Reference Index services associate information with a canonical identifier. A canonical identifier is either a CTS URN or a Collection identifier. We could describe a reference index mathematically as a relation where information in some domain is mapped to a range of canonical identifiers.
Reference indexing encompasses mappings that we traditionaly call "indices" (such as an index of proper names, an index of geographic names, or a lemmatized index of a text that maps a lemma onto each word of a text), as well as other kinds of mappings: a commentary, for example, could be considered an application relying on an index mapping a potentially complex data structure (the comment) onto a passage of text; a schedule of assigned reading for a class could be considered an application based on an index mapping dates and instructions onto passages of texts. A printed reader aligning passages in the three synoptic gospels could be considered an application mapping parallel themes onto textual references.
An instance of Reference Index Services supports one or more
Reference Indexes.
The services' indexes are described in an IndexInventory
document that is returned by the
GetCapabilities request.
Information about each index
includes a unique identifier and Dublin Core metadata
about the index.
Each index is described by
an idxstructure element that defines
the type of the two parts of the index:
All Reference indexes have one or more record elements composed of
associated pairs of a ref element
from the range of the index, and
a value element from the domain of the index.
Reference Index Services support the following four requests:
GetCapabilities
QueryIndex
ref or value
DownloadIndex