Re: [MMUSIC] Resolving IESG issues with RFC4566bis-35: "a=type" impact on media direction default

[Paul Kyzivat <[e-mail]>]

Once more, I'm looking for insight, especially old timers (Mark, Van, 
Colin, Magnus, Henning) from RFC2327 days. If I don't hear anything I'm 
going to change the text so this "type-specific" defaulting of 
directionality no longer appears to be normative.

	Thanks,
	Paul

On 6/6/19 12:56 PM, Paul Kyzivat wrote:
> I've been working to resolve the IESG issues with RFC4566bis-35.
> I think I have resolved most of them, but I have a couple where I feel I 
> need input from the wg. This is one of those:
> 
> "a=type" impact on media direction default:
> 
> While I was resolving a comment on Section 6.9 (a=type) I noticed what I 
> think may be a new issue. The text says:
> 
>     "recvonly" should be the default for "type:broadcast" sessions,
>     "type:meeting" should imply "sendrecv", and "type:moderated" should
>     indicate the use of a floor control tool and that the media tools are
>     started so as to mute new sites joining the multimedia conference.
> 
>     Specifying the attribute "type:H332" indicates that this loosely
>     coupled session is part of an H.332 session as defined in the ITU
>     H.332 specification [ITU.H332.1998].  Media tools should be started
>     "recvonly".
> 
> This seems to imply that the *default* media direction 
> (sendonly/recvonly/...) will vary depending on the value of the type 
> attribute. Yet it isn't written in a normative manner. (This language 
> dates back to RFC2327.)
> 
> My first reaction to this is that while it is written oddly the intent 
> is only that these are suggestions for what the creator of the SDP 
> should use. But then I also found language about this in section 6.7 
> (Media Direction Attributes):
> 
>     If none of the media direction attributes is present at either
>     session level or media level, "sendrecv" SHOULD be assumed as the
>     default for sessions that are not of the multimedia conference type
>     "broadcast" or "H332" (see below).
> 
> This language was not in RFC2327 - it was added in RFC4566. (Probably by 
> me!)
> 
> My current thinking is that this is messed up. At the time of the 
> writing in RFC2327 the use case in mind was announcements via SAP. This 
> really doesn't work right in an offer/answer context.
> 
> I'm inclined to take the language about type-specific directionality out 
> in both places. Or, if that can't be done for reasons of backward 
> compatibility, then perhaps qualify it as only applying in the case of 
> announcements.
> 
> I need feedback here!
> 
>      Thanks,
>      Paul
> 
> 
> 
> _______________________________________________
> mmusic mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/mmusic
> 

_______________________________________________
mmusic mailing list
[e-mail]
https://www.ietf.org/mailman/listinfo/mmusic