[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Fwd: Handle System RFC Note

Bob,
First, no one need have 30 years of experience in the
Internet industry to question a decision of the IESG; anyone can
do it at any time (and many do!). The decision taken by the IESG to ask you
to insert an IESG message to the handle drafts was done under the
auspices of 2026, section 4.2.3, and like any other decision under
the auspices of 2026, it is subject to appeal. If you or Bob Kahn
would like to appeal the decision, then you can, but since this
is a decision of the IESG at this point, I don't think I personally
have the right or ability to reverse it. Note that I have cc'ed
the IESG on this response in order to allow others to make any
other points.
Second, if Bob Kahn or anyone else makes a technical
contribution to the IETF process, I do my utmost to hear their
points. Since I started working to move these documents forward,
Bob made no technical contributions to the process prior to the
IESG's decision. I worked with the authors, and got agreement from
them; since authors represent themselves as individuals in the IETF process,
I believed and continue to believe that this is appropriate. I understand
from the email below that Bob had a negative reaction to the wording
the authors agreed to. The email I saw is not a technical contribution
related to the work, but a statement about his reaction to the IETF note.
If there have been technical contributions, I have not seen them. Obviously,
no one can take into account technical points which are not made to them.
Third, you have asked, in essence, why this work is "something that
conflicts with, or is actually inimical to, an established IETF effort".
Possibly our effort to make the IESG note polite and acceptable to
the authors has taken to much of the conflict out of the picture.
If so, I regret the lack of clarity, but this work is clearly in conflict
with established IETF efforts. The only oddity here is that it has
been in conflict with the established IETF efforts so long that the
working groups involved have concluded their work and shut down.
This was work extensively discussed in the IETF, including in
the URI and URN working groups; the experience there was that the
working groups' efforts to bring the handle system into the larger
system resulted in changes in the working group documents but
no movement on the part of the document authors. As it stands
now, the documents represent the view of a "loyal opposition" and
contain views of how resolution and identifiers should or could work
on the Internet that are clearly not the consensus view; as such, I