Early Review of draft-ietf-netconf-yang-push-06
review-ietf-netconf-yang-push-06-yangdoctors-early-wijnen-2017-05-16-00

Request Review of draft-ietf-netconf-yang-push-04
Requested rev. 04 (document currently at 22)
Type Early Review
Team YANG Doctors (yangdoctors)
Deadline 2017-02-28
Requested 2017-02-07
Requested by Mehmet Ersue
Other Reviews Secdir Early review of -00 by Takeshi Takahashi (diff)
Yangdoctors Last Call review of -15 by Martin Björklund (diff)
Yangdoctors Last Call review of -21 by Martin Björklund (diff)
Review State Completed
Reviewer Bert Wijnen
Review review-ietf-netconf-yang-push-06-yangdoctors-early-wijnen-2017-05-16
Posted at https://mailarchive.ietf.org/arch/msg/yang-doctors/PKnkH2R9gMcwy06x61pHjj9Unww
Reviewed rev. 06 (document currently at 22)
Review result On the Right Track
Draft last updated 2017-05-16
Review completed: 2017-05-16

Review
review-ietf-netconf-yang-push-06-yangdoctors-early-wijnen-2017-05-16

- last para of sect 3.5.
  This seems to me to make it difficult to create interoperable
  implementations. Or is there a way for a client to figure out what
  is or is not support, other than tryal and error?
- page 41:    

     /* YANG Parser Pyang crashing on the following syntax below

  So does the definition get skipped? Or what needs to happen here?

Consistency

- last bullet on page 7 talks about "YANG subtrees". I do not see that term
  in netconf or yang documents. Those just talk about "subtrees". Maybe I am
  not looking good enough?
- top of page 8 I see the words "xpath", "Xpath" and "XPath"
  is there a difference?

Nits

- you may want to check the reference/citation occurrences of [subscribe]
  at several places it points to
     draft-ietf-netconf-yang-push-06#ref-subscribe
  whereas I think it intends to point to the [subscribe] in the
  normative references section
- first bullet on page 5:
     Enhancements to filters. Specifically the filter MUST at identify at
     least one targeted yang
   s/at//  -- the first "at" seems superfluous
   plus, you are using capitalized MUST with out reference/citation of
   RFC2119
- page 36:

    leaf dependency {
      type sn:subscription-id;
      description
        "Provides the Subscription ID of a parent subscription which
         has absolute priority should that parent have push updates
         ready to egress the publisher. In other words, there should be
         no streaming of objects from the current subscription if of
         the parent has something ready to push.";
      reference
        "RFC-7540, section 5.3.1";
    }

     s/if of/if/ ??