Renditional defaults

rendition renditional default
tagsDecl rend rendition tagUsage id

Methods of setting renditional defaults, using the tagsDecl in the TEI header

In cases where the rendition of a particular textual feature is consistent, the rendition can be expressed as a default rather than being represented through the rend attribute on each individual encoded instance of that feature. For instance, if all personal names in the document are set in italics, this fact can be captured as part of the document’s metadata, rather than by encoding each name as persName rend="slant(italics)". Even if the rendition is not perfectly consistent, it may still make sense to encode a default; individual exceptions can be captured using the rend attribute, which will override the default rendition.

Default renditions are specified in the tagsDecl section of the teiHeader. The tagsDecl starts with a series of rendition elements which contain a rendition ladder as their content. Each of these rendition elements is given an id attribute.

The second part of the tagsDecl is a series of tagUsage elements. Each tagUsage element associates a specific element (e.g. persName) with a single rendition element that expresses its default rendition. The gi attribute of the tagUsage element gives the name of the element whose rendition is being defaulted, and the render attribute of tagUsage points to the id of the rendition element that describes the default rendition (as shown in the example).

Individual elements that do not conform to these defaults can override them using their own individual rend attribute. Any renditional information on a particular element will override the renditional information specified in the default.

Designing a set of renditional defaults which addresses the needs of the particular document is an important part of document analysis. It is not necessary that the renditional defaults apply to more than one document, and in fact this may be impossible if your documents have very different typography. However, if your documents are all quite similar then developing a set of defaults that apply to your entire collection would be advantageous.

In addition to document-level defaults which are expressed explicitly, it is also helpful to identify broader default assumptions that determine what does and does not need to be said about document rendition. For instance, for many projects it is unnecessary to say explicitly that paragraphs start on a new line: this is such a common way of formatting paragraphs that it does not need to be expressed as a default, let alone encoded using the rend attribute for individual paragraphs. (For some projects, there may be paragraph-like features whose boundaries are identified by a symbol rather than by a line break, and in these cases a default rendition might well be set explicitly.) By identifying these global assumptions about the presentation of your documents, you can help your encoders identify what aspects of rendition need to be expressed using the rend attribute, and which can be silently assumed. If you treat rendition as an important component of your encoding, we recommend that you document it in detail, including your global defaults, your typical document-level defaults, and the range of features that you have decided to encode using the rend attribute. This will help ensure consistency across documents and will make it much easier for you to write stylesheets that use this renditional information meaningfully.


Example 1.

The following set of defaults would specify that all name and emph elements in the document are in italics, and all mcr elements are in small capitals. The choice of values for the id attributes is arbitrary; they serve only to identify the rendition element so that it can be pointed to.

<rendition id="rend.italics">
<rendition id="rend.smcp">
<tagUsage gi="name" render="rend.italics">
<tagUsage gi="emph" render="rend.italics">
<tagUsage gi="mcr" render="rend.smcp">