[xsl] Transforming milestone tags

[Dieter Köhler <d.k@xxxxxxxx>]

I have a source document which uses a hierarchical to markup the structure
of the text of a manuscript (<div> for the big divisions and <p> for the
paragraphs) and milestone tags for page breaks (<pb>) and line breaks
(<lb>), which may occur in virtually any place inside the hierarchy, for
example:

<doc>
  <pb n="1" />
  <div>
    <p>Line A
    <lb/>Line B
    <pb n="2" />
    <lb/>Line C
    </p>
    <p>Line D
    <lb/>Line E
    <lb/>Line F
    </p>
    <pb n="3" />
    <p>Line G
    <lb/>Line H
    <lb/>Line I
    </p>
  </div>
  <div>
    <p>Line J
    <lb/>Line K
    <lb/>Line L
    </p>
  </div>
</doc>

I would like to transform this document into a nested structure of <page>
and <line> tags and markup the textual divisions as milestones:

<doc>
  <page n="1">
    <newdiv/>
    <newp/>
    <line n="1.1">Line A</line>
    <line n="1.2">Line B</line>
  </page>
  <page n="2">
    <line n="2.1">Line C</line>
    <newp/>
    <line n="2.2">Line D</line>
    <line n="2.3">Line E</line>
    <line n="2.4">Line F</line>
  </page>
  <page n="3">
    <newp/>
    <line n="3.1">Line G</line>
    <line n="3.2">Line H</line>
    <line n="3.3">Line I</line>
    <newdiv/>
    <newp/>
    <line n="3.4">Line J</line>
    <line n="3.5">Line K</line>
    <line n="3.6">Line L</line>
  </page>
</doc>

What is the best strategy to do this? (My main problem is to get a
selection of nodes spanning between <pb> tags appearing on different levels
in the hierarchy.)


Dieter Kvhler