps and referenced content to the Sakai collaboration and learning environment in support of information sharing and collaboration. Perhaps the most visible outcome of recent design and development activity, and that which is currently attracting the greatest user attention, is VUE’s support for presenting content maps. In direct response to faculty suggestions, VUE now provides a toolset for creating interactive presentations from content placed in VUE maps. VUE’s familiar flexibility in structuring digital content was an important design criterion for this new feature and, as a result, presenters may now style and display information as linear sequences or in a nonlinear, interactive manner. An illustration of VUE’s presentation feature is provided in Figure 1 on the following page.

Figure 1: VUE Presentation Tools

Deliverables & Milestones

Below is a summary of the main deliverables of the past 12 months and planned future VUE development.

Semantic Mapping Tools

The VUE 2.0 project has introduced tools for constructing maps from information ontologies. This functionality enables groups of scholars to build maps using a controlled vocabulary which may then be analyzed, compared and merged based on shared concepts. Through VUE, the elements of OWL or RDF-S ontologies may also be associated with visual characteristics or styles for display. An important test case for VUE’s support for ontologies was successfully integrating VUE with the FEDORA digital repository system. Collections of objects may now be created within VUE using the FEDORA relationship ontology and published directly to the repository.

New keyword “tagging” and search functionality was introduced this year to facilitate information discovery and map filtering. Metadata about map content, including relationships among resource nodes, is represented in RDF and managed by an in-memory triple store. VUE’s updated query tools have been built using the Jena semantic web framework which provides an important foundation for importing, mapping and exploring a variety of RDF data structures.

Presentation Tools

The presentation features that were implemented in prototype form last year have been completed along with several enhancements to this toolset. Using VUE’s pathway feature, any node on a content map may now become an element in one or more VUE presentations and displayed in full-screen mode. Map nodes can have multiple representations where embedded content such as images and online resources are styled for presentation purposes. By constructing presentations as pathways, scholars can prepare guided tours through their materials for external audiences without impacting the underlying semantic structure of their content map. Figure 2 illustrates the revised pathways panel along with a portion of a map where some of the image nodes, Olympia and Déjeuner sur l'Herbe, appear on two separate presentation pathways. Using VUE’s presentation controls, scholars can zoom to edit the slide previews or launch directly into full-screen mode from the pathway.

          Figure 2: Multiple Presentation Pathways

The new presentation tools also provide several ways of navigating through VUE material in real time. One can choose to simply follow a pathway in linear fashion or choose to branch out to other connected map nodes or to different presentation pathways. Figure 3 captures two options for nonlinear navigation of VUE content, an interactive map overlay and the side menu that reveals all map content connected to the item currently displayed. These features enable presenters and their audiences to focus on a particular slide or concept while maintaining a sense of the overall context.

Figure 3: Nonlinear Navigation

Integration with Learning Management Systems

In addition to publishing maps and content to FEDORA repositories, VUE users may now publish their maps and related digital resources to various course spaces within the Sakai collaboration and learning environment. This functionality greatly facilitates the sharing of VUE maps among students and faculty and embeds VUE into many online teaching and learning activities. When a local VUE map is published to Sakai, the map and its digital resources are uploaded to the Sakai repository. The map’s resource pointers are updated to reflect the new online location (URLs) of the content. This recent development work involved integrating VUE with the resource management components of Sakai via an OKI-based digital repository plug-in. This integration also required exposing Sakai’s repository and course management services to external applications such as VUE. Below is a screen capture of the publishing interface and process for sending VUE maps to Sakai.

         Figure 4: Publishing Maps to the Sakai CLE

Looking Ahead

This current phase of VUE development will end with the production release of VUE 2.0 scheduled for the end of March 2008. In addition to refining some of the existing features and fixing user-reported bugs, VUE 2.0 will introduce a revised method of packaging VUE files for easier transport and exchange. The release of VUE 2.0 in March will accomplish the goals outlined in the VUE 2.0 project proposal to the Andrew W. Mellon Foundation and represents a significant event in VUE’s evolution.

Looking ahead, we plan to expand on VUE’s strong foundation of user-centered tools for visually expressing and annotating relationships among digital documents by focusing future development efforts on making research data sets more accessible and meaningful to scholars. There is currently a gap in open research tools that enable manual mapping and annotation of data and those that rely solely on computer guided data display. We plan to combine methods for semi-automated map generation with VUE’s user-driven annotation and manipulation techniques to support a wider range of research use cases. Rather than creating content maps by hand, one resource at a time, a scholar may wish to import a set of data for initial display within VUE and then use the application’s rich suite of visual annotation tools to correct, refine and structure information in a meaningful manner.

