The sole purpose of the <body>
of a class-1 file is to contain an ordered, segmented transcription of a single
version of a single work from a scriptum. <body>
must take @xml:lang
, specifying the predominant language of the text. If a
change in language occurs in a descendant <div>
, ensure that its @xml:lang
also changes.
<body>
takes one or more
<div>
s, each of which govern
either other <div>
s, or text (or
TEI elements), but never both. TAN files adopt a non-mixed content model (see the section called “Mixed, non-mixed, and semi-mixed content”).
The term leaf div refers to those <div>
s that contain only text, and not
other <div>
s.
Within this treelike structure of <div>
s, the concatenation of @n
values, starting from the most rootward
<div>
, provides the
reference system used by class-2 files to refer to parts of TAN-T(EI) files. A given
<div>
may have more than one
reference, if its @n
or any
@n
it inherits has multiple
values. Every permutation is calculated, and they are treated as synonymous ways to
refer to that <div>
.
The rule of combinatorial inheritance also applies when @n
has as its value a range of numbers. For
example, if @n
has the value "1-3"
then it will match for 1, 2, and 3. Such ranges are important for translations, where
there might not be precise one-to-one correlation with the divisions in the original.
Applications that handle texts with one-to-many alignment mappings can used different
strategies to reconcile the differences. See tan:merge-expanded-docs()
for discussion.
In previous versions of TAN, there was a requirement that each leaf <div>
should have a unique reference.
That requirement has been relaxed, because there are cases where non-unique leaf
<div>
s are required.[16]In a TAN-T(EI) file, for any two <div>
s that share the same reference, it is not allowed that
one be a leaf <div>
and the other
not. To do otherwise would entail a mixed content model. It is also further assumed
that all <div>
s that share the
same reference are consecutive, constituent parts of the same <div>
. That is, any two <div>
s with the same reference are not
alternatives to each other, but are rather disjoint parts. For true alternatives, see
discussion above on using variant
in @type
.
[16] Some scripta are encoded such that leaf divs are broken up (see Bodëús's edition of Aristotle's Categories, at 2a35, 2b5, and 2b6b). And some translations must be encoded so that leaf divs interleave. Further, one TAN-T's leaf divs might easily become another TAN-T's non-leaf divs, and vice versa. The distinction between leaf and non-leaf div is arbitrary, so both types should be expected to adhere to the same kind of reference system rules.