Many DLP projects involved controlled vocabularies. These may be compiled internally, borrowed in their entirety from an authority (LCSH, Getty, etc) or some combination of the two. The vocabulary services portion of the infrastructure project is to develop and maintain systems for managing these thesauri and providing consistent and useful machine access to the information.
OCLC Terminology Services
In 2008, OCLC Research implemented a series of SRU/W search targets from which vocabulary information could be gleaned. These services were used to implement a basic search-enhancement prototype.
Local Terminology Services
Local services are being developed to expose vocabularies we have rights to using SRU/W and the Zthes context set. Java libraries that can easily issue requests and parse responses from these services are already in the works and can be used to develop vocabulary-enhanced applications.
- If a query is expanded to include more than several hundreds of clauses (ie, "people" is expanded to the 985 narrower terms) the query takes too long to be performed. This is yet another argument for index-time term expansion.
- The usability issues associated with vocabularies that have multiple terms with the exact same spelling and representation have not been addressed.