A new XML indexing, conversion and retrieval engine called XRS has been developed by Dongwook Shin, a visiting scholar at the US National Library of Medicine. It appears to have a few advantages over existing engines. XRS allows several retrieval options, including retrieving any elements, the weighting of elements, and a query building tool that can develop a structured query. It uses a BUS (bottom-up scheme) devised by Shin, which minimises the indexing overhead to between 20 per cent and 45 per cent. Another innovation is a Java component that converts XML to HTML, but without the need for an XML-enabled browser. If a search result is in XML, the query mediator servlets forces it to pass through the rendering component, or sends the result to the GUI. The GUI in XRS is programmed as a Java applet, so that a query can be formulated with the query-building tool which then communicates via the browser and so to a servlet on the server. Readers who have an aversion to XML, or IE come to that, may find this development useful for converting XML documents. ® Further information