[xsl] Building complex, hierarchical html datasets

[Mark <charltonrainbird@xxxxxxxxxxxxxx>]

Hi XSL-List

I'm using Saxon 9 to process xslt 2 stylesheets from the command line.

I use xslt to transform either large single xml files or sets of xml
files using document().

These are all complex, hierarchical publications (usually with a
printed equivalent) delivered as html datasets with subcontents pages,
breadcrumb links and extensive xrefs.

I break data down into manageable-sized html pages, using logical
break-points in the structure eg. chapter 3 breaks into 3.1, 3.2 etc.
3.1 breaks into 3.1.1, 3.1.2, 3.1.3 etc. xsl:for-each-group is
invaluable here.

However, some of the dtds allow quite loose use of elements that make
the structure far harder to predict/analyse. eg. <law> elements may
sit at the top of the hierarchy, at the second level or third level.
They in turn may contain further <law> elements or <law-level>
elements. Alternatively they may sit within a <group> element or the
<law> elements may contain <group> elements. In turn <group> elements
may sit at the top of the hierarchy and contain, amongst many others,
<law> elements.

Therefore structure is very difficult to predict and producing a
routine for every variation that does (or could) exist would be very
arduous and probably unreliable. What I am after is not a specific
solution but any pointers anyone can give or locations of information
elsewhere on the type of approaches that could be used for this type
of work. Obviously I've tried searching for this myself but have been
unable to turn up much so far that covers my specific situation.

Regards

Mark