Jeff Beal Filenames are computed by templates in the "chunk-filename" and
"recursive-chunk-filename" modes located in 'chunk-common.xsl'. Most of the
work for determining the name of the file happens in
recursive-chunk-filename, which follows these precedence rules (for
"chunking" elements).
1. If there is a filename specified by a <?dbhtml?> processing instruction,
use it
2. If the element has an *explicit* id (i.e. it uses the @id attribute, not
the generate-id() function) and the use.id.as.filename parameter is
non-zero, construct a filename by appending the $html.ext parameter to the
element's ID
3. Otherwise, construct a hierarchical filename that (roughly) follows the
pattern bk??ch??s??.html where the ?? indicates the element's position
within the parent. For example, bk04ch12s03.html is the file generated by
Section 12.3 in the 4th book of a set. If you are building from a book
instead of a set, the bk?? is left off. Different elements will generate a
different prefix, <article/>, for example, inserts an 'ar??' into the
filename, <preface/> is 'pr', <part/> is 'pt'.
This method can certainly create long filenames. We build some of our
documentation using a section chunking level of 3 and we get files named
'bk03ch07s01s02s07.html'. If, however, you're only building a single book
and chunking down to top-level sections, your filenames would be
ch??s??.html. |