---- Original Message -----
From: "Cullen Jennings" <fluffy(_at_)cisco(_dot_)com>
To: "Christian Huitema" <huitema(_at_)microsoft(_dot_)com>
Cc: <tsvwg(_at_)ietf(_dot_)org>; "Paul Hoffman"
<paul(_dot_)hoffman(_at_)vpnc(_dot_)org>; "Chris Benson"
<cbenson(_at_)adax(_dot_)com>; "IESG IESG" <iesg(_at_)ietf(_dot_)org>; "Sam
Hartman"
<hartmans-ietf(_at_)mit(_dot_)edu>; "IETF discussion list"
<ietf(_at_)ietf(_dot_)org>
Sent: Wednesday, February 16, 2011 5:31 AM
I've been thinking more about this thread and my concerns about this draft. I
was originally looking for the draft to have advice for the expert review team
that gave them guidance on what the IETF thought was all right to approve or not
approve. It's become clear that this draft does not have that advice and is not
likely to get it in the very short term. This BCP will empower the expert
reviewer to reject or approve just about any request. Appeals are not the best
way to balance putting that power because they are incredibly corrosive and time
consuming to everyone involved. I think this thread somewhat suggests an
alternative approach for a check and balance.
Cullen
My understanding of this draft is that Expert Review only applies when "IETF
Review" or "IESG Approval" does not apply, so a duly approved Standards
Track RFC trumps all Experts.
I proposed text on 2nd February to make this clearer, and understand that this
change was accepted.
Where the request is not via "IETF Review" or "IESG Approval", then I have
less concern as to what the Expert may or may not allow.
My text is
"For most IETF protocols, ports in the User Ports range will be assigned under
the "IETF Review" or "IESG Approval" procedures [RFC5226] and no further
documentation is required.
Where these procedures do not apply, then the requester must
input the documentation to the "Expert Review"
procedure
[RFC5226], by which IANA will have a technical expert review the
request to determine whether to grant the assignment. The
submitted documentation MUST explain why using a port number in
the Dynamic Ports range is unsuitable for the given application."
Tom Petch
What do people think of the idea of: for all ports requests, the request and
the expert reviewer reposes including reason for accepting or rejecting them
need to be posted to a public email list. This seems like a simple way to help
mitigate this issue and it will help educate people writing a port request to
know what types of issues they need to address and what would be appropriate or
not.
Pros & cons of this idea?
On Feb 8, 2011, at 1:41 PM, Christian Huitema wrote:
I don't see that "public identity" (of expert reviewers) is required for
"interactive discussion".
Or would anonymous interaction fail a Turing test of some kind?
Public identity is required for reviewer accountability. It is easy to
imagine how withholding registration of some required numbers may delay a
competitor's products. The best protection against shade is sunlight.
-- Christian Huitema
_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/ietf