Re: [payload] [Gen-art] Gen-ART review of draft-ietf-payload-rtp-opus-08

["Ben Campbell" <[e-mail]>]

On 15 Apr 2015, at 4:38, Christer Holmberg wrote:

> Hi,
>
>>> Minor Issues: I previously gave the following comment:
>>>  
>>> “Regarding SDP, I think it would be good to have the ABNF syntax 
>>> for
>>> the a=fmtp parameter (currently you only have descriptive text of 
>>> the
>>> different parameters). It makes the life for the parser implementers
>>> much easier :)”
>>>  
>>> I guess one, by reading section 7 and the examples, can figure out 
>>> how
>>> to encode the a=fmtp parameter, but I think it would to explicitly
>>> define the syntax.
>>
>> (For the record, I just recently took over responsibility for 
>> PAYLOAD, so if I'm misinterpreting things, someone please tell me :-) 
>>  )
>>
>> While I can see that as a "might be nice" addition, I don't think 
>> it's something that we have required of other payload drafts.
>
> After all my years of taking drafts through the IETF publication 
> process, I have learned that there is no such thing as consistency :)
>

That is absolutely true. And I do think it's worth discussing whether 
future payload specs should do that. It's just a matter of who gets 
stuck with a new documentation requirement :-)

>> draft-ietf-payload-rtp (in RFC ed queue) says the following about 
>> ABNF for SDP parameters:
>>
>> "Not that commonly used in RTP payload formats but may be useful when 
>> defining Media Type parameters of some complexity."
>>
>> If I'm reading correctly, the FMTP parameters in this draft fit the 
>> pretty common "semicolon delimited list of parameter=value pairs", so 
>> I don't think this rises to the level of "some complexity".
>>
>> So unless you think the parameters in this draft are more complex 
>> than average, I don't think we need to add them at this late stage. 
>> It might be worth discussing whether we should ask authors of future 
>> payload drafts to include ABNF for this sort of thing.
>
> I can live without any changes.

Thanks!

>
> Regards,
>
> Christer