As an example of this planned functionality, imagine a scholar who has obtained three sets of tabular data that she wishes to use as the basis for constructing a VUE map about Rockwell Kent’s artistic production by year, location and painting medium. Using VUE’s data mapping control panel, she imports the first data set that contains a chronology of Rockwell Kent’s travels and chooses to map only the locations where he painted during his 70-year career. VUE represents each record in the data set as a node, drawing labels from the “location” column of the table. With the locations of Kent’s travels now mapped, she now imports the records for each of his paintings and chooses to join this data to the existing map using the “date” column of the record set. The map expands to accommodate the new nodes representing Kent’s paintings, which are visually clustered around the location nodes. As a final step in the process, she incorporates a data set that lists Kent’s work by medium onto a new VUE layer, choosing to style this new information by shape and selecting the title of work as the joining key. The result is another cluster of information illustrating Kent’s choice of medium not only for each painting but also, by association, for the dates and places where he worked. Below is mock-up of the map for this scenario.

         Figure 5: Dynamic Mapping of Tabular Data

The way in which software is designed and packaged is a key factor in sustaining and evolving an application over time. Thus, in addition to accommodating research data, we also plan to make VUE functionality available to other applications, web-based or desktop, by developing smaller, task specific components derived from VUE’s existing code base. An important motivation for pursuing this objective is our interest in exploring VUE’s potential interaction with web-based tools such as Zotero and as components within SEASR workflows. One important component serving many use cases is the development of a lightweight VUE applet for viewing and manipulating VUE data within a web browser.

Community & Sustainability

Our goal for sustainability is ultimately one of impact – that the essential work of more scholars is advanced by VUE. Sustaining the project into the future will require increasing VUE’s reach and adoption along with a combination of institutional support, grant funding and active participation by technologists and scholars. Having an institutional home, Tufts University, VUE has an important advantage which many open-source projects lack. As a supported application in use by Tufts students and faculty, assistance and support is available to resolve basic technical problems, fix bugs and maintain a distribution of the application. Equally important is direct access to a user community willing to assist the development team in advancing the product. VUE enjoys solid institutional support, which is one of the key factors for sustaining the project.

Institutional support alone, however, is insufficient for making substantial changes to the VUE code base. Compelling suggestions made by VUE users for new features or strategic opportunities to extend VUE functionality in support of the emerging needs of an academic community cannot be pursued at present without external resources. Our most recent program accomplishments along with planned development activities are designed in part to address this issue. Positioning VUE as a tool for mapping and annotating data will dramatically increase the number of research activities for which VUE is the ideal tool. Likewise, packaging VUE components for integration into web application frameworks and creating a browser-based applet will increase VUE visibility among the general public. While increasing the size of the VUE user community is no guarantee that people interested in contributing to VUE’s development will come forward, we believe it is probable. A relatively small group of contributors is sufficient to sustain VUE development activities. Establishing VUE as a flexible tool for data visualization and annotation would also provide a foundation for proposals to relevant federal funding programs. Examples include the NSF Human-Computer Interaction (HCI) program and the NSF Foundations of Data and Visual Analytics (FODAVA) program. Finally, we believe that outreach to potential VUE partners is one of the most effective means of establishing future productive collaborations. As discussed below, these include partners funded by the Mellon Foundation including ARTstor, SEASR, Simile, Sophie and Zotero.

Synergistic Opportunities
I believe there are a number of shared interests among the VUE team and other Mellon funded projects. This is by no means an exhaustive list, but the following possibilities seem particularly compelling. Both VUE and Zotero are focused on providing tools for annotating and organizing digital resources in support of research. There are mutually beneficial opportunities for connecting VUE to Zotero that would provide VUE with a means of storing web-based resources locally and, at the same time, offer a rich visual interface to Zotero users for organizing and establishing relationships among their resources. VUE is also well positioned to partner with ARTstor. VUE has long provided access to ARTstor content via a repository plug-in and VUE’s new presentation tools include flexible tools for presenting digital images. Extending VUE to support ARTstor’s encrypted high-resolution images would be a natural next step and would provide scholars with a new option for interacting with ARTstor collections. Members of the VUE and SEASR teams have been exchanging technical information over the past few months and discussing ideas of where SEASR and VUE could come together in support of research. As discussed above, VUE’s future development plans call for wrapping core VUE functionality as components that could be used within the SEASR workflow framework. We would also like to explore connections between VUE and Sophie.  Both projects are developing client applications for creating presentations of one form or another using digital resources. In addition to sharing GUI conventions and advances, one could imagine exporting a VUE presentation as a Sophie book or creating a map component that could be manipulated within a Sophie book. We also have much to learn from and share with the Simile project at MIT. An important goal for VUE’s future development is accommodating research data expressed in RDF and building new tools for visualizing and exploring data. Finally, with VUE connected to FEDORA and Sakai, we would hope to see use cases emerge that would form the basis of future collaboration. External to the Mellon projects, VUE has been used as a partner application for constructing activity flows for virtual patient cases within Labyrinth, an online activity modeling system. Future collaboration with Labyrinth would result in tighter integration of the two systems based upon VUE’s new support for ontology-base constrained mapping and RDF.