Metadata formats and SRU/SRW
Ross wants to use A9’s OpenSearch for metasearch, but he doesn’t want his client to have to handle four metadata standards. Bruce doesn’t want to maintain his own citation database at all, and proposes XSLT extensions to handle retrieving citations from remote servers. Why not insert remote services into the pipeline, that retrieve records in any format and pass them on in any other? A metadata-crosswalk proxy? YAZ Proxy does some of this for MARC already, and XSL lets us build easy-to-maintain tools.
《》
Dunno if you are familiar with the work that my colleagues are doing with schema transformation. Project page (is being updated): http://www.oclc.org/research/projects/mswitch/1_schematrans.htm Get in touch with Jean Godby if you would like some more info. http://www.oclc.org/research/staff/godby.htm See a demo of the repository of crosswalks at: http://www.oclc.org/research/researchworks/schematrans/default.htm Jean and Devon are in the process of overhauling their web presence to expose current work. Lorcan
I was also suggesting that RDF is an important tool to enable my goal of not having to maintain mmy own citation metadata. SRU could be important too, of course (SRW less so). What do the two share in common? Rich and flexible simplicity. With SRU, you have a rich query language united to underlying storage technology or format, and easy-to-implement RESTful approach. With RDF you have a standard data model and ability to seamlessly mix-and-match different vocabularies. I know it's not fashionable to say that RDF is simple, but I'm starting to think an RDF approach is much simpler than writing and maintaining three or four large and complex XML schemas in XSD.