Two further points on RDF/A - ancestors and qnames/URIs

I wrote a document about how to implement RDF/A, various approaches not 
just XSLT - also how to merge RDF metadata into head and/or body of 
XHTML2 doc.

It is here:

http://lists.w3.org/Archives/Public/www-archive/2004Nov/att-0027/xhtml2.pdf

(sorry only in PDF)

I draw your attention to section 11, which raises two new? points, and 
reiterates old issues. The new stuff is:

[[
Another aspect of RDF/XML that some people find difficult is
to know when to use a qname and when to use a URI. This problem
is repeated in RDF/A, with predicates and datatypes expressed
by qnames, and subjects and objects with URIs. This might be a
mistake which could cause confusion.
A further problem was that adding structure to an XHTML document
by putting in container elements such as a <div> without
any RDF/A attributes may modify the meaning of RDF/A attributes
on the child elements. This could be avoided by changing the rules
for finding the subject for an RDF/A element without a about or
id to not just look at the parent node, but to search up the XML
tree until some ancestor node with an explicit object or subject is
found.
]]

On the last point, there is an example earlier (page 8 column 2) where 
adding a <table> to  better structure the document content means that 
the RDF/A markup stops working as naturally.

Jeremy



If just dipping into this version of this doc, note that it starts by 
presupposing my proposed simplifications - I am expecting to have to 
revise it at some point.

Jeremy

Received on Thursday, 18 November 2004 13:21:11 UTC