Networked PIM using PDMS

Personal information management (Pim) is a promising new type of application allowing not only to search a desktop, but to pose complex, structured queries against the data on ones computer. We propose to remove the confines of Pim and make selected data available to a network of peers using peer data management system (Pdms) technology. The result is an application for collaborative information management within workgroups. To achieve this vision, several participating tools and technologies must be adapted: Pim systems must be augmented with privacy concepts to protect non-public data, which in turn must be interpreted by the Pdms query rewriting mechanisms. The entity resolution methods of individual Pim systems must be extended across multiple Pim systems with possibly heterogeneous schemata and must support ad-hoc queries. Finally, traditional Pdms are designed to find the complete and correct query result. In a networked Pim application it is not necessary to find all results to a query and it is acceptable to respond with inexact results. 1 Personal information management across peers The range of applications to save and manage personal data is wide and ranges from email, address book, calendar, and to-do-lists up to presentations and publications. Additionally, there is a variety of platforms to perform these applications, ranging from the classical personal computer, to PDAs, and to webspaces with personal blogs. This huge number of applications and platforms leads to a heterogeneity and fragmentation of data digitally available. There are several approaches to sum up and query this data in so called personal information management systems (Pims). A first approach is to enable full text search within the entire amount of data (emails, files, bookmarks, etc.) as supplied for instance by Google Desktop Search or Apple Spotlight. More complex queries are made possible by system