[#issue-mt-ns] Four choices

So, I'd like to get substantive discussion (and if we can, close) the
#issue-mt-ns soon, so I've made what I think are the four obvious choices:

1) Any data served as application/xml and not as rdf/xml does not
license the triples therein. This makes a namespace document in RDF
served without rdf/xml not valid, but furthermore does not license any
hypothetical RDFa triples served in XHTML, since that would obviously be
served as application/xhtml+xml. Another note would be that this MIME
type business only applies to the Web, so that if we converted RDFa in
XHTML to RDF/XML via a (client-side) GRDDL, we would be licensed to
interpret the triples, since there would be no MIME-type per se for the
result of the GRDDL.

2) Any data served with the RDF namespace *any where* in the document
does license the triples therein. However, this does mean that XSLT
transforms hosted with the application/xslt+xml that contains RDF
triples does indeed license those triples, which seems a bit illogical.

3) This is Danny's suggestion [1]. That any document served with
application/xml does in fact license a single triple upon encountering

<dataview:namespaceTransformation
       rdf:resource="http:/example.org/trnsfrm.xsl"/>

in a namespace document. The triple it licenses is this:

<> dataview:namespaceTransformation <http:/example.org/trnsfrm.xsl> .

QUESTION: Which then GRDDL can run to get any RDF out of the namespace document and do a graph merge with any RDF from the GRDDL source document itself?


4) This is Sean Palmer's suggestion [2]. - that an application/xml
document is interpreted as given by the namespace of its root  (document
node). This would mean if the root node has the RDF namespace, as in the
RDF namespace document example, then we could interpret the triples, but
if the root node was not (say in XSLT with embedded RDF example) then we
could not interpret the triples. Furthermore,we could only license the
triples from RDFa in XHTML if either the MIME type or the namespace doc
of XHTML (or the  XHTML profile) allows us to.

Anyways, hope those clarify some of the options!

         cheers,
             harry

[1]
[2]

-- 
		-harry

Harry Halpin,  University of Edinburgh 
http://www.ibiblio.org/hhalpin 6B522426

Received on Tuesday, 14 November 2006 03:48:22 UTC