* Danny Ayers <[EMAIL PROTECTED]> [2004-11-11 11:56+0100]
> 
> If we can have a http: scheme URI for the base of link rel, then why
> not for the namespace?
> 
> Personally I would prefer a http: URI for the reasons Dan mentioned.
> The use of these is widespread - every single namespace URI I have
> seen in the context of syndication follows this pattern. Is current
> practice compelling at all?

On the extensibility front, there has been occasional mention here of 
possibility for transforming (by XSLT, or even XQuery) from Atom's
format into RDF/XML. W3C has a Note on one technique for doing so,
called GRDDL, which makes use of metadata about namespaces to 
figure out which XSLT-defined mappings are appropriate to get RDF/XML 
from some instance data. 

http://www.w3.org/TR/2004/NOTE-grddl-20040413/
http://www.w3.org/2003/g/data-view

esp. http://www.w3.org/TR/2004/NOTE-grddl-20040413/#ns-bind

I believe there's an implementation that consults namespace docs for 
this info 'while you wait', though to be fair I couldn't find it 
just now.

So the motivating scenario is that the AtomPub WG could make an
atom2rdf.xsl available and cite it from the Atom format's namespace
document in a machine-friendly way. This only becomes compelling if 
RDF toolkits start to acquire the ability to get their RDF via XSLT 
transforms, rather than just parsing RDF/XML notation. I think this 
is starting to happen, eg.
http://people.w3.org/~dom/archives/2004/11/grddl-support-in-rap/

Most of the attention around GRDDL to date has been w.r.t. the 
XHTML-oriented version, but it seems to have wider potential.

Dan

Reply via email to