Re: [core] [netmod] WG Last Call on draft-ietf-core-sid-21

[Andy Bierman <[e-mail]>]

On Sun, Sep 10, 2023 at 11:59 AM Michael Richardson <[e-mail]>
wrote:

>
> Andy Bierman <[e-mail]> wrote:
>     > I do not have time to do another full review of this draft.
>     > These are high-level comments.
>
>     > ## document scope is too large
>
>     > The draft should be split into 2 or 3 drafts.
>
>     > 1) SID File Definition
>     > 2a) SID Management Procedures
>     > 2b) Initial SID File Definitions
>
> It sounds like a good idea in principal, and if it were 2018, I'd say,
> sure.
> But I'm not really sure we are going make anything easier by splitting
> things
> up now.
>
>     > ## document scope is not CORE-specific
>
>     > Several WGs and SDOs want to use SID files.
>     > The 13 modules listed in sec 6.4.4 are a small fraction of the IETF
> YANG
>     > modules
>     > that should have SID files.
>
> I don't understand why it matters where we publish the work, as long as we
> do.
>
>

I don't understand where the SID files are for the modules in the initial
registry (sec. 6.4.4).
Is that for another WG to create the SID files and publish them in an RFC?


    > ## SID files are normative
>
>     > - The algorithm is not stable enough yet (both specification and
>     > implementation).
>     > - Manual editing is allowed.
>     > - There is no indication in the SID file that it is automatically
> generated
>     > or manually edited.
>     > - Only the authoritative SID file can be trusted to have the correct
>     > assignments
>
> These seem like reasonable complaints, but I'm not sure we can get past
> this
> until we publish a documents.  It's a *PROPOSED STANDARD*
>

The standard is fairly dependent on availability and conformance to
authoritative SID files.


>     > ## Initial SID files are missing
>
>     > The normative SID files for all YANG modules listed in sec 6.4.4 are
>     > missing.
>     > The assignment range is insufficient for interoperability.
>
> I hadn't imagined that we'd have to do all that work before publishing, but
> rather that we couldn't do any of that work until we published.
>


IMO this draft should include the normative SID files for the initial
registry contents.



>
> --
> Michael Richardson <[e-mail]>   . o O ( IPv6 IøT consulting )
>            Sandelman Software Works Inc, Ottawa and Worldwide
>
>
>
>
Andy