The aim of the proposed research is to integrate flexible hypermedia infrastructures to automated capture environments in order to support long-term use. Previous work by the group at GATECH has aimed at making capture a ubiquitous service at any point in time and over a short period of time, The real use of the infrastructure has created the interesting problem of having too much information available, The objective of the proposed joint research is to find flexible and customizable information access solutions that scale over time, increasing the benefit of captured memories for the human as time passes and more information is captured. Our approach to this problem is to view capture as only one part of the information acquisition process in any given application. We introduce a spiral model for capture that stresses how the capture repository must be viewed as an ever-growing container for activities and knowledge that occurs before, during and after live capture sessions. As the repository grows over time, specialized automated services must be created that add structure by highlighting inter-relationships between the captured experiences and activities that occur outside of capture. Moreover, flexible services for structuring and accessing the information must be provided. Work that involved Abowd and Pimentel demonstrated how the digital media of the captured lecture can serve as an effective anchor for collaborative discussions and Web-based homework assignments - Abowd et al. (1999). The four main activities of the joint research plan are:

1. The definition of a spiral model to describe the research challenges of an automated capture service to support long-term activities,
2. A generic Infrastructure for Capture and Access - lnCA. lnCA provides not only the framework for a run-time environment to support the capture of many live activities, but also an extensible information model for captured experiences that will be used to link activities occurring during a live session to those that occur outside of the live session, as described by the spiral model. 3, The development of a generic and open infrastructure to Store, Extend, Retrieve, and Visualize Evolutionary information, or to SERVE captured information to humans. The SERVE infrastructure will support specialized services that occur outside of live capture and serve to augment the information contained in a capture repository and to generate automatic associations between related experiences.
3. The continued experimentation within the educational domain as well as the development of expertise in a new domain, distributed collaborative design meetings. These activities will be split between researchers in the lCMC-USP, Brazil, and researchers in the Future Computing Environments Group at GATECH, U.S,A, These institutions are particularly well positioned to do meaningful research on automated capture environments. The FCE Group, lead by Dr. Gregory Abowd, invented the Classroom 2000 system, the foremost demonstration of a large-scale automated capture environment. The USP Group, with extensive experience in the modeling and design of hypermedia systems, is lead by Dr. Maria da Graca Pimentel, who has worked closely with the FCE Group to extend the capabilities of the Classroom 2000 system over the past year. This collaboration has involved 8 different graduate students. 3 from USP, and will continue at least for the next 15 months, with Dr. Pimentel and another faculty member from USP, Renata PM. Fortes, working at Georgia Tech. This proposal, if funded, will continue to feed the momentum of this collaborative activity. in terms of the work proposed in this document and its companion proposal submitted by Dr. Abowd to the NSF in the US, the FCE Group will be responsible for the development of the lnCA framework for developing capture applications for different domains. The group at USP will be responsible for the development of the SERVE infrastructure and will actively participate in the use of the revised Classroom 2000 educational capture system at lCMC-USP. Both groups will collaborate in the development of a new educational capture environment and one to support distributed design meetings. It should be evident from this proposal and the one submitted by Dr. Abowd to NSF that this research could not be carried out at this scale without effective collaboration between these particular partners. The results of the research applied to the educational domain will benefit directly part of the teaching activities in both USP and GATECH, while the research applied to distributed collaborative design meetings could benefit any other software development teams.