YANG Library
draft-ietf-netconf-rfc7895bis-07

Thanks for the work everyone did on this document.

ID Nits reports:

  ** Obsolete normative reference: RFC 5246 (Obsoleted by RFC 8446)
  ** Obsolete normative reference: RFC 6536 (Obsoleted by RFC 8341)

---------------------------------------------------------------------------

Page 16:

>      leaf checksum {
>        type string;
>        mandatory true;
>        description
>          "A server-generated checksum of the contents of the
>           'yang-library' tree.  The server MUST change the value of
>           this leaf if the information represented by the
>           'yang-library' tree, except 'yang-library/checksum', has
>           changed.";

I suspect that changing the name of this node in the tree would be disruptive
at this point in time, but this is clearly not a checksum ("There is no
requirement that the same information always results in the same 'checksum'
value"). I would suggest updating the description to use the term "version
identifier" or something similar.

---------------------------------------------------------------------------

§8.2:

>  [RFC8340]  Bjorklund, M. and L. Berger, Ed., "YANG Tree Diagrams",
>             BCP 215, RFC 8340, DOI 10.17487/RFC8340, March 2018,
>             <https://www.rfc-editor.org/info/rfc8340>.

Since RFC 8340 is required to understand the syntax used in the tree diagrams
used by draft-ietf-netconf-rfc7895bis, RFC 8340 should be normative rather
than informative.

Just a few minor comments:

Substantive Comments:

§2, 2nd bullet: "Each YANG module and submodule within the library SHOULD have a revision."
Why not MUST? Does it ever make sense not to have a revision? What are the consequences?

Editorial Comments:

§1, last paragraph: Missing article before "YANG Library" in first sentence.

§2, list item 1: "Efficient for a client to consume." - sentence fragment.
-- List item 3: Why is "NOT" in all-caps?
-- List item 6: The first sentence, while not technically a fragment, seems to use an understood "you" as the subject. I doubt that is the intent.

It's interesting that multiple entries for full-fledged module
implementation are forbidden, but import-only modules can have multiple
(different) entries.  I'm not familiar enough with the YANG ecosystem to
understand why one is more common or more reasonable to permit than the
other.