Re: [core] Review of draft-ietf-core-multipart-ct-02

Klaus,

I think that I would prefer that the choice between ordered or unordered be
a property of the application that is using the multipart collection rather
than hard coding it in.  I can think of use cases where both are things to
be desired and thus would not like to rule out either one.

Jim

> -----Original Message-----
> From: core <core-bounces@ietf.org> On Behalf Of Klaus Hartke
> Sent: Wednesday, November 7, 2018 9:26 PM
> To: Fossati, Thomas (Nokia - GB/Cambridge) <thomas.fossati@nokia.com>
> Cc: core@ietf.org WG <core@ietf.org>
> Subject: Re: [core] Review of draft-ietf-core-multipart-ct-02
> 
> I think the draft needs to be much clearer on whether a multipart
collection
> is an unordered multiset of representations (which I had assumed so far)
or
> an ordered sequence where applications attach additional out-of-band
> meaning to the positions.
> 
> Speaking of additional out-of-band meaning, I wonder if there should be
> identifiers in the collection explicitly describing the semantics in the
interest
> of self-description. There are also considerations related to evolving the
out-
> of-band information, such as when an application wants to change the
> meaning of a position but continue using the "application/multipart-core"
> content format ID.
> 
> Klaus
> 
> _______________________________________________
> core mailing list
> core@ietf.org
> https://www.ietf.org/mailman/listinfo/core