I'm using role attributes and the docbook stylesheets (successfully) to use one image format when rendering to FO, and other when rendering to HTML (i.e. epub).
<mediaobject>
<imageobject role="fo">
<imagedata fileref="images/DIT-local.pdf"/>
</imageobject>
<imageobject role="html">
<imagedata fileref="images/DIT-local.png"/>
</imageobject>
However I have a couple of images using callouts, and using roles like that doesn't seem to work:
<screenshot>
<mediaobject>
<imageobjectco role="fo">
<areaspec>
</areaspec>
<imageobject>
<imagedata fileref="images/MVC-DIT-structuresvg" scale="50"/>
</imageobject>
</imageobjectco>
<imageobject role="html">
<imagedata fileref="images/MVC-DIT-structure.png" scale="50"/>
</imageobject>
</mediaobject>
</screenshot>
The SVG is just a wrapper that references the similarly named PNG and then adds some callout bugs.
Rendering that to FO results in two images side-by-side. Not sure if that's a bug, or expected behaviour. It isn't what I want though.
However if I change the second imageobject to an imageobjectco, which seems the logical structure, the is problematic - I have a single calloutlist that needs to refer to the "same" callout in either image. And I can't have two s with? have the same xml:id :-(
I hate image callouts! :-)
Is there a more correct way to organize my two alternate images?
Cheers,
Chris