Page 231 - Building Digital Libraries
P. 231

CHAPTER 9



                                          <schemaInfo>
                                          <schema identifier=“info:srw/schema/1/marcxml-v1.1” sort=“false”
                                          name=“marcxml”>
                                          <title>MARCXML</title>
                                          </schema>
                                          <schema identifier=“info:srw/schema/1/dc-v1.1” sort=“false” name=“dc”>
                                          <title>Dublin Core</title>
                                          </schema>
                                          <schema identifier=“info:srw/schema/1/mods-v3.5” sort=“false”
                                          name=“mods”>
                                          <title>MODS v3.5</title>
                                          </schema>
                                          </schemaInfo>



                                                   Given all the benefits associated with SRU/W, it has been puzzling why so
                                                   few applications, tool sets, or organizations currently support the protocol.
                                                   By contrast, after nearly fifteen years, the Z39.50 protocol, despite its cryp-
                                                   tic communications format and spotty character-encoding support, still
                                                   remains the predominant method for searching and retrieving bibliographic
                                                   data in the library environment. This isn’t to say that a handful of large
                                                   SRU/W implementations doesn’t exist. The Library of Congress provides a
                                                   robust SRU/W API for its library catalog, as do projects like the European
                                                   Digital Library (www.theeuropeanlibrary.org). Additionally, vendors like Ex
                                                   Libris encourage users to utilize their built-in SRU service over the tradi-
                                                   tional Z39.50 connection (though both exist). However, after being available
                                                   and in production for nearly fifteen years, the tools or software to support
                                                   the SRU/W standard remain few and far between. What’s more, given the
                                                   inconsistent implementation of SRU services and the trend among digital
                                                   library developers to create more RESTful-based services, it appears that
                                                   the desire to provide long-term support and maintenance for existing tools
                                                   capable of working with the SRU/W standard is very much on the wane.




                                                   OpenSearch
                                                   OpenSearch is an intriguing protocol when one considers how quickly it has
                                                   been adopted by developers outside the library community. OpenSearch was
                                                   developed in early 2004 by Amazon.com to allow resources to be integrated
                                                   into its A9 search engine. However, the simplicity of the protocol’s search
                                                   and response syntax has led to a very quick adoption within the corporate
                                                   community. So ubiquitous has OpenSearch become that all major web
                                                   browsers use it to configure new search endpoints in the browser.
                                                      As a protocol, OpenSearch functions very much like SRU/W in that it
                                                   works over an HTTP GET connection utilizing a very simple URL query
                                                   structure. What makes OpenSearch different is the response format. Open-
                                                   Search utilizes RSS (Really Simple Syndication) and the Atom Syndication
                                                   format as the response format. By utilizing these two formats, OpenSearch
                                                   has been able to successfully leverage tools that have traditionally been used


            216
   226   227   228   229   230   231   232   233   234   235   236