The goal here is to utilize the new facilities embodied in the prototype Interspace environment to build new applications with revolutionary functionality. Unlike the above two efforts that are primarily performed by our team, this effort is primarily done via collaborations with other large projects specializing in the applications. Our efforts within the contract would be to complete the packages for application kits and work with external applications groups to build a complete application and run a large-scale trial of the Interspace in different applications. This is the technology transfer into the digital library community in several different aspects. The documents application will be done in collaboration with the $4M NSF/ DARPA/NASA Digital Library Initiative (DLI) project at the University of Illinois. This project is building software to search distributed repositories for thousands of documents in scientific journals on thousands of users in midwest universities.
The document applications kit will enable analysis across repositories to be supported, as opposed to merely search. Since the principals on the Illinois DLI are the same as those on this proposal, close collaboration is easy to achieve. Schatz is also PI on the DLI as well as this proposal.
The maps application will be done in collaboration with the DLI project at the University of California at Santa Barbara. This project is building software to search large collections of spatial data within maps of different types across the country. The maps applications kit will enable analysis across repositories to be supported, by comparing textures from one region to textures within another. We have a close collaboration with this project, as evidenced by a supplement to the UCSB DLI grant specifically for the collaboration on the texture images with Illinois and Arizona. UCSB researchers are enthusiastic about using our analysis software in their GIS applications on real users. Their corresponding PI is Terry Smith.
The software application will be done in collaboration with the Patterns Group at the University of Illinois, led by Ralph Johnson. This group is internationally known for their work in practical specifications for software engineering and for their uses of Smalltalk. For example, they host the annual Pattern Languages of Programming (PLoP) conference. We are already actively collaborating with them, in using patterns as the specification for the Interspace architecture, and our coPI Charles Herring is a member of both groups. There was a joint submission to the Software Engineering Group of DARPA ITO BAA 95-40 entitled "Pattern Languages for System Design and Evolution: A Case Study and Hyperprogram Prototype", with Johnson and Schatz as coPIs, which was selected as deserving of funding. The effort here is to investigate which units are appropriate to use concept spaces to help locate appropriate program fragments for particular applications.
The proposed concept space research has also been applied to other collaborative activities, especially in the context of synchronous electronic meeting facilitation [Chen, Hsu, Orwig, Hoopes, and Nunamaker 1992] [Chen 1994]. Using the Arizona GroupSystems electronic meeting system (EMS) as a research testbed, we adopted concept space extraction and categorization techniques to facilitate meeting convergence, a process often considered difficult for collaborative systems (most collaborative systems are suited for the divergence process) [Nunamaker , Dennis, Valacich, Vogel, and George 1991]. Our system was able to analyze and extract concepts in electronic brainstorming comments and summarize and present ranked topics in a graphical representation, all in less than 10 minutes on 66-MHz PCs [Orwig, Chen, and Nunamaker 1995] [Chen, Houston, Yen, and Nunamaker 1996@]. Based on techniques similar to those described for the proposed research, the algorithms were designed as a set of "intelligent agents" (helpers) for meeting facilitation.
A BAA96-06 proposal (Collaboration Infrastructure track) entitled "Electronic Meeting System Support for DOD Real-Time Decision Making in Crisis Management and Planning" has also been prepared jointly by Dr. J. F. Nunamaker (University of Arizona, Center for the Management of Information) and Dr. Hsinchun Chen (University of Arizona, AI Group). This EMS research plans to create dynamic, concept-based "intelligent agents" for information importation and filtering during electronic meetings and to provide real-time information to support time-critical group decision making in crisis management and planning. We believe the proposed concept protocols research will be synergistic with this EMS research (with Chen as the key technical lead for algorithms in both projects). Our studies should significantly enhance the development of effective and "intelligent" collaborative infrastructures, enabling us to produce practical results from our research for both asynchronous and synchronous collaboration.
4.
DESCRIPTION OF RESULTS,
PRODUCTS, TRANSFERABLE TECHNOLOGY,
AND TECHNOLOGY TRANSFER PATH
5. COLLABORATIONS
7. TECHNICAL RATIONALE, APPROACH, AND PLAN
8. COMPARISONS TO OTHER RESEARCH
