Page 123 - JOURNAL OF LIBRARY SCIENCE IN CHINA 2018 Vol. 42

122 Journal of Library Science in China, Vol. 8, 2016

was developed by OCLC based on 10 controlled vocabularies including LCSH during 2004 to
①b
①a
2008 (Vizine-Goetz, 2008); 3) UMLS terminology services: a terminology service system which
was constructed by the U.S. National Library of Medicine based on UMLS and other terminology
①c
resources in 2012 (“UMLS Terminology Services”, 2012); 4) Open Metadata Registry (OMR): a
large-scale vocabulary and metadata registry which was constructed by the U.S. Natural Science
Digital Library (NSDL) project in 2005 (Phipps & Hillmann,2011). In addition to these practical
systems, some research projects also involved the study and development of terminology services.
In the HILT (High-level Thesaurus) project, which is a series of collaborated projects between the
University of Strathclyde and the University of Edinburgh during 2007 to 2009, seven terminology
services were implemented for term retrieval using the SOAP and SRU/SRW protocol in the
①d
Phase IV (Nicholson, McCulloch, & Joseph, 2009) . In the STAR (Semantic Technologies for
Archaeological Resources) project, which is a collaborated project between the University of Glam
organ, English Heritage and the Royal School of Library and Information Science, Denmark during
2007 to 2010, seven terminology services were implemented based on the thesauri and glossaries
about English heritage (Binding & Tudhope, 2010). In the ADL (Alexandria Digital Library)
project, two protocols were developed respectively until 2009 for accessing distributed gazetteers
and thesauri: one is the ADL thesaurus protocol which provides three SOAP-based terminology
services for querying the attributes and place names of gazetteers (Janee & Hill Linda, 2015), and
the other is the ADL gazetteer protocol which provides five SOAP-based terminology services for
querying and browsing the content of thesauri (Janee, Ikeda, & Hill Linda, 2015).
The earlier terminology services, such as OCLC, ADL, HILT and UMLS, usually used the
traditional XML or relational database format to represent and store vocabulary data. With the
emergence and development of the semantic representation language SKOS, the later terminology
services, such as AGROVOC, OMR and STAR, usually used SKOS/RDF as the representation
format of vocabularies, RDF triple stores to store vocabulary data, and SPARQL queries to access
RDF data. Even the systems that did not use SKOS as an underlying format, such as OCLC and
HILT, also can output the retrieval results with the SKOS format.
Except for the UMLS terminology services which applied both the SOAP and REST
architecture, the earlier terminology services, such as HILT, STAR and ADL, mainly applied the
SOAP architecture, whereas the later services, such as OCLC and OMR, usually applied the REST
architecture. This is consistent with the recent trend that the REST architecture is more and more
applied in Web services. In addition, OCLC and HILT used the SRU/SRW protocol, among which
SRW is SOAP-based and SRU is REST-based, to support interoperability among heterogeneous

①a OCLC refers to Online Computer Library Center, a non-profit organization providing WorldCat, the largest online public access
catalog (OPAC) in the world.
①b LCSH refers to Library of Congress Subject Heading.
①c UMLS refers to Unified Medical Language System.
①d SRU/SRW: SRU (Search/Retrieve via URL) and SRW (Search/Retrieval Web service) are a pair of web-based information
retrieval protocols, which use the architecture of Web services to realize some basic functions of Z39.50.