On Fri, Jan 9, 2015 at 9:05 AM, Jeffrey Haas <[email protected]> wrote:
> On Thu, Jan 08, 2015 at 05:53:08PM -0500, William Atwood wrote:
>> No, I don't have anything specific---my expertise is not in software
>> evaluation.  I was just concerned that Jeffrey's comment implied that 3
>> Routing Area directors could solve this problem.
>
> Note that I didn't say that a third AD will solve this problem.  I referred
> people to the discussion about the motivations for adding one.
>
> In general, document review in IETF should follow a well defined process
> similar to what is seen in many large software organizations:
> - Reviewed by those working on it. (Authors)
> - Reviewed by their peers. (Working Group)
> - Reviewed by targeted experts, especially for cross-area work.
>   (Directorates)
> - Reviewed for big-picture architecture. (IESG)
>
> Where things have the appearance of breaking down is when there has been
> insufficient review at the lowest layers.  WG chairs should not send along
> documents for directorate or AD review without having some certitude about
> the document being of high enough quality to do so.  In project management
> parlance, they're the gatekeepers.
>
> Yang modules, and MIBs before them, have a few interesting issues that
> complicate the normal review process:
> - Modeling as a discipline is something that not everyone is good at.  IETF
>   tends to attract people who like working bits-on-the-wire, not management
>   and operations front-ends.
> - The recognition and inherent rewards for participating in such models in
>   IETF lacks quite the same level of prestige from many people's
>   perspective.  Basically, "it's not sexy".
> - Review of models requires not only the depth of experience in the thing
>   being modeled, operations for that thing (which may not have practices if
>   it is new) but also strong familiarity with the modeling language and its
>   semantics.
>   + Yang is new.  That's going to cause some bumps as people learn.
>   + As MIBs have demonstrated, sometimes it's less whether something is
>     syntactically or semantically correct, it's whether it follows a
>     best-practice style.  Those styles are still evolving for Yang.
>
> If the above hold true, the likely fixes are:
> - Get many more people familiar with Yang.
> - Make sure as best practices evolve that they are rapidly and widely
>   disseminated.  (See rtg-yang-coord as one example.)
> - Provide the same level of prestige for doing this hard work as for doing a
>   new protocol.  "Egoboo"/"Karma" is a valid currency in volunteer run
>   organizations such as ours, even though most of our membership tends to be
>   sponsored via their jobs.
>


I also would like to clarify that my mail was in no way against the
Yang work being done in i2rs, I have great respect in the draft
authors, they are doing great admirable work.

Regards,

Behcet
> -- Jeff
>
> _______________________________________________
> i2rs mailing list
> [email protected]
> https://www.ietf.org/mailman/listinfo/i2rs

_______________________________________________
i2rs mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/i2rs

Reply via email to