Re: [Fwd: draft-reschke-http-addmember-00]

Helge Hess wrote:
> 
> On Feb 15, 2005, at 14:03, Julian Reschke wrote:
> 
>>> If we create a new method, I would like to ask for the possibility to 
>>> return multiple resources. The one situation which doesn't fit well 
>>> with PUT+location is a PUT of a cyclic event which is flattened by 
>>> the server to multiple resources.
>>
>> So how would you define the semantics for that in a generic way? A 
>> multi-part request body?
> 
> 
> No, the request body is just one entity. The difference is that the 
> server might decide to split the resource provided by the request into 
> multiple resources on the server.
> 
> The example is in the quote above, you ADDMEMBER an iCalendar entity 
> which contains a recurrence rule. A lot of servers do not work with 
> calculated recurrences as in iCal, but rather maintain a recurrence as a 
> set of individual events. So if the requests arrives at such a server, 
> it would create n different resources for a single request entity.
> It would be nice to be able to report that fact to the client.
> 
> Actually I'm not sure whether something like this belongs into a core 
> HTTP draft, I think the mentioned example is the only one which comes to 
> my mind which requires that type of a response.

Yup, and that's why I think that a supposed-to-be generic method would 
be the wrong choice here.

> ...

Best regards, Julian

-- 
<green/>bytes GmbH -- http://www.greenbytes.de -- tel:+492512807760

Received on Tuesday, 15 February 2005 13:36:03 UTC