Renditional prefixes and postfixes

rendition prefix postfix
pre post rend

Use of the pre and post keywords to capture characters printed before or after an element (used as delimiters)

Prefixes and postfixes are characters (or strings of characters) which precede or follow a textual feature or element and are directly associated with the element’s presentation: for instance, quotation marks, or a square bracket preceding a turnunder in poetry, or a pair of parentheses enclosing a stage direction. Identifying a particular character or string of characters as a prefix or postfix, in effect, expresses the judgment that these characters serve a primarily renditional function in the document, rather than being constitutive parts of the document’s linguistic content. Clearly this very distinction is subject to criticism, and in many projects may simply be inappropriate or unsustainable. In all cases, the boundary between what is renditional and what is textually essential (if those categories are taken to be meaningful) is subject to interpretation and strategic consideration.

In practical terms, however, it is often very useful to encode prefixes and postfixes as renditional information using the rend attribute, rather than as character data that is transcribed directly into the document. Encoding them with rend provides some flexibility in the formatting of your output (for instance, to regularize the treatment of quotation marks or footnote numbering).

Prefixes and postfixes are encoded using the pre and post keywords within the rend attribute on the element that motivates their appearance. Common cases include stage directions, continued verse lines, divisions such as chapters which are separated by ruled lines or ornaments, and quotations. Title pages often include delimiters (such as ornaments or ruled lines) which are most sensibly encoded using pre and post.

In general, we recommend that any delimiters which indicate the boundary of an element (such as ornaments, rules, brackets, parentheses, quotation marks, etc.) should be encoded using pre and/or post. We do not recommend encoding delimiters that are part of the ordinary sentence punctuation in this way . Thus for instance we would not encode the parentheses in the following example using pre and post:

I hope (mon dieu!) that you are joking!

Note that because parentheses are part of the syntax of the rendition ladder itself, they need to be preceded by a backslash (as an escape character) when they appear as the value of pre or post:

<speaker rend="pre(\()post(\))">

Because the pre and post keywords are part of the markup of the document, it is not possible to apply further markup to the character strings that are given as their values. This imposes certain limitations: it is not possible, for instance, to provide any explicit information about the formatting of the delimiter itself, if it differs from the element it accompanies. In cases where this is a problem, it is better to transcribe the prefix or postfix directly as content rather than encoding it as renditional information.