A Space-Efficient Model for Sharing Personal Knowledge Objects in Peer Communities

In the future it is likely that peer communities will be routinely established for the purpose of sharing electronic resources and targeted information among groups of peers with common interests. This sharing will be possible by building on existing Internet technologies, such as web applications and services, and peer-to-peer networks. Today many peer-to-peer networks do not work well for sharing large heterogeneous collections among common home users. In this paper we present a model for organizing personal knowledge and a two-fold framework for improving methods of information sharing. Our framework is based on a) the creation of compact, portable, organizational objects through the ad-hoc structuring of knowledge on a personal computer, and b) the aggregation of these organizational objects into a visual catalog that can provide a peer community with a conceptual view of the information resources residing in the community. We call our model the BookQuest framework, and we show how it addresses several key information sharing issues, including space-efficient summarization, naming and locating resources in a peer community, and user interface and visualization issues. Our model introduces a variety of techniques for space-efficient representations, including use of generalized Bloom filters to support relevancy ranked keyword searches, and statistical sampling methods to support similarity comparisons between collections. We describe applications of our model to peer-based, inquiry orientated educational