Re: [AVTCORE] AD Evaluation draft-ietf-avtcore-rfc5285-bis-08

[Roni Even <[e-mail]>]

Ben,
See inline
Roni
> 
> 
> Hi,
> 
> This is my AD evaluation of draft-ietf-avtcore-rfc5285-bis-08. These are
> mostly editorial, but I'd like to resolve at least the substantive
> comments prior to IETF last call.
> 
> Thanks!
> 
> Ben.
> 
> - Substantive:
> 
> -- 6, paragraph 3, "... unless the
>     transmitter has some (out of band) knowledge..."
> Why bother with that exception? It seems like the negotiation is easy
> enough. If people write code assuming that out-of-band knowledge will be
> available, interoperability will suffer.
[Roni Even] We need it since RFC3264 is not the only usage. For example for declarative SDP there is no negotiation
> 
> -- 9, 2nd paragraph:
> 
> 3264 does not require integrity protection to be _used_, just _provided_
> (read: MUST implement). Does this mean to promote that to "MUST use"? If
> so, please be clear that this is a new requirement.
> 
> But from a practical perspective, do you believe people will really
> follow a "MUST use"? Is it a MUST (BUT WE KNOW YOU WON”T) [RFC6919] ?
[Roni Even] I am OK with changing it to" MUST be implemented"
> 
> 
> - Editorial:
> 
> -- 3, third paragraph:
> 
> This changes the "... mechanism may be available..." from the original
> to "... mechanism can be available...". I think "may be" is more correct
> here.
[Roni Even] OK
> 
> -- 4.1.1, first paragraph, first sentence:
> This is a statement of principle, and as such shouldn't use 2119
> keywords.
[Roni Even] OK, is changing to lower case enough? I can add a sentence to section 2 that normative language is written in capital letters in the document.
> 
> -- 4.1.2, 2nd paragraph: s/len/length  ; or "len field"  (2 occurances)
> 
[Roni Even] OK
> -- 4.1.2, 3rd paragraph, sentence starting with "Only the extension..."
> 
> “Only…MUST” constructs are ambiguous. The can be interpreted to
> say the MUST only applies to the matching set, and the rest are
> undefined. Please recast as “MUST NOT ”. (Note that this construct
> appears elsewhere in the draft, but I think the other occurrences are in
> the original text. I don't expect this update to fix those.)
[Roni Even] The text says that only one byte or two byte rtp header extensions MUST be used in an RTP packet. I am not sure what is confusing here?
> 
> --- "A transmitter MAY be aware...": That's a statement of fact; the MAY
> should not be capitalized.
[Roni Even]OK
> 
> -- 4.2, first sentence, first paragraph: The REQUIRED seems to be
> talking about an external requirement from RTP. If so, it should not be
> normative here.
[Roni Even]OK
> 
> -- 7, 11th paragraph, "Local identifiers in the valid range inclusive in
> an offer or answer
>     MUST NOT be used more than once per media section"
> 
> The MUST NOT was not capitalized in 5285, probably because it was the
> 4th time the requirement was mentioned :-). I don't think it needs to be
> promoted here.
[Roni Even]OK
> 
> -- 9, 1st paragraph: I don’t think this is an appropriate use of a
> 2119 MUST. (It wasn’t capitalized in the original.)
> 
[Roni Even]OK
> 
>