Re: Instant RDF

Ora,


> Aaron,
>
> Aaron Swartz wrote:
> > Of course, the idea would be to write an XSLT add-on for
> > each XML format, being careful to generate valid and useful
> > RDF, thus preserving the semantics. This, IMO, is much more
> > useful than generating an RDF set from any old random XML.
>
> Hmm... OK, that's possible, but leads to an awful lot of XSLT. And I agree
> that "any old random XML" shouldn't automatically be RDF either.

Yes it does lead to alot of XSLT. And we need to make a clear distinction
between

a) "any old random XML" being automatically RDF
b) making it really easy to generate RDF from "any old random RDF"

personally I prefer:

c) making it really easy to "RDF tag" appropriate XML documents without
mandating too much about the structure of such documents (i.e let's assume
the document designers are otherwise intelligent).

>
> Personally I am not opposed to a new RDF syntax (the current looks a bit
> like it was "designed by committee" :-). But ultimately the syntax
shouldn't
> matter all that much since I am sure everyone is hoping that most of RDF
> will be both read and *written* by machines (not humans).
>

Firstly, many things that are otherwise easily human readable are also
easily machine readable.

Secondly what we are really talking about is whether RDF should be
completely distinct from other documents, or ought generaly documents
contain RDF metadata. If RDF is meant to be totally distinct, it doesn't
matter what the syntax is because only specialized RDF processors will be
processing it --- the argument has been made that XML itself, in this
scenario, might not be the best serialization format.

On the other hand if we wish to "RDF enable" general XML applications (and
lets include XHTML, ebXML, XSQL etc), then we ought carefully consider the
extent that RDF ought mandate a particular document structure. If this is
the case, as has been suggested by several people, then allowing RDF
shorthands such an "rdf:about" attribute in lieu of an rdf:Description
element, or an "rdf:instance", rdf:for etc. etc. may facilitate the adoption
of RDF itself.

As I see it, the role of XSLT would be to transform this RDF shorthand into
"real" RDF (in which case we might only need several XSLTs).

e.g. something like:

<xsl:template match="*[name() != 'rdf:Description' and @rdf:about]">
    <rdf:Description rdf:about="{@rdf:about}">
        <xsl:apply-templates/>
    </rdf:Description>
</xsl:template>

In this case the XSLT can serve as an unambigous mapping between the RDF
shorthand and the actual RDF.

Jonathan Borden
The Open Healthcare Group
http://www.openhealth.org

Received on Friday, 25 August 2000 10:22:50 UTC