Nested <refsection> tags to not get treated the same
ways that <refect1> ... <refsect2> ... <refsect3> ...
with respect to the size of the title font decreasing,
this makes it hard to see the differences in section
levels in generated output.
This problem has been observed in XSL:FO, XHTML (I
assume HTML) and ManualPage outputs (I have submitted
via email a very simple patch to the ManualPage author
(lotr), unfortunately I got lost in the docbook xsl
attributes and could see how the FO code was derived,
so sorry no patch at this time)
I think that changing the xslt so that the following
are equivalent is a simple and workable solution, but
the authors may think this hack undesirable.
refsect1 == refentry/refsection
refsect2 == refentry/refsection/refsection
refsect3 == refentry/refsection/refsection/refsection
refsect4 == refsection
so the templates for refsect could add an "|" and
the second item in the above list - I have no idea how
efficient such deep matching is, but I figure there are
plenty of abstraction levels in the DocBook XSL already
so, this should not be too horrible. (I hope)
Better solutions more than welcome.