IMPETUS / HOW PROJECT BEGAN: Ever larger collections of documents and non-numerical data are becoming accessible to casual users via powerful search and retrieval algorithms. For instance, a user's query to a search engine can often result in hundreds of potentially relevant documents. We believe that interactive 3D visualization techniques, used correctly, can be a powerful medium in which large amounts of abstract information can be comprehensibly presented, just has scientific visualization has been used to present large numeric data sets. The goal is to give users 1) a seamless view of text document result sets, at both a general and detailed level and 2) a powerful set of operations through which the user can organize, filter, and inspect groups of documents.

Objective: Our initial goal is to demonstrate when three dimensional visualization can be a valuable interactive medium between the user and a retrieved set of documents. It should help the user perform tasks (such as finding those documents which are most relevant or understanding general properties of the retrieved set) more quickly than a traditional scrolled list. In particular, we want to give users 1) a seamless view of text document result sets, at both a general and detailed level and 2) a powerful set of operations through which the user can organize, filter, and inspect groups of documents. Longer term we are are exploring how to measure effectiveness of visualization to support the exploration of documents in collaborative settings with interactive devices such as large screen displays.

ARTIFACTS:

• NIRVE User Guide
• Project overview
• Snapshots of various models
• Software distribution

PUBLICATIONS: