RDDL proposal from Chris Wilper

Here is a proposal for a syntax for resource description, 
based on putting RPV assertions in the XHTML HEAD.

<h:html xmlns:h="http://www.w3.org/1999/xhtml">
  <h:head>
    <RPV xmlns="http://www.rdf.net/rpv/">
      <R>
        <PV p="http://www.rddl.org/#resource" 
            v="/schemas/L.rng"/>
        <PV p="http://www.rddl.org/#resource" 
            v="/style/L.css"/>
      </R>
      <R r="/schemas/L.rng"/>
        <PV p="http://www.rddl.org/#title" 
            v="RelaxNG Schema"/>
        <PV p="http://www.rddl.org/#nature" 
            v="http://relaxng.org/ns/structure/1.0"/>
        <PV p="http://www.rddl.org/#purpose" 
            v="http://www.rddl.org/purposes#validation"/>
      </R>
      <R r="/style/L.css"/>
        <PV p="http://www.rddl.org/#title" 
            v="CSS Stylesheet"/>
        <PV p="http://www.rddl.org/#nature" 
            v="http://www.isi.edu/in-notes/iana/assignments/media-types/text/css"/>
        <PV p="http://www.rddl.org/#purpose" 
            v="http://www.rddl.org/purposes#render"/>
      </R>
    <RPV>
  </h:head>
  <h:body>
    <h:h1>The L Resource: A Descriptive Representation.</h:h1>
    <h:p>Human-readable description.</h:p>
  </h:body>
</h:html>

Pros:
  p1) simple; easy to read and write[1]
  p2) extensible (add property assertions at will)
  p3) explicit about the fact that we're describing "related" resources.
  p4) useful in other contexts.

Cons:
  c1) not XHTML-valid[2].
  c2) Not based on a widely known(yet?) syntax.

Note:
    I think it's more appropriate for this kind of information 
    to be given as a "media type" descriptive representation
    rather than a "namespace" description.
    The desire to use namespace uris as the anchor for this
    type of information seems to be driven by the lack of
    a standard way to assert (especially in-line) the "media type"
    of an instance document.  The tension caused by
    this desire is illustrated well by Patrick Stickler at:
    http://lists.w3.org/Archives/Public/www-tag/2002Dec/0036.html


[1] See Tim's http://www.textuality.com/xml/RPV.html 
    Note that I made up an element to contain RPV elements for
    ease in processing.  I think this makes sense, as
    this element is useful as a container in a plain RPV 
    document as well.  Also, I'm not sure if the intent with 
    RPV is to support multiple-valued properties the way I did it 
    with http://www.rddl.org/#resource above.

[2] Creating a new variation on the XHTML format 
    (or even doing the "xhtml modularization" thing) 
    just to include this info doesn't make sense.
    I would rather see support for some kind of
    "resource-property-value" assertion syntax
    in the XHTML format itself, as a more expressive 
    META alternative.

___________________________________________
Chris Wilper                                                    
Cornell Digital Library Research Group
http://www.cs.cornell.edu/~cwilper/

Received on Tuesday, 3 December 2002 19:36:33 UTC