Hi, Tom: Thanks for your valuable review. See my reply inline. Regards! -Qin -----邮件原件----- 发件人: Tom Taylor [mailto:tom.taylor.stds@xxxxxxxxx] 发送时间: 2014年7月7日 10:37 收件人: xrblock-chairs@xxxxxxxxxxxxxx; draft-ietf-xrblock-rtcp-xr-psi-decodability@xxxxxxxxxxxxxx; Alissa Cooper; Gen Art; The IETF 主题: Gen-ART IETF LC review of draft-ietf-xrblock-rtcp-xr-psi-decodability-04 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-psi-decodability-04 Reviewer: Tom Taylor Review Date: 6 July 2014 IETF LC End Date: 7 July 2014 IESG Telechat date: (not known) Summary: basically ready with very minor issues and a number of editorial suggestions. Major issues: none. Minor issues: (1) It might be helpful to add text in Section 3 explaining that PAT_error_2_count and PMT_error_2_count are actually replacements for and improvements on PAT_error_count and PMT_error_count respectively and are therefore preferred in future implementations. [Qin]: Okay. (2) Condition (2) of PAT_error_2_count: "one table with table_id other than 0x00" is more precise than intended by [ETSI]. s/one/a/. This comment also applies to PMT_error_2_count (third from last line of first paragraph) and CAT_error_count (both conditions). [Qin]:Accepted. Nits/editorial comments: General: blanks are missing in a number of places, typically following a comma or preceding a parenthesis. [Qin]: Okay. Abstract -------- "statistics metrics" seems a bit redundant, but I wonder if "metric" has a special meaning to people working in this area. To me, "metric" is another word for "measurement result". So its use to describe the contents of the XR block makes sense. However, when we get to Section 3, "metric" is used in place of "indicator". Is that really correct usage? [Qin]: To avoid confusion, we can make the following change OLD TEXT: " ETSI TR 101290 [ETSI] generally defines metrics related to error events while this document contains counts of those metrics defined in [ETSI]. " NEW TEXT: " ETSI TR 101290 [ETSI] generally defines parameters related to error events while this document contains counts of those parameters defined in [ETSI]. " s/Program specific information/Program Specific Information/ [Qin]:Okay. Section 1.1 ----------- Some redundancy with the opening paragraph of 1.1, some cramming together of different ideas. Suggested alternative: OLD This memo is based on information consistency tests and resulting indicators defined by ETSI [ETSI] and defines a new block type to augment those defined in [RFC3611] for use with MPEG2 Transport Stream (TS) [ISO-IEC.13818-1.2007]. The new block type supports reporting of the number of occurrences of each Program Specific Information (PSI) indicator in the first and second priorities that supplements information from PSI independent Decodability Statistics Metrics Block [RFC6990]; third priority indicators are not supported. NEW This memo defines a new block type for use with MPEG2 Transport Stream (TS) [ISO-IEC.13818-1.2007], to augment those defined in [RFC3611]. The new block type supports reporting of the number of occurrences of each Program Specific Information (PSI) indicator in the first and second priorities listed by [ETSI] sections 5.2.1 and 5.2.2 respectively. Third priority indicators are not supported. The metrics defined here supplement information from the PSI-independent Decodability Statistics Metrics Block [RFC6990]. [Qin]: Accepted. Section 1.2 ----------- s/defined/defines/ on second line for consistency with the other sentences. [Qin]: Okay. Section 1.3 ----------- s/Architectures [RFC6792]/Architecture [RFC6792]/ s/guideline/guidelines/ s/for reporting block format using RTCP XR/for RTCP XR reporting block formats/ [Qin]:Okay. Section 1.4 ----------- s/;/,/ on second line. s/;/./ on third-last line. [Qin]: Okay. Section 3 --------- See remark on use of "metric" above (Section 1.1). Could the first sentence be rewritten: OLD ETSI TR 101290 [ETSI] generally defines metrics related to error events while this document contains counts of those metrics defined in [ETSI]. NEW ETSI TR 101290 [ETSI] generally defines indicators related to error events, while the XR block defined in this document contains counts of occurrences of the [ETSI] indicators. [Qin]: I am okay with this proposed change. Fifth line: s/PSI independent/PSI-independent/ (add hyphen) [Qin]: Okay. Paragraph below the CRC and CAT bullets: (1) What do you mean by: "scrambling may be considered"? Do you mean that the presence or absence of scrambling is part of the error checking, or something else? [Qin]: This sentence did introduce a few confusion, what about the following change: NEW TEXT: " Measurement results for some of these parameters (e.g.,PAT error or PMT error) may be different based on whether scrambling is employed " (2) I'd suggest expanding "The other parameters ..." to "The other parameters defined in [ETSI] Section 5 [or whatever scope you intended] but not listed above ...". [Qin]: Okay. Section 3, PID_Error_Count -------------------------- Second sentence is not quite accurate. It should read: OLD A PID_error occurs when MPEG TS streams are remultiplexed and any PID doesn't refer to an actual data stream, as defined in the section 5.2.1 of [ETSI] NEW A PID error occurs [is indicated?] when no data stream is present corresponding to a given PID. This may be caused by multiplexing or demultiplexing, then remultiplexing. See section 5.2.1 of [ETSI]. [Qin]: Accepted. For consistency, we prefer to use 'occurs' instead of "is indicated"