We have a catch22 problem: We want to retrieve a potentially large number of records from a range of target repositories but without limiting the set of results displayed to the user. How do we filter the results without losing valuable records? Are there any software components that can take out duplicate records, apply best fit algorithms to the search results etc We obviously don't want to do anything crude like cutting off authors with surnames in the range m-z! Howard Noble Educational Interoperability Specialist Oxford University, Learning Technologies Group (LTG) 13 Banbury Rd, Oxford. OX2 6NN. England Tel. +44 (0)1865 273211 internal 73211 email: [log in to unmask]