Re: Content model for SVG

I think, in doubt, especially in SVGT1.2 author have to
move structured information anyway into the metadata
element and using the new RDFa attributes to indicate
the role or property represented by the structured 
metadata.

Because the structure of titles is typically simple 
(and authors should prefer the metadata with a role
like tooltip anyway to provide a tooltip information 
instead of (ab)using the title element for a tooltip 
and to reduce the risk to corrupt the title
with inproper content), the new possibilities with
systemLanguage look pretty useful - of course,
older viewers (what are currently mainly all viewers)
do not care about this new feature.
Therefore this is currently only a theoretical 
possibility - hopefully it can be used in the near
future.

There are more problems however with the desc
element in SVGT1.2, because it cannot contain
a structure anymore. Therefore authors have to
move such structured descriptions in the metadata 
element, providing a proper CURIE for role or property:

<metadata role="http://www.w3.org/TR/SVGTiny12/struct.html#DescElement" 
property="http://www.w3.org/TR/SVGTiny12/struct.html#DescElement">
<!-- structured description -->
</metadata> 

This looks funny within SVG, but there is no better way to
do it in SVGT1.2 ;o)
Because it is much more complex for viewers to
identify this as a structured description, the probability is
significantly lower, that this is available for a wide 
audience.

With all this the complexity of content increases, even more
if such constructions have to be used in SVG in 'HTML5',
having not even included such important RDFa features ...

Received on Tuesday, 24 March 2009 10:57:18 UTC