ietf
[Top] [All Lists]

Re: registries and designated experts

2012-06-12 13:14:18
On 2012-06-12 17:31, SM wrote:
Hi Peter,
At 07:19 12-06-2012, Peter Saint-Andre wrote:
By my reading, the "happiana" discussions [1] over the 12+ months have
led most participants to the conclusion that registration does not imply
standardization, and that it's not the role of the designated expert to
act as a gatekeeper with respect to the technical merits of the
technologies that trigger registration requests. It might be good to
have a wider discussion about the purpose of registries and the role of
designated experts, but IMHO it's not correct to conclude that a
technology is acceptable just because the designated expert didn't
object to the registrations related to that technology.

I'll +1 the above.

In a recent review the path followed by the draft is Standards Action
whereas the assignment policy is Expert Review.  Explaining to the
authors that they should not use the assigned value isn't a worthwhile
effort given that they have already been through the gate to get the
value.  The Designated Expert did his job; that is to see that the
requirements were met instead of acting as gatekeeper.  If you reject
assignment requests people will find it simpler not to register the
values.  If you accept the request people might consider that the
specification is fine.

The reasons provided for managing a namespace are:

  - prevent the hoarding of or unnecessary wasting of values

  - provide a sanity check that the request actually makes sense

  - interoperability issues

The above is at odds with standardization.  The last reason does not
apply for Expert review.

I don't understand that statement. RFC 5226 says, in Section 2 about
"Why Management of a Namespace May Be Necessary":

"  A third, and perhaps most important, consideration concerns potential
   impact on the interoperability of unreviewed extensions."

One of the specific considerations for designated experts in section 3.3
is

"      - the extension would cause problems with existing deployed
        systems."

It seems clear that interoperability is a primary concern for any
expert review.

    Brian

<Prev in Thread] Current Thread [Next in Thread>