Novice questions (repost)

[Mads Ipsen <mpi@osc.kiku.dk>]

Dear DocBook users,

This is my second attempt at posting these questions; I believe that
my first attempt ended up as a reply to a previous posting.

If you believe that this forum is out-of-scope for the questions I
raise, please let me know, and I'll try to find help elsewhere.

First of all, these are real novice questions, but I hope some of you
will have the time to provide some feedback to my questions:

1. Using the style sheet

     /usr/share/xml/docbook/stylesheet/nwalsh/xhtml/docbook.xsl

   DocBook generates html files where the whole string ass. with a
   figure title gets formatted in bold. I prefer: Figure + number in
   bold and the caption text in roman, eg.

     <b>Figure 1</b>: Blah. blah.

   Here's my approach for dealing with this from a user supplied style
   sheet:

  <xsl:template name="formal.object.heading">
    ...
    <p class="title">
      <xsl:choose>
        <xsl:when test="contains($title,'Figure')">
          <b>
            <xsl:value-of select="substring-before($title,'.')"/>
          </b>:
          <xsl:value-of select="substring-after($title,'.')"/>
        </xsl:when>
        <xsl:otherwise>
          <xsl:value-of select="$title"/>
        </xsl:otherwise>
      </xsl:choose>
    </p>
  </xsl:template>

  It works, but I am starting to choose a wrong track here?

  2. For the html representation of my documentation, I need to insert
     a graphical header and the beginning and end of each sect1 <div>
     to achive a rounded appearance of the block. Something along the
     lines:

       <div class="sect1" lang="en" xml:lang="en">
         <div id="contentUpper">
           <img src="images/main_center_header_660px.png" alt="" />
         </div>
         ...
         <div id="contentLower">
           <img src="images/main_center_footer_660px.png" alt="" />
         </div>
       </div>

     It took me 5 min. to write a Python script that uses regular
     expression for matching and inserting these items - but are there
     a cleaner style sheet based solution? I tried fiddling with

     <xsl:template match="sect1">

     but that completely messes up the chunking.

  3. At some time soon, I will have to start caring about print as
     well, i.e. PDF reps. Do you have any suggestions of the
     postprocessors that I should use (commercial or free). Its
     imperative that they can handle MathML.

Thanks for taking your time to read this. If you have any feedback,
please let me know.

Best, Mads