Re: [xrblock] Last Call Review: draft-ietf-xrblock-rtcp-xr-post-repair-loss-count-07

[Tom Taylor <[e-mail]>]

Replies with [PTT] below.

On 25/12/2014 3:21 AM, Huangyihong (Rachel) wrote:
> Hi Tom,
>
> Thank you for the nice Gen-ART review. Please see my replies inline starting with "[Rachel]".
>
> BR,
> Rachel
>
>
>> -----Original Message-----
>> From: Tom Taylor [mailto:[e-mail]]
>> Sent: Thursday, December 25, 2014 10:42 AM
>> To: Gen Art; [e-mail];
>> [e-mail]; [e-mail]; Dan Romascanu
>> Subject: Last Call Review: draft-ietf-xrblock-rtcp-xr-post-repair-loss-count-07
>>
>> I am the assigned Gen-ART reviewer for this draft. For background on Gen-ART,
>> please see the FAQ at
>>
>> <http://wiki.tools.ietf.org/area/gen/trac/wiki/GenArtfaq>.
>>
>> Please resolve these comments along with any other Last Call comments you
>> may receive.
>>
>> Document: draft-ietf-xrblock-rtcp-xr-post-repair-loss-count-07
>> Reviewer: Tom Taylor
>> Review Date: 2014-12-24
>> IETF LC End Date: 2014-12-26
>> IESG Telechat date: 2015-01-08
>>
>> Summary: This document is mostly to publish as a Proposed Standard, with
>> minor issues verging on editorial corrections.
>>
>> Major issues: None.
>>
>> Minor issues:
>>
>> 1) Second paragraph of Introduction: last sentence introduces the "repaired
>> loss count". I found this confusing. Figure 1 shows you really meant "unrepaired
>> loss count".
>
> [Rachel]: This report block introduces two metrics : "unrepaired loss count" which is specified as the total number of packets finally lost after applying loss-repair methods. And "repaired loss count" is specified as the total number of packets which are fully repaired after applying loss-repair methods. And in this sentence, I meant "repaired loss count" because this metric is used to help calculating pre-repair loss count (pre-repair loss count = unrepaired loss count + repaired loss count). Maybe the name of "unrepaired loss count" is confusing. Would it better changing to "post-repair loss count"?

[PTT] Yes, "post-repair loss count" sounds good.
>>
>> 2) First paragraph of Section 3, third sentence: the ending currently reads:
>>      "Some packets will not be repaired in the current
>>      RTCP interval."
>> I think your point is that some of these could be repaired later, so I suggest the
>> text should be extended:
>>      "Some packets will not be repaired in the current
>>       RTCP interval, but could be repaired later."
>>
>
> [Rachel]: Right. Will fix it in the new version.
>
>> 3) I question whether RFCs 4588 and 5109 are normative for this document,
>> but I know this sort of thing is a difficult decision and I'll accept whatever the
>> final outcome is.
>
> [Rachel]: I agree with you that these two should be informative.
>>
>> 4) Appendix A: The "Units of Measurement" are packets.
>
> [Rachel]: According to RFC6390, Section 5.4.5, there's an example to show how to define a metric. Here's how it defines the "Units of Measurement":
> "
> Units of Measurement:  This metric is expressed as a fixed-point number with the binary point at the left edge of the field.  For example, a metric value of 12 means a loss rate of approximately 5%.
> "
> I consulted it to specify mine.
>
[PTT] But what you are measuring according to your description and text 
in the preceding body of the document is number of packets, not a loss 
rate as shown in the RFC 6390 example. So what you should say is:

   "This metric is expressed as a 16-bit unsigned integer giving the 
number of packets ...".
>>
>>
>> Nits/editorial comments:
>>
>> 1) First paragraph of introduction:
>>         s/contains/contain/ at end of first line.
>>
>> Fourth sentence:
>>
>> OLD
>>
>>      However, this metric is measured on media stream before
>>      any loss repair mechanism, e.g., retransmission [RFC4588] and Forward
>>      Error Correction (FEC) [RFC5109], is applied.
>>
>> NEW
>>
>>      However, this metric is measured on the media stream before
>>                                          ^^^
>>      any loss repair mechanism, e.g., retransmission [RFC4588] or Forward
>>                                                                ^^
>>      Error Correction (FEC) [RFC5109], is applied.
>
> [Rachel]: Okay. All the above nits will be fixed.
>
>>
>> I would break this paragraph into three parts. The first three sentences
>> introduce the existing metric. The next two sentences, from "However, ..." to
>> "... [RFC3550].", introduce the problem. The following two sentences, from
>> "Consequently ..." through "... higher overhead.", give an attempted solution.
>> The final sentence, I think, is the motivator for this document and I would
>> merge it into the next paragraph.
>
> [Rachel]: Good suggestion.
>
>>
>> 2) First paragrpah of Section 3: again, I suggest it be broken into three parts.
>> The second part would begin with "Thus it is RECOMMENDED ...", and the
>> third with "The sequence number range ...".
>
> [Rachel]: Okay. So you're suggesting to break the big paragraph into 3 small ones, right?
>
[PTT] Yes.
>>
>> 3) The sentence preceding Figure 1 should identify "Figure 1" explicitly (via
>> XML2RFC cross-reference if that is what you are using). It should also note that
>
> [Rachel]: Okay. Will do.
>
>> bit positions are given in octal -- the first time I've seen that in a document, but
>> it's OK with me.
>
> [Rachel]: This draft is just an RTCP XR extension to augment the report blocks defined in RFC3611. So the format is just in a manner consistent with other RTCP XR packets. All the RFCs produced in XRBLOCK wg specifying new report blocks have the same pattern. That's why I think it doesn't need to be clarified in this draft.
>
>>
>> 5) Section 5, third line: s/confidentially/confidentiality/
>
> [Rachel]: Okay.
>