Re: ISSUE-57: The use of HTTP Redirection

Ed Davies wrote:
>
> Xiaoshu Wang wrote:
>> I think a less intrusive way is to give the RDF mimetype 
>> representation (application/rdf+xml, text/rdf+n3,...) a unique 
>> status.  Because RDF document always talk about something-else.  
>> So, GET (rdf) http://example.com/
>
> Sorry, but this seems like a very poor solution to me.  It
> complicates the meaning of GET (to add the possibility of
> returning metadata instead of a representation of the
> resource).  It seems to be quite reasonable to ask for
> metadata about an RDF document; for example, you might want
> provenance or validity period information.
I don't understand what you mean by complicating the meaning of GET?  
What is the meaning of GET? GET, for me, means give me something 
relevant.  You mean GET means specifically GET data or metadata?

Provenance and validity do make it a bit complicated.  But since 
Provenance and validity will be talked in RDF, so we can always makes 
our meanings clear with specialized vocabulary.  The use case I am 
trying to think of is the non-RDF technology talks about RDF.
> Also, there could be horrible interactions with content
> negotiation.  Some content types (e.g., application/xhtml+xml
> with suitable GRDDL markup) could be considered "RDF
> mimetypes" and hence be quite sensible responses to a request
> for metadata yet also be used for likely representations of
> the resource itself.
I don't see why this confuses or complicated anything.  If a client can 
only process RDF, he sets his content header to get only RDF.  If he can 
process GRDDL of some sort as well, he put in his Accept header with 
different priority.  On the server side, if the server process it as it 
is before.  But when the agorithsm deciding the returned MIME type is 
RDF and he can generate one, just 200 it back.  Otherwise, 303 it to 
somethingelse like HTML where GRDDL is embeded.  Which part it will mess 
it up?

Cheers,

Xiaoshu

Xiaoshu

Received on Monday, 27 August 2007 17:20:44 UTC