RE: httpRange-14: Use Case for RDF

There are at least 4 best practices questions around the TAG's
httpRange-14 issue
(http://lists.w3.org/Archives/Public/www-tag/2005Jun/0039.html ):

1. How should http URIs be minted?  Two styles of http URIs could be
used: Hash URIs and 303-redirect URIs. What are the implications, pros
and cons of these approaches?

2. How should servers respond to HTTP GETs?  Content negotiation, 303
redirects.  

3. How should clients interpret HTTP responses?  What can a client
conclude from the response.

4. What kind of document should be returned?

David Wood's message below (and at
http://www.w3.org/2002/02/mid/6F9F6968-CB73-427A-8682-AF9AB0F2E9C2@softw
arememetics.com;list=public-swbp-wg ) addresses questions #2 and #3
above.  My comments on his message are below; I'll respond separately to
#1.

> -----Original Message-----
> From: public-swbp-wg-request@w3.org 
> Sent: Monday, September 05, 2005 4:54 PM
> To: public-swbp-wg@w3.org
> Subject: httpRange-14: Use Case for RDF
> 
> 
> Hi all,
> 
> This message suggests a use case for SemWeb URIs to be resolved via  
> 303 responses.  David Booth, Ralph Swick and I contributed.
> 
> This message is in response to the following action item:  "Ralph,  
> DavidW, and DavidB to an initial draft of TAG httpRange-14 
> resolution  
> impact on semweb application developers" [1].
> 
> Use Case:
> 
> A SemWeb application includes RDF data.  The application allows each  
> URI in the RDF to be resolved, and possibly grounded in the Web.   
> Information regarding those URIs may be found as follows:
> 
>    Following the TAG's resolution of httpRange-14 [2]:
>    - Resolution may be attempted for any HTTP URI (an RDF subject,  
> predicate or object).
>    - Resolution of an HTTP URI resulting in a 2xx response 
> determines  
> that the nature of the resource identified by that URI is an  
> information resources;
>    - Resolution of an HTTP URI resulting in a 4xx ("error") response  
> determines that the nature of the resource identified by that URI is  
> unknown.
> 
>    Extending the resolution of httpRange-14:
>    - Resolution of an HTTP URI resulting in a 303 ("see other")  
> response determines that the nature of the resource identified by  
> that URI may be determinable by redirection.  That is, the URI  
> contained in the 303 response may be resolvable on the Web.
>      o If the URI returned in the 303 response is resolvable,  
> resolving that URI would result in an information resource 
> describing  
> the original RDF URI (either in human- or machine-readable form).   
> The SemWeb application should treat the RDF URI as known and may  
> optionally resolve the URI in the 303 response to obtain information  
> regarding it.  The server returning the 303 response is, in effect,  
> saying, "I'll confirm the existence of that name, and provide a  
> pointer to more information regarding it".

I think the above is too strong.  The 303 response by itself does not
confirm the existence of anything.  For example, thing-described-by.org
will return a 303 response for any query string that is a URI, regardles
of whether that URI exists.

>      o If the 303 response does not contain a URI (or the URI is not  
> resolvable), then the SemWeb application should treat the RDF URI as  
> known, but accept that further information regarding it is not  
> available.  The server returning the 303 response is, in effect,  
> saying, "I'll confirm the existance of that name, but nothing more  
> about it".

Ditto, I think the above is also too strong.  If the client is unable to
retrieve anything from the redirect target location, then nothing
further can be inferred about the existence or meaning of the original
URI.

>    - Resolution of HTTP URIs resulting in 4xx ("error") responses  
> should be treated by SemWeb applications as determining that no  
> information is known about the URI.  That is, the RDF URI is  
> unknown.  The server returning the 404 response is, in effect,  
> saying, "No such name exists".
> 
> 
> Example:
> 
> An RDF statement includes a predicate, http://purl.org/dc/elements/ 
> 1.1/creator.  Resolution of that URI results in a 303 response.  The  
> 303 response includes a URI, http://dublincore.org/2003/03/24/ 
> dces#creator.  A SemWeb application can determine that the predicate  
> is known and that further information regarding it is available at  
> the second (dublincore.org) URI.
> 
> Comments welcome!  Any errors in this message are mine; Ralph and  
> David have not yet reviewed my summary.
> 
> [1] http://www.w3.org/2005/06/27-swbp-minutes.html#action02
> [2] http://lists.w3.org/Archives/Public/www-tag/2005Jun/0039.html
> 
> Regards,
> Dave

Thanks,
David Booth

Received on Monday, 12 December 2005 08:19:04 UTC