Position Paper by Thomas Baudel, ILOG

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?

Even though an InfoVis infrastructure should be centered around a particular rendering engine (data processing, view generation, view transform and interaction), this engine is only a minor aspect of the infrastructure, pretty much like the X windows rendering engine is only a small part of the Xwindows system.

An infrastructure should rather be centered around an introspectable notification network that connects together:

  1. the rendering engine
  2. the data model and input model
  3. interaction handling
  4. a collection of predefined actions and commands (undoable actions)
  5. predefined support editors and dialogs.
  6. predefined support widgets (range sliders...)
  7. document/project model.

I think a general InfoVis infrastructure should be built as an extension of an existing infrastructure, such as the Eclipse or Visual Studio development framework (or another suitable infrastructure), rather than be standalone.

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)?

Understanding what InfoViz is for.  I approve the idea to group efforts to reach common ground on a variety of issues which are already solved, albeit differently, in many environments. However, I think it is too early to think a unique visualization infrastrucure could meet the goals of the various participants. Most notably, I am interested by the integration of InfoVis features within existing application frameworks, to make the InfoVis component only a service/component part of larger application infrastructure.

Part II

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

II.1) Project Name and Web Address

ILOG Discovery, http://www2.ilog.com/preview/Discovery

II.2) Core Team Members

chief architect: Thomas Baudel, baudel@ilog.fr
developer, product integration: Bruno Haible, haible@ilog.fr
developer, data management: Robert Dupuy, dupuy@ilog.fr
consulting, graph layout: Georg Sander, sander@ilog.fr

+ additional staff members for QA, packaging and documentation.

II.3) Project Start Date

Oct. 2001

II.4) Targeted User Group

Product targeting is being defined.

II.5) Supported User Tasks

From a visual data analysis application, we are moving towards a view definition application and a software component to enable providing sophisticated data views inside any kind of (business?) application.

II.6) Major Features of the System Architecture

The engine follows more or less a formal, declarative visualization model which enables depicting a wide range of algorithms (class of "data linear" visualization) with a fixed and ordered number of parameters.

This enables providing many sorts of connected visualizations and refine the view parameters quite easily to litterally "browse the visualization space" rather than the data space.

II.7) Algorithms Provided

Many preset views including 2d graphs, various kinds of grids, treemaps, parallel coordinates, histograms... The data flow model can be extended to reach a wider design space.

II.8) Snapshot of the Interface

II.9) Development Platform

Java.

II.10) Supported Operating Systems

Windows, Linux, Solaris, MacOS X

II.5) Software Dependencies/Required Libraries

SUN JDK 1.3 or above.

II.5) Current License

Free, renewable 3 months licences. Beta program for developers wishing to evaluate the component rather than the application.

II.5) Number of Users/Downloads

650 downloads as of Aug. 2004. The component is used daily in 2 internal applications and 

II.5) Pros and Cons

Cons: still in development, and positioning is unsure

Pros: robust, documented and quite flexible.

II.5) Planned Work

test scenarios and examples of uses ; usability studies.

Part III

I am curious to know of other's plans and views on what InfoVis infrastructures should be like, and to this effect I am willing to share my own view and our current undertaking in this respect.

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.


Created by Jean-Daniel Fekete and Katy Börner on Thur Aug 12 11:15:27 2004