Re: Putting context in RDF serialization

"Sean B. Palmer" wrote:

> >     context="contextUri1"
> >     context="contextUri2"
>
> You can't repeat attributes

Woops, forgot about that :(

> - maybe you mean:
>
>      context1="contextUri1"
>      context2="contextUri2"

Well that might work, we could then specify where the statements belonged in
a N dimensional space.  See [1] http://www.cyc.com/publications.html   But I
don't think think there will be enough agreement about such a contextual
space in the open environment of the SW to justify the added complexity.
Consequently I would opt for a single tag - perhaps "contextFor" - which
could then be defined in a single place in our schemas.  This means that we
would factor the parallel context specifications (if needed) into the context
node itself.  Like the following:

<?xml version="1.0"?>
  <RDF
    xmlns="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
    xmlns:s="http://description.org/schema/"
    contextFor="contextUri3"
   >
    <Description about="http://www.w3.org/Home/Lassila">
      <s:Creator>Ora Lassila</s:Creator>
    </Description>
  </RDF>

And the context node might look like:

[contextUri3]--subContextOf-->[contextUri1]
[contextUri3]--subContextOf-->[contextUri2]
[contextUri3]--contextFor-->[id1,http:..../Lassila, Creator, "Ora Lassila"]

> Hmmmm, it's a very good idea. In the example you gave, what would be to
> examples of the types of context the triples (O.K. quads) may take on?

Doug gives several examples in [1] .. but source, time, category, come easily
to mind.

> Also, if (or rathr, "as") the context URI's are RDF Schema, could we use
> classes/properties of the Schema (i.e. further pointing out the context),
> and if so how?

I think I would prefer doing it outside the schema as illustrated above.
What do you think?

Seth Russell

Received on Tuesday, 19 December 2000 14:38:38 UTC