xml-dev - RE: RDF? TM? (was Re: Didier's lab report)

[Uche Ogbuji <uche.ogbuji@fourthought.com>]

> You now Uche what could be very useful? it is to have the possibility to
> inherit the capabilities of an rdf: description. For instance, just imagine
> for a moment that you want to encode a list of resources pertinent to a
> topic, get this topic to be potentially used as a link and as a RDF
> description. Said differently used as a link and as a frame. In that case, I
> may get an element to inherit from both the xlink and the RDF capabilities
> some thing like:
>
> <netfolder xlink:type="extended" xlink:title="my title"
> rdf:type="a_file_system_folder">
> <resource xlink:type="locator" rdf:type="a_resource"
> xlink:title="mydoc.xml">
> 	<author>Uche Ogbuji</author>
> 	<modified> dec 14 2000, 11:01AM</modified>
> 	<MIME-type>text/xml<MIME-type>
> 		... other properties .....
> </resource>
> <resource xlink:type="locator" rdf:type="a_resource"
> xlink:title="Anotherdoc.xml">
> 	<author>Didier PH Martin</author>
> 	<modified> dec 14 2000, 10:01AM</modified>
> 	<MIME-type>text/xml<MIME-type>
> 		... other properties .....
> </resource>
> </netfolder>
>
> What we would have now would be something that could potentially be
> processed by an xlink processor and something which can potentially be
> processed by an RDF processor. Note that the link now becomes a smart link
> containing meta data about the resource referred by it. In the Antique
> Greece I would have gone to the temple and ask the gods to get some good
> sense and help us poor xml developers to bring us some good and useful
> stuff. In the modern times I have to ask the same thing to the modern
> gods... the W3C consortium :-)) And you know what is even funnier? my
> outlook spell checker want to change the word W3C by WACO :-)))

I just hope you're not mistaking me for anyone in the W3C.  I'm just a
lowly developer as you are.

> Seriously, it would help tremendously if the rdf capabilities could be
> inherited like the linking capabilities are. This would turn links into
> smart links. So in the example above, an rdf interpreter would know that a
> <resource> element is an rdf description - would know that it is about the
> resource identified by xlink:href. This implies that an RDF processor has to
> be aware that any href with or without a namespace prefix is a reference to
> a resource (href seems to be a de facto way to refer to a resource in
> several SGML or XML based languages). Then finally, an RDF processor would
> know that the children are to be properties. Obviously the rule would be
> that only properties could be set as child of a description. Otherwise, we
> would have to use namespaces to differentiate the properties and the
> non-properties. This could work but wouldn't improve the document's
> readability.

I agree that some aspects of RDF's heavy reliance on namespaces can be a
weakness.  But I want to point out that although the solution does impair
readability, I think it's still a valid solution.

Besides, if you don't mind one level of additional processing, it's quite
easy to embed the rules about the non-differentiated properties into an
attendant stylesheet.

In fact, the product I work on, 4Suite Server, provides a mechanism for
establishing RDF metadata from arbitrary XML using rules encoded with
XPath.


-- 
Uche Ogbuji                               Principal Consultant
uche.ogbuji@fourthought.com               +1 303 583 9900 x 101
Fourthought, Inc.                         http://Fourthought.com
4735 East Walnut St, Ste. C, Boulder, CO 80301-2537, USA
Software-engineering, knowledge-management, XML, CORBA, Linux, Python