> This combination appears to have been missed by the FO stylesheet. But
> some clarification is in order.
>
> DocBook 5 supports two table models: CALS-markup and HTML-markup tables.
> When you use the HTML-markup model, its caption element replaces the title
> element, based on how W3C HTML tables are modeled. So you cannot use both
> title and caption with an HTML-markup table. That version of the caption
> element has a content model like title, allowing only text and inlines.
>
> If you use CALS-markup, you can put a title at the beginning and a caption
> at the end. That version of the caption element can contain para and other
> block elements. Currently the FO stylesheet does not format the caption at
> the end, which I believe is an oversight. The HTML stylesheet does.
>
> Thanks for the bug report. I just checked in an update for the FO
> stylesheet to process caption in a CALS-markup table. That fix will appear
> in the next release.
>
> If you want to add the fix to your customization layer, you would need to
> copy the template named "calsTable" from fo/table.xsl to your customization
> layer and add this line at the end of the template:
>
> <xsl:apply-templates select="d:caption"/>
>
The following also needs to be added:
<xsl:template match="doc:caption">
<fo:block xsl:use-attribute-sets="table.caption">
<xsl:attribute
name="keep-together.within-column">always</xsl:attribute>
<xsl:apply-templates/>
</fo:block>
</xsl:template>