Hi folks

I am making a set of roles of accessibility and DHTML to make content more accessible and adaptable ....

I am running up with some best practice issues.
Part of the background and (self imposed) requirements of this work are important - on the one hand (R1) we need it to be easily to parse for a first time RDF user agent.
Secondly (R2) We need it to be really easy for the author - who has never heard of RDF, to use XHTML roles as Qnames
Thirdly (R3) we need it to be easy to extend, for a user group or special interest group such as bloggers    or DAISY
(R4) Simple things simple and difficult things need to be supported..
 (R5) We want user agents to be able to process new extensions of the ontology that they have not yet heard of or created full support by easy deducing what it is close to and similar to (or derived from).
Lastly (R6) we want to make it a "good" (what ever that means)  rich hi quality ontology and way of describing roles of content   - so all the usual concerns in design fit hear..(possibly    ;)

A bit more ..
we need to include
I would also add
and then we need to do the same kind of thing for state (possibly the rest of the interactivity model...)
And now to the main...

Option 1:  I would prefer a taxonomy to be rich via rich inheritance structure and less properties. I would make it stand alone, and not to reliant of outside concepts and ontologies.
outside concepts instead  be referenced as related to DC: related to


I see that as good for R1, R2, R3 and R4 and who knows about R5

 example

<owl:Class rdf:ID="LinksPage">

                        <rdfs:subClassOf rdf:resource="#PageType"/>

            </owl:Class>


<owl:Class rdf:ID="SiteMap">

                        <rdfs:subClassOf rdf:resource="#LinksPage"/>

            </owl:Class>



Option 2: A few base classes and types, and rich in properties -so the author defines the characteristics of a specific  usage by setting property values - and  not by use of a predefined  instance that describes it

This makes the whole this the most flexible as two richness of ontologies

eg:

  <accs:widgit rdf:ID="radio">

       <dc:creator>W3C WAI PF Working Group</dc:creator>

       <dc:description xml:lang="en">A radio button</dc:description>

       <accs:hasState rdf:resource="&waistate;selectable"/>

       <accs:hasState rdf:resource="&waistate;selected"/>

       <accs:hasState rdf:resource="&waistate;checked"/>

       <accs:nameFromSubtree rdf:datatype="&xsd;boolean">true</accs:nameFromSubtree>

  </accs:widgeit>

Option 3
a taxonomy to be rich via rich inheritance structure and less properties. But - using concepts that other people have already defined... like option 1 but the inheritance comes from wordnet etc...


Option 4  -the  one that I have not thought of....


Your comments...

All the Best
Lisa Seeman