Page 227 - Building Digital Libraries

CHAPTER 9

that are being developed today. And while many tools continue to support
and rely primarily on Z39.50, the execution and processing of the protocol
is disruptive, since these resources cannot be fully integrated into an XML-
based system. Moreover, while Z39.50 does include authentication control
mechanisms, they pale in comparison to the authentication methodology
currently being developed and deployed within today’s current-generation
web services. In response, the Z39.50 International Maintenance Agency
began the development of ZING, the next generation of Z39.50. ZING was
to be an XML-based, object-oriented form of Z39.50 that could be utilized
within today’s web services infrastructure. Originally developed as a web
services protocol, ZING was eventually expanded and renamed as SRU/SRW.
SRU (Search/Retrieval URL) and SRW (Search/Retrieval Web Service)
represent the next generation in Z39.50 development. It’s important to note
that SRU/W was not developed with the intent to supersede the Z39.50
protocol, but to provide a protocol that would give users and developers
easier access to bibliographic data. SRU/W continues to share a number of
concepts with Z39.50, most notably the EXPLAIN, SEARCH, and SCAN
services—as well as retaining data in results sets.

EXPLAIN

The SRU/W EXPLAIN protocol provides developers with an
avenue for querying a server for the types of metadata sche-
mas and query methods currently supported by the server.
Moreover, this option provides users with administrative
information about the server—providing contact information
and information relating to the current configuration of the
SRU/W instance. While this command set can be utilized by
developers attempting to design a service around an SRU/W
server, its greatest use is as an automatic discovery method for
machine-to-machine processing. Since SRU/W utilizes a form
of response structure, automated processes, much like the web
crawlers used by search engines, can be used to automatically
capture and utilize an SRU/W server’s configuration informa-
tion for current and future queries.

SCAN
The SCAN command is a holdover from Z39.50, providing
access to a quick hits list. SCAN searches return a minimal
amount of metadata (primarily title and creator), as well as
information regarding the number of items in the SCAN list.
Within Z39.50, the SCAN command provided a lightweight
mechanism for querying a Z39.50 server for fast results, forgo-
ing the more time- and research-intensive process of retrieving
actual datasets. However, within the SRU/W context, the SCAN
command’s value is lessened due to the inexpensive nature of
the request, coupled with the absence of a significant perfor-
mance penalty for retrieving an item’s full metadata.
212