[occi-wg] Where is the scheme ?

[Alan Sill]

So what I get from this discussion is that for the infrastructure URL semantics, the blank-document approach is OK.

What about the others - for example, http://schemas.ogf.org/occi/core (#resource and #link) - which still return 404 Not Found ?  Should the same solution be applied?

Alan


On Apr 6, 2011, at 3:51 AM, Andre Merzky wrote:

> Ralf, thanks for clarifying the HTTP semantics for the fragments.
> That semantics makes sense I think, as the whole document needs to be
> retrieved anyway.
> 
> Cheers, Andre.
> 
> 
> On Wed, Apr 6, 2011 at 8:54 AM, Ralf Nyren <ralf at nyren.net> wrote:
>> 
>> On Wed, 6 Apr 2011 07:46:36 +0200, Andre Merzky <andre at merzky.net> wrote:
>> 
>>> One could also link the normative documents for the
>>> extension at http://schemas.ogf.org/occi/infrastructure#specification;
>>> one could verify the #target elements; etc.  However, that is as of
>>> yet unspecified AFAICS, and is likely a rather naive proposal, given
>>> my somewhat limited knowledge of HTTP itself.
>> 
>> There is a little problem with the fragment part of the URI, i.e. the
>> #xxxx. According to the RFCs it is not part of the URL sent to the HTTP
>> server. It is something that the client has to handle by itself.
>> 
>> For example:
>> - Clicking on the http://schemas.ogf.org/occi/infrastructure#compute URL
>> in your favorite browser will yield an HTTP request as follows:
>> GET /occi/infrastructure HTTP/1.1
>> 
>> - In other words the _whole_ infrastructure "document" will be retrieved.
>> The client (in this case your browser) will attempt to translate the
>> #compute part into something useful with regard to the content-type of the
>> document received. In case of text/html it is to jump to the fragment
>> marker in the document.
>> - How to handle the URL fragments in a future OCCI schema definition must
>> be defined together with the format. However it is still the client which
>> has to do the processing.
>> 
>> regards, Ralf
>> 
>> 
> 
> 
> 
> -- 
> So much time, so little to do...
> [Garfield]
> _______________________________________________
> occi-wg mailing list
> occi-wg at ogf.org
> http://www.ogf.org/mailman/listinfo/occi-wg