Re: [bfcpbis] Draft new version: draft-ietf-bfcpbis-rfc4583bis-25

[Paul Kyzivat <[e-mail]>]

On 9/26/18 9:47 AM, Mary Barnes wrote:
> We definitely need Paul Kyzivat to review and ensure that he's okay with 
> the changes, since he did the SDP directorate review. 

I just did a review of the changes since my review of -22, looking at 
the SDP aspects.

Thanks for updating the format of the SDP specifications!

* General comment:

I have a slight preference for doing references to rules defined in 
other documents differently. E.g.:

    DIGIT = <DIGIT defined in [RFC5234]>

rather than:

    ;DIGIT is defined in [RFC5234]

The advantage is that when mechanically verifying the SDP, the comment 
form results in an undefined symbol while the <> form doesn't. (No, I 
didn't mechanically verify the SDP this time.)

* Section 5.2:

Running both confid and userid into a single subsection is confusing 
rather than helpful. Can you split them please?

* Section 5.3:

The following is wrong:

    floor-id = "a=floorid:" 1*DIGIT SP "mstrm:" token *(SP token)

A minimal fix would be:

    floor-id = 1*DIGIT SP "mstrm:" token *(SP token)

But that reads funny. ISTM the string of digits is the floor-id and the 
sequence of tokens is something else. The attribute seems to be poorly 
named, but you probably can't change it now. I don't have a good 
suggestion on how to improve it.

* Section 5.4:

As in 5.3, the following:

    bfcp-version = "a=bfcpver:" version *(SP version)

ought to be:

    bfcp-version = version *(SP version)

> Note, he 
> mentioned  in a posting that he thought the changes warranted another 
> last call and I don't recall seeing a response to that.   As shepherd, I 
> kindof agree, although I'm okay as long as everyone that reviewed 
> previously confirms on the list that they have reviewed the changes and 
> think the doc is ready to go.

I'll leave that to you all to decide.

	Thanks,
	Paul