Standards Classification and Reality Problem Statement (was Re: Not a problem statement [ was Re: Killing old/slow groups - transition thinking)

James Kempf kempf@docomolabs-usa.com
Fri, 13 Dec 2002 09:35:29 -0800


> Possibly that's because we don't know which problem we are trying
> to solve; see the name of this mailing list.
>

The problems we are trying to solve, as I see it, are the following;

    - Increasing formalization of the WG ID "status" as being preordained for
standards track, despite official IETF policy
      and ID template language to the contrary, as a result of:

        . Publicity by document authors and editors that they have achieved WG
ID "status" in an attempt to lock in their
          authorship and basic ideas.

        . Implementation and interoperatability testing of WG ID drafts, based
partly on the assumption by others as a result
          of authors' media campagin that the draft, will, in fact, become
standard, and also in some cases to the authors funding
          for outside implementations in an effort to get a broadly distributed
code-base buy-in that will make it hard to
         remove the draft from WG status.

        . The underlying market need, which is behind the standardization effort
in the first place, for a solution, and, in many cases a
          solution right away. People doing products are willing to put up with
some churn in the spec if they feel the need is great
          enough, what the IETF considers to be formally the case is of little
importance, especially the DS and S classifications.

 - The result of the above is the hardening of the design by the time the ID
becomes WG draft around the original idea of the authors
      with little opportunity for WG input, especially with respect to
architectural changes that might make the resulting design a better
      fit with the Internet architecture in general.

  - Lack of requirement for interoperability when a draft becomes PS means that
many years may go by with informal agreement
     about interoperability or even a lack thereof in certain areas, because for
many IETF standards, PS is in reality the final phase
     and really what most customers of IETF standards care about.

    - Draft Standard and Standard have little or no meaning for customers of
IETF standards. They are
      kind of a historical recognition,  and therefore of little practical use
for day to day engineering.

In summary, the process has become considerably more front-loaded since the
original standards classification scheme was developed.

                        jak