Re: Trust, Context, Justification and Quintuples

Sandro Hawke wrote:

>>This would work by convention instead of by spec...
> 
> 
> How else would you read RFC 2396 (especially bis), RFC 2616, and
> WebArch, if not to say that the RDF triples serialized and transmitted
> in a "200 OK" response to an HTTP GET of some URI were the resource
> (or the current state of the resource) identified by the URI?
> 
> That is: if you get a "200 OK" with Content-Type: application/rdf+xml,
> you know the URI (according to its host) identifies a knowledge base,
> a time-varying collection of RDF triples, the current contents of
> which are exactly what was transmitted to you.  That's the only
> sensible reading I've been able to make of the relevant specs, so I
> think that means I'm operating according to spec.  I'd agree the specs
> aren't terribly clear on the subject, though.
> 

Well, considering all the discussion about whether an RDF URI that 
identifies an RDF resource should return anything, and if so, what... 
and the fact that some RDF URIs are likely not to be GETTable and when 
they are, may not return RDF... how are we to know when an RDF 
resource's URI is to be treated in the favored way and when it is not?

Since the answers to these issues is not spelled out in the RDF spec 
package, nor really in the other specs you mentioned, it will have to be 
settled by common usage based on what seems sensible (like your 
argument) - to me, that is by convention.

Cheers,

Tom P

Received on Friday, 19 December 2003 23:02:44 UTC