Re: [MMUSIC] AD Evaluation of draft-ietf-mmusic-rfc4566bis-32

[Paul Kyzivat <[e-mail]>]

On 3/11/19 1:43 PM, Colin Perkins wrote:

>>> §5 has several changes to normative language. Most are okay, but I think the change from “all MUST appear in exactly the order given here” to “all must appear” weakens it too much, and I’d prefer that to remain a MUST.
>>
>> The reason for changing that is because the *normative* specification of the ordering is from the ABNF. The text here is explanatory and non-normative. (Note that a couple of paragraphs prior to this is a new statement emphasizing that the ABNF is normative.)
> 
> Sorry, but I think this change is problematic. The text needs to use normative language that is consistent with the ABNF.

My thinking is that we don't like to have redundant normative 
specification, just in case they aren't consistent. The ABNF is 
normative. This section is of necessity an approximation.

But I defer to Ben or whoever.

>>> This is a normative requirement of RTP, however. If we want to avoid normative examples, which I’d agree makes sense, then this needs to be rewritten as just “An RTP-based system in recvonly mode SHOULD…”.
>>
>> That is the way it was. The change from "SHOULD" to "should" was to make it non-normative.
>>
>> Or were you requesting to remove “e.g."?
> 
> Remove the “e.g.”, yes, but also change “should” back to “SHOULD”.

I thought the point was that the RTP specs are normative and this is 
only an example, and so shouldn't be normative.

	Thanks,
	Paul