Re: Re[4]: AW: Content negotiation flamewar (was: Re: "Hash URIs" and content negotiation)

Xiaoshu,

On 14 Nov 2006, at 22:18, Xiaoshu Wang wrote:
>>  From RFC 2854 [1]:
>>
>> | For documents labeled as text/html, the fragment identifier
>> designates
>> | the correspondingly named element; any element may be named
>> with the
>> | "id" attribute, and A, APPLET, FRAME, IFRAME, IMG and MAP
>> elements may
>> | be named with a "name" attribute.
>>
>> So, the frag id names an *element*, a structural part of the  
>> document.
>>
>> This increases my conviction that, if #Bob is a person, a 303
>> should be done before we serve HTML.
>
> The semantics you refered is the semantics of text/html but not the  
> HTTP
> protocol. As far as the HTTP protocol is concerned, the #Bob is never
> requested.  What is requested is the http://example.com/resource,  
> which is
> an information resource.

True.

> How to interpret the semantics of #Bob is at the
> client side and it is not coverred by the httpRange-14.

It's not covered by httpRange-14, but by the URI spec. The URI spec  
says that the content type of the response to http://example.com/ 
resource determines what #Bob identifies.

I don't think it makes sense to say that the semantics of #Bob is  
interpreted at the client side. By serving some content type, the  
naming authority at example.com communicates something about the  
meaning of #Bob. By serving HTML, it communicates that #Bob  
identifies a part of an HTML document. By serving a 303, it  
communicates nothing except what is said in the target document of  
the redirection.

> If the agent
> "thinks" it has requested the http://example.com/resource#Bob, then  
> it is
> the wrong implementation of the agent that leads to the wrong  
> conclusion,
> but not the httpRange-14 resolution, don't you think?

The client has to request http://example.com/resource, but the  
response will tell something about #Bob.

Richard






>
> Xiaoshu
>
>

Received on Thursday, 16 November 2006 05:17:05 UTC