Heterogeneous Data Management on Large Networks

We propose an information retrieval product for networked information systems using a new technique called "reference based retrieval" (RBR). The product is a smarter server/navigator for selecting the most appropriate objects for a user by using adaptive techniques using feedback from the users. By using the reactions of users, a wide feature space of matching can be consulted to achieve optimal results. This product is appropriate to the subtopic because it makes heterogeneous collections of data that are distributed over large networks manageable by giving users a powerful tool to find the best objects as defined by that user. Only with this type of tool will the large quantities of information in multiple formats be effectively ranked individually for a user. The product would extend the usefulness of Z39.50 standard and be part of the Wide Area Information Server (WAIS) architecture to allow maximum utilization of the tool.