For the Workshop on "Information Visualization Software Infrastructures" at IEEE 2004 Visualization,
Organized by Katy Börner, Indiana University, USA and Jean-Daniel Fekete, INRIA, France

Part I

I.1) What functionality should a general InfoVis infrastructure provide?

• Rich data structures: It is important that the concerns of the underlying data be seperated from the concerns of the visualization of the data. Doing so ensures that those interested only in the analysis and modeling of the data need not be forced to address issues of how it should be visualized. It is also important that users be allowed to associate arbitrary meta-data with the underlying entities and relationships reflected in their data so that the data structures can support all of the richness of the data provided by the user. Finally, it is important that the data structures provide mechanisms by which constraints are enforced with regards to the entities and relationships in the data.
• Separation of concerns between the layout, rendering, manipulation, and storage of data: It is important that the modules for determining where the entities should be laid out on the screen, those for determining how the entities and relationships appear on the screen, methods for changing the data itself, and the storage of the data all be seperated so that each module can be substituted for another of the same type and no unnecessary dependencies in the codebase are required.
• Event notification support: It is important that event listeners can be notified when changes caused by alteration of the data structures occur both on-screen and off so that interactivity is fully supported. Allowing developers to define their own events and listeners allows for a truly customized interactive experience.
• Filtering support: By providing an API which allows for structured transformations of the data, common sequences of state changes in the data can be macroized and reused.
• Implementation of standard visualization algorithms and renderers
• Implementation of standard file formats and data representations for storing the data in memory.
• Implementation of default widgets for viewing the data

I.2) What do you see as the main technical challenges for creating a central but flexible and universally useful (information) visualization software infrastructure (as opposed to 100 different ones)?

Part II

Please describe the (information) visualization software infrastructure you are working on.

II.1) Project Name and Web Address
The Java Universal Network/Graph Framework (JUNG), http://jung.sourceforge.net/.

II.2) Core Team Members
Co-founder & developer, Scott White, scott@ics.uci.edu
Co-founder & developer, Danyel Fisher, danyelf@acm.org
Co-founder & developer, Joshua O'Madadhain, jmadden@ics.uci.edu
Webmaster & Release engineer, Yan-Biao Boey, ybboey@users.sourceforge.net

II.3) Project Start Date
February 11, 2003

II.4) Targeted User Group
Educators, researchers, and industry professionals

II.5) Supported User Tasks
As an open-source library, JUNG provides a common framework for graph/network analysis and visualization. For purposes of visualization, JUNG provides a framework that makes it easy to construct tools for the interactive exploration of network data. Users can use one of the layout algorithms provided, or use the framework to create their own custom layouts. In addition, filtering mechanisms are provided which allow users to focus their attention, or their algorithms, on specific portions of the graph.

II.6) Major Features of the System Architecture
The JUNG architecture is designed to support a variety of representations of entities and their relations, such as directed and undirected graphs, multi-modal graphs, graphs with parallel edges, and hypergraphs. It provides a mechanism for annotating graphs, entities, and relations with metadata. This facilitates the creation of analytic tools for complex data sets that can examine the relations between entities as well as the metadata attached to each entity and relation.

II.7) Algorithms Provided
The current distribution of JUNG includes implementations of a number of algorithms from graph theory, data mining, and social network analysis, such as routines for clustering, decomposition, optimization, random graph generation, statistical analysis, and calculation of network distances, flows, and importance measures (centrality, PageRank, HITS, etc.).

II.8) Snapshot of the Interface
JUNG is a library and not a GUI so this does not apply.

II.9) Development Platform
Java J2SE 1.4

II.10) Supported Operating Systems
OS independent (all that support Java)

II.5) Software Dependencies/Required Libraries

1. JDK 1.4 or newer
2. Apache Jakarta Commons Collections 3.0
3. Cern Colt Scientific Library 1.0.3
4. Xerces (for GraphML reading and writing)

II.5) Current License
BSD License

II.5) Number of Users/Downloads
As of the end of September 28th there have been 9,673 downloads, 121,107 page views, and north of 50 active users

II.5) Pros and Cons

• Customizability: Because JUNG, for the most part, provides clean interfaces for modeling, analysis and visualization, it easy to substitute different widgets, algorithms and data structures for most tasks. Furthermore because JUNG provides a mechanism for annotating graphs, entities, and relations with metadata, it easy for users to create analytic tools for complex data sets that can examine the relations between entities and the metadata attached to each entity and relation.
• Rich graph data structures: JUNG provides rich data structures for accomodating many different kinds of graphs one might want to use (e.g. directed vs. undirected graphs, graphs with parallel vs. non-parallel edges, etc.)
• Wealth of algorithms for data analysis: JUNG provides a wealth of algorithms developed in the fields of social network analysis, information visualization, knowledge discovery and data mining.
• OS agnostic
• Open-source

• Poorly written visualizatoin API: The latest release of JUNG has a poorly written visualization API that many users have found cumbersome. There are plans, however, to release a newly written viz API in the near future.
• Weak support for different file formats: JUNG currently only supports a few different file formats and most of those formats are not fully supported.
• Layout algorithms: JUNG currently only provides a couple of layout algorithms and none of them can be used to visualize more than around a thousand nodes at a time in real-time.
• No support for relational databases

II.5) Planned Work
There are plans to release a significantly new visualizatoin API which provides much richer support for user interactivity as well as richer data structures for supporting layout and rendering. There are also plans to integrate code submitted from some of our users as well as provide enhanced documentation.

Part III

Please describe your main interest in participating in the workshop

Determining the feasibility of combining efforts to create one common, shared IV infrastructure as opposed to 100s of underfunded or proprietary toolkits, platforms and frameworks. Scouring for ideas for a common data protocol for communication between plugins. Eliciting feedback about the IVC software architecture with regard to extensibility and ensuring that it is future-proof.

Please use no more than 4 pages, in this HTML format if possible.
Send the completed paper by Sept. 30, 2004 to katy@indiana.edu and Jean-Daniel.Fekete@inria.fr.