OASIS Darwin Information Typing Architecture (DITA) TC

The following is intended for the introductory topic for the Index markup section:

----------------

For the purposes of the DITA index markup, an "index" is a mapping from index terms to locations in DITA content.

Index terms are organized into hierarchical collections of index terms, where sibling index terms are unique within the scope of their common parent. The precise rules for determining that two <indexterm> elements define the same index term are necessarily implementation dependent. For example, case and whitespace may or may not be significant based on the editorial rules for the publications that use the index, or the language of the index may require the use of <sort-as> to determine the text used to compare index terms (i.e., for Japanese index terms that include kanji characters). Index terms may be associated with other index terms by "see" or "see also" relationships.

Locations are established by the source locations of <indexterm> elements, either within map or topic prologs or within topic short description, abstract, or body content. An index term may be associated with zero or more locations. An index term that has a "see" relationship to another index term is not directly associated with any location.

An index-term-to-location mapping may be used or rendered in many different ways, including as traditional "back of the book" indexes in paged or web-based publications and as search indexes in online delivery systems.

The published representation of locations is entirely dependent on the nature of the publication. in a paged rendition it could be page numbers or it could be paragraph or clause numbers if the publication uses detailed numbering, such as a standard, regulation, or codified law, In a typical HTML rendition, the locations could be rendered as direct links to the locations or, if the publication is also paginated, the HTML could reflect the page numbers from the paged rendition, or it could be paragraph or clause numbers if the publication uses detailed numbering.

------------------

The intent of this language is to make it clear that:

1. "index" in this context is not the computer science concept of "index" as for databases.

2. That the source markup is defining an abstract term-to-location mapping that could be used in many different ways, only one of which is back-of-the-book indexes.

3. That "location" rendering is highly implementation dependent.

Cheers,

E.

LinkedIn | Twitter | YouTube | Facebook