Re: [core] multipart/core

[Carsten Bormann <[e-mail]>]

Hi Alexey,

On Jun 6, 2018, at 16:44, Alexey Melnikov <[e-mail]> wrote:
> 
> Hi,
> I would like to quickly insert AD interrupt.

getting input from the media type (and core-parameters) experts would have been next on the list.
Thank you for accelerating that input a lot :-)

> I had a quick look at draft-fossati-core-multipart-ct-04 and I think the proposed definition of multipart/core is syntactically incompatible with MIME. Please read sections 5.1 and 5.1.1 of RFC 2046. In particular, all multipart media types have to use syntax specified in RFC 2046, which this proposal violates.
> 
> To be clear, I am not against registering a media type which uses CBOR syntax for a collection of other media types. However this media type name can't use the "multipart" top level media type. You can either define a new "application" media type (e.g. application/multipart-core) or you can define a new top level media type different from "multipart". The process for definining a new top level media type is slightly more comlicated, so you should evaluate whether you want to do this.

There is no particular need to use the top-level “multipart” (or any other top-level for that matter), so we would go for an application/ top-level.

(The draft from 2013 that this new version continues wasn’t suggesting a media type name, so the not-so-well-thought-out suggestion to use multipart/ was new in this version.) 

Questions to the WG:

Anyone have a problem with the name application/multipart-core?
(It’s not visible in the CoAP exchanges, we just see the content-format identifier there.
We could discuss whether that should be a single-byte identifier; the current draft goes for a two-byte one.)

Grüße, Carsten