When speaking of “index terms”, we may think of a back-of-the-book index in a print work, where there are categories of index terms, and each term points to a page where the term occurs, or where something associated with that term is discussed. To create such a back-of-the-book index, the passages on the pages need to be marked, for example by underlining a term, or by writing a term in the margin. Such a set-up can also work in a digital environment, using hyperlinks, with the difference that page numbers cannot be used; instead the exact location, or range, is given.
This system can also work in TEI, using <index> and other elements (see http://www.tei-c.org/release/doc/tei-p5-doc/en/html/CO.html#CONOIX). However, this set-up is designed to digitize existing back-of-the-book indexes, and is therefore less suitable for digitally-born publications like Brill MRWs. Moreover, it does not readily allow unique identification of index terms (esp. according to some external scheme, for example Library of Congress Subject Headings) or the insertion of canonical terms, for example the uninflected lemma or regularized version that needs to be included in a back-of-the-book index. For these reasons, Brill uses the <rs> element instead, which stands for referring string.
Referring strings are covered in the TEI Guidelines in Chapter 3.5 Names, Numbers, Dates, Abbreviations, and Addresses. As always, TEI offers a host of options. For reasons of simplicity and uniformity, Brill chooses a minimum number of generic elements based on concrete needs.
The <rs> element tags the phrase in the base text that needs to be indexed. Within this tag, idno@sortKey can be nested. This sets the sequence of index terms in a list, such as a back-of-the-book index or a browse list. (Not all index terms are necessarily alphabetical, or to be ordered alphabetically). Also, the <reg> element can be nested within <rs> to document the canonical term. The @type attribute indicates the category of index terms, for example “Index Nominum”, i.e. index of personal names.
Then the identifiers: @xml:id identifies an element. This allows, for example, to refer from a browse list of index terms to a phrase in the base text. (This attribute may be used in combination with <rs>, <idno>, and <reg>). The rs@key attribute identifies a term according to some external defined scheme. It makes sense to declare such a scheme in a <taxonomy> in <teiHeader>, and use #values that are equally defined in the <teiHeader>. Lastly, rs@ref identifies terms with a URI, for example a CITE URN.
… A <rs xml:id="brill00001" type="subjects" ref="http://id.loc.gov/authorities/subjects/sh85018111"><idno sortKey="burlesque"/><reg>Burlesque (Literature)</reg>burlesque</rs> is a literary, dramatic or musical work intended to cause laughter …