Re: Semantic content negotiation (was Re: expectations of vocabulary)

On 27/07/06, Xiaoshu Wang <wangxiao@musc.edu> wrote:

> > The webserver serving the resource <http://gmuer.ch/>
> > delivers two different XML documents depending if the Accept
> > header is set to "application/atom+xml" or to
> > "application/rss+xml". XML clients get what they want,
> > whether they prefer RSS or Atom.
>
> Again, they are two independent documents representing the same resource.
> They are not two parts residing in one document, with part-1 being in atom
> and part-2 in RSS.

I can't help feeling that the notion of "document" is causing some
problems here. Surely the URI indicates a resource, not (necessarily)
a document. The representation returned is a document. Similarly the
resource isn't (necessarily) an RDF graph (itself already different
from a document containing a serialisation of this graph.

I can choose what my URI denotes.

If my URI denotes a file, then I can't do anything with
language-related headers. In effect the URI indicates a series of
bytes.

If my URI denotes an abstract "document", and translations of this
exist, I can negotiate about which translation I want.

If my URI denotes an RDF graph, then I can get a serialisation of that
graph. I can negotiate about the syntax of the serialisation.

If my URI indicates something else, then why can't I specify all
manner of properties which I desire of the returned *representation*?

Cheers,
Hamish

-- 
Hamish Harvey
Research Associate, School of Civil Engineering and Geosciences,
Newcastle University

Received on Thursday, 27 July 2006 15:21:26 UTC