Re: [Ace] ace-coap-est-08: using /skg with Accept Option set to TBD287

> CoAP is not aware that the representation happens to contain embedded representations and therefore the content negotiation mechanism cannot be used directly to negotiate the formats of those. 
> The value of the Accept option in the request needs to be registered in the IANA registry and the value of the Content-Format option in the response must be the same as Accept value.
> Of course, one possible solution is to drop the use of content format ID 62 entirely and just register one ID for each possible combination. (But then the client can still only include at most one Accept option in its request.)

Hmm, that is a fair point. I don't think it is warranted to register four more content formats for all possible format combinations in the multipart response. 

It looks to me that your proposal of using Uri-Query in the request in order for the client to define the supported formats of the requested resource/response is a good one.

-----Original Message-----
From: Ace <ace-bounces@ietf.org> On Behalf Of Klaus Hartke
Sent: Tuesday, February 12, 2019 4:36 PM
To: Panos Kampanakis (pkampana) <pkampana@cisco.com>
Cc: Esko Dijk <esko.dijk@iotconsultancy.nl>; ace@ietf.org
Subject: Re: [Ace] ace-coap-est-08: using /skg with Accept Option set to TBD287

Panos Kampanakis wrote:
> Well, RFC7252 refers to a singular content format. In our case we are talking about a dual content format (286 or 281 and 280 or 284) returned in a 62 multipart-content. Would it be a violation of RFC7252, since RFC7252's text had single content format responses in mind only?

>From the point of view of CoAP, there is just a representation with
content-format 62. A client can indicate that it accepts a representation with content-format 62; the server then is required to return either a representation with content-format 62 or an error.
CoAP is not aware that the representation happens to contain embedded representations and therefore the content negotiation mechanism cannot be used directly to negotiate the formats of those.

>>  Maybe the draft-ietf-core-multipart-ct should extend the semantics of "Accept" to cover this case?

A content format is not a protocol extension and cannot override the protocol definition.

> I think that is good idea. The simplest way to do that would be encode the 3 content formats (for example 62, 286 and 280) into a single CF included in the Accept option which tells the server what combination of content formats to send back. Would that violate RFC7252 because the Content-Formats needs to be actual CFs defined in the IANA registry and not a combination of them?

The value of the Accept option in the request needs to be registered in the IANA registry and the value of the Content-Format option in the response must be the same as Accept value.

Of course, one possible solution is to drop the use of content format ID 62 entirely and just register one ID for each possible combination.
(But then the client can still only include at most one Accept option in its request.)

> From a previous thread with Jim S., I was under the impression that In the virtual CoAP WG meeting a month back we went through in some explicit detail that both Content-Format and Max-Age have no meaning when appearing on a request and therefore should not be there.

Max-Age doesn't have a meaning in requests and therefore must not be there. I'm not sure where that about the Content-Format option comes from. If a POST request has a payload, then the format of that payload is described by a Content-Format option. (A GET request doesn't have a payload and therefore must not include a Content-Format option.)

Klaus

_______________________________________________
Ace mailing list
Ace@ietf.org
https://www.ietf.org/mailman/listinfo/ace