3 XML Binding Language for SVG

Hi-

I have some comments, questions, and suggestions for sXBL.

In 3.4, it says, "Thus, when the custom element is within a rendering
context (versus non-rendering contexts, such as described under sXBL
bindings for SVG resources and sXBL bindings for visual effects), then the
custom element behaves as if it were a g element with no attributes and the
nodes on xblChildNodes were the children of the g."

I'm concerned that semantic data (or if you prefer, data that might be rad
and interpreted by a DOM-reading AT UA, for example) might be lost if the
wording for treating the custom element as a <g> with no attributes or
children is not revised. Implementors may interpret that to mean that the
rich semantics of the original custom element are not available in the DOM.
It seems pretty clear that the intent is to outline how the visual rendering
is handled, but I want to make sure that implementers expose the custom
elements.

Is there any dispensation for reusing script from one element to another?
That is, if I write complex drag-and-drop code that can apply to multiple
'xbl:definition' elements, can I reference it only once, rather than repeat
it each definition? Note that I don't mean referencing script from the bound
document, but rather from the binding document.

Finally, my recent discussion with Ian on semantics, on and off-list, has
given me a thought. I would like to see title, description, and hint
elements on sXBL definitions themselves, not only on their content. It's  my
view that this metadata, if accessed through the proper UA, could give the
user a short and a detailed description of the custom element that is being
bound, including information about its namespace and intent, and
instructions on how to use it. Since this would reside in the
'xbl:definition' element, it would not have to repeated in the SVG content,
and could be very different to the content's own title, desc, and hint. This
may help to bootstrap the adoption of a wide variety of novel XML dialects,
since the user (or author who is reusing 3rd-party sXBL components) could
discover the nature of the dialect's underlying domain. Obviously, the UA
would have to allow access to this metadata in order for this to be useful.

Regards-
-Doug 

Received on Thursday, 25 November 2004 15:00:36 UTC