Basic document structure
The basic document structure recommended by EpiDoc follows exactly that laid out for
TEI.2. There are two main sections of the document, both of which are required: the
teiHeader and the text. The teiHeader is
essentially the title page and descriptive catalog record for the electronic version of
the text, providing information about the source of the text, the way it is digitally
represented and any revisions it has undergone. The text element, on the
other hand, embodies the structure and content of the text itself, including
introduction, textual edition, translation, critical apparatus, commentary, notes and
the like. The basic, top-level structure of any such TEI document would look like that
sketched in [Example 1](example1).
This simple representation of the top-level tagging hides the structural complexity and
semantic power that can be achieved with TEI. Within each of the two major elements, TEI
permits (and sometimes requires) the use of a wide range of other elements. Some of
these elements have very precise rules governing their use. Others permit adaptation to
meet the needs of particular projects or types of textual materials.
The
text element and its subordinate
divs
When compared with the teiHeader, the top-level text element has a more
flexible content model, as do many of its subordinate elements. This flexibility is
necessary to accomodate the wide range of textual materials that might require digitization. [Example 3](example3) provides an abbreviated EpiDoc representation of a boundary marker for an imperial estate in North Africa.
Within the text element of [Example 3](example3), one finds the mandatory TEI body
element. TEI provides that this element can be subdivided further into any number of logical
divisions using the div element. TEI specifies the use of the div
element for the hierarchical grouping of divisions and subdivisions within a text. Depending
on the genre of the text, these 'groups' can have different common names, such as 'chapter'
in a novel, 'scene' in a play, or 'book' in an ancient literary work. Rather than elaborate
differently named elements for each and every possible such grouping, TEI employs the
generic div element for all such divisions. The TEI specification for the
div element provides a standard attribute named type. This
attribute is used to categorize these div elements independently of their
hierarchical relationships using a set of attribute values appropriate to the genre and
audience of the text. In this way, the nesting of div elements can be used to
indicate the hierarchical relationship between a chapter and a subsection, while the
type attribute can be used simultaneously to indicate the semantic difference
between a chapter and an appendix. Note: [Example 3](example3) does not 'nest' its div
elements, so there are not any significant hierarchical relationships between divisions of
this particular text. But there are significant semantic differences, as indicated by the
use of the type attribute.
In this example, the type attribute of the div element is used to
distinguish between six different types of divisions:
a description of the text and/or monument;
the edition of the epigraphic text itself;
a translation of that text;
a scholarly commentary related to the text and its unique problems and items of interest;
a history of the discovery, documentation, and interpretation of the text leading to its
present publication; and
a bibliography relevent to this text.
These attribute values (and two more: metadata and diplomatic) have been introduced as part
of the EpiDoc standard (see Section 5.5: EpiDoc divisions (tei:div) of the text element
(tei:text) for the formal definition of the div/ element and its attributes). TEI
places no limits or requirements on the use of the type attribute; it merely
makes type available "to provide a name or description for the division." We have
deliberately kept the number of EpiDoc values for type small to ensure the
greatest possible degree of smooth interchange between EpiDoc-conformant projects.
EpiDoc exploits the flexibility and specificity of the div/ element and its
type attribute to distinguish between the various components commonly found
in the scholarly presentation of epigraphic texts. Compare the conventional print
representation of ALA 2 (Example 4) with the XML version in Example 3. Although styles and
presentation differ from one scholarly epigraphic publication to the next, at a basic level,
this example is representative of the historiographic genre. Typographic cues are generally
used to alert the reader to a change in the character of information being presented for
each text, but the exact nature of such changes is often simply implied. We assume that our
reader understands the conventions of epigraphic publication. In the XML counterpart, the
semantic differences between each division are made explicit using standard values for the
type attribute.
These semantic distinctions, encoded into a digital text from its creation, facilitate a
number of important possibilities when the time comes to prepare the text for presentation
to a user or incorporate it into another publication or project. Standard XML software can
use combinations of element and attribute values to select, compare, combine and reformat
semantically discrete portions of documents in complex ways, providing the semantic
distinctions have been explicitly encoded in a standard manner. The typed divisions of TEI,
when categorized according to a generally-agreed and epigraphically relevant scheme,
constitute such a standard.
Divisions with the same type attribute value can be repeated as necessary, and further
semantic distinctions between them drawn at the editor's discretion using the standard TEI
attribute n. These arrangements permit extensive, project-specific flexibility in
searching, formatting and reusing documents. There are no EpiDoc-mandated values for n, but
projects are encouraged to develop standard usage and publish documentation thereof as part
of their dataset. The EPAPP project has developed such a system of attribute values, as
demonstrated by the text in Example 5. The EPAPP system of values for n on the
div element has been adopted by these guidelines as a recommendation for projects
which find it helpful.