Re: [MMUSIC] Possible BUNDLE alternative syntax: explicit m-line for bundled session

[Christer Holmberg <[e-mail]>]

Hi,

>>> Instead, Let us discuss the generic categories of attributes and mention in this draft.
>>
>> I am not sure I understood.
>>
>> For sure we will have to discuss how/if attributes are affected, but we need to have a general idea on how the multiplexing is going to be offered. Same ports, m=bundle, or something else...
>
> One of the major benefits of the "pure" m=bundle proposal is that we *don't* have to discuss categories of attributes (other than to make recommendations to 
> implementors).  The attributes inside each unbundled m= line need to be self-consistent, and the attributes inside the m=bundle line also need to be self-consistent, 
> but there doesn't need to be any standardized mechanism to infer the one from the others.

That is my current understanding also.

And, IF we need to define new attributes, we will of course define the usage of them from scrath :) 

For example, the m=bundle draft I'm working on defines a new SDP attribute, which maps a PT to a specific media type (audio, video etc), as it cannot be assumed that the rtpmap attribute always can be used to identity the media type (in most cases it can, though).

Regards,

Christer