I've got a problem that I've been pondering over and I'm sure there's a
pattern for doing this.
I have a hierarchical structure that I need to break out, but things seem to
get confusing when mixed content models come into play.

<para>
      This is some above image text<sup>See footnote 1</sup>
   <image src="image1.gif" alternative="My Image"/>
       This is some text in between the image and table.
   <table>
       <tr>
           <td>...
       ...
   </table>
       Some more text to finish the paragraph off.
</para>

<p>This is some above image text <sup>See footnote 1</sup></p>
<img src="image1.gif alt="My Image"/>
<p>This is some text in between the image and table.</p>
<table>
   ...
</table>
<p>Some more text to finish the paragraph off.</p>

<xsl:template match="para">
   <p>
       <xsl:apply-templates />
   </p>
</xsl:template>

<xsl:template match="image">
   </p>
       <img>
           <xsl:attribute...
           ...
       </img>
   <p>
</xsl:template>