xml:href, xml:rel and xml:type

["Rushforth, Peter" <Peter.Rushforth@NRCan-RNCan.gc.ca>]

Hi there,

I posted this to an 'official' w3c list but didn't get any response.  That could mean
one of several things... ;-).  

I am interested in "XML on the web".  Some say it failed!  I beg to differ.  But, how could
the story improve, I wonder?  

I am interested in RESTful applications, and in thinking about the space, I have come
to think that XML, while wonderful for creating your own domain specific vocabulary,
also suffers from that very strength:  too many re-inventions of the same thing (because
its so easy to reinvent) leads to no standardization/interoperability at all.

One of the really really really important aspects of REST is that 
links, link relations and media types are central concepts.

In Atom, there is the successful atom:link element.  It is so successful, it is used
in other vocabularies via the atom namespace.  So that makes me wonder if 
XML would benefit from a _single_ simple link standard, inherited by all XML vocabularies
in the same manner as xml:base,  xml:lang . 

My question to the XML community is this:

What would be the value and cost of @xml:rel, @xml:href and @xml:type, 
where "xml:" is automagically mapped to the XML namespace [1], 
"rel" is a link relation [2] and "type" is a MIME media type [3] and xml:href is a 
URI.

xml:type would be 'advisory', in that it would not guarantee what the server would serve
at that URL, but could be used to negotiate the response format for that URI.  This
might also necessitate considering a "xml:hreflang".

Is it enough (would we need xml:hreflang, or more)?  Is it too much (security, impact on parsers, XSLT/XQuery/XForms/XOther processors)?  
Is the use of the xml: namespace impossible at this point?  

Thanks for your thoughts or links.

[1] http://www.w3.org/XML/1998/namespace
[2] http://tools.ietf.org/html/rfc5988#section-5.3
[3] http://tools.ietf.org/html/rfc2046

Cheers,
Peter Rushforth