Re: RCC Extension

Hi Alexander,

> As there’s an invitation of feedback, I’ll try to give one.

Thanks!

> We’re currently exploring the possibilities to integrate RCC into an 
> authoring tool environment.

That's good news.

> This is pretty hard as we must for example know which scripts to 
> execute and which not.
> But there’s one big main problem: We need to know the attributes of an 
> element. I thought
> on some metadata, something like:
>
>   <elementDef name="progressbar">
>      <model>
>         <property name=”x” type=”SVGCoord” description=”X-Coordinate”/>
>         <property name=”color” type=”SVGColor” …/>
>       </model>
>      <prototype>
>        …. Visual Appearance
>      </prototype>
>   </elementDef>

> I please you to carefully consider something like that. Without such a 
> runtime information,
> it’ll be unimpossible to implement RCC really good into Authoring 
> Tools. On the other side,
> this should be a requirement. So that when there’s no property 
> defined, you cannot use it
> by this.self.getAttribute(“”) otherwise most authors would leave out 
> the model part and this
> would result that a lot of RCC “components” would be incompatible to 
> authoring or re-use.
> I think this is also very useful for the re-use if components by 
> others.

This type of issue has already been raised within the working group.

One possibility is to require in your authoring tool that each 
extension is documented with an accompanying schema (XSD, RNG, you name 
it) which would handle not only definition and typing of attributes on 
an element, but also allow for defining more complex structures where 
attributes can be defined once and re-used, etc. I'll save you all the 
benefits from using a schema language. I believe that right now, the 
RCC syntax is not meant to serve as a schema and leveraging other 
existing technologies could complement the necessary framework you need 
on an authoring perspective.

However, there would be possibly very interesting new possibilities if 
there were an RCC mechanism to specify what certain attributes have as 
a datatype. One issue that has been raised is the poor RCC integration 
with animation as you would have to know the nature of the custom 
attribute's value to animate it. The working group is currently looking 
at how we can add a notion of attributes specification to custom 
elements so that this would be possible.

Stay tuned.

Antoine
-- 
Antoine Quint <aq@fuchsia-design.com>
W3C SVG Working Group Invited Expert
SVG Consulting, Teaching and Outsourcing
Fuchsia Design <http://www.fuchsia-design.com/>

Received on Tuesday, 11 November 2003 11:36:52 UTC