[Date Prev][Date Next] [Chronological] [Thread] [Top]

Re: [Gen-art] A *new* batch of IETF LC reviews - Sept 10th

At 07:22 PM 9/10/2005, Joel M. Halpern wrote:
>[This is a gen-art review prepared for the IETF chair by a pseudo-randomly selected member of the gen-art team.
> This review is intended to help the IETF Chair / General AD.  It is provided directly to the other affected individuals to
> save confusion.]
>
>This document is nearly ready for publication as a BCP.
>
>Question: Based on the issues that have been raised about reviews recently, should this document be more specific as to what kinds of issues the expert reviewer is to look for in each kind of item to be reviewed, and why?  (I am not arguing with the WGs choice of mechanism, just trying to head off trouble that can be foreseen.)

As the IESG-designated 'Expert Reviewer' for LDAP parameters,
I don't forsee there being significant trouble due to the
level of guidance given.  I think the BCP64 adequately covers
specific issues.  While there might be some general issues:
"do no harm (to the internet)"), "focus on the request, not
the specification", etc., I don't see the need to discuss
these in BCP64 as none of its specific to LDAP registrations.

Its my view that one of the key reasons we rely on IESG
selected experts is to reduce the need to be overly precise
in the guidance.

I suggest that we focus only on LDAP-specific issues.
If you have any, please elaborate.

>Question: Is the last sentence of the second paragraph of 3.4 intended to call for expert review of all descriptors, or only of descriptors referencing different object identifiers from an already registered descriptor with the same name?  While I find this verbiage awkward in the other places it is used, this particular usage is less clear than the others.

The intent of the paragraph:
  While the protocol allows the same descriptor to refer to different
  object identifiers in certain cases and the registry supports
  multiple registrations of the same descriptor (each indicating a
  different kind of schema element and different object identifier),   
  multiple registrations of the same descriptor are to be avoided. 
  All such registration requests require Expert Review. 

is to note that while the registry supports non-unique descriptors
(but unique within their category of use), registration of a
non-unique descriptor is to be avoided and, hence, requires
expert review.  That is, if you want to register e-foo
both as an attribute type descriptor and as an object class
descriptor, the policy is Expert Review (instead of
First Come First Server for other e- descriptors).

Please feel free to offer a specific suggestion (e.g., text) to
help clarify this paragraph.   The current text is the best I
could come up at the time (and at this time).

Editorial/nits issues noted will be appropriately addressed.

Thanks for the review, Kurt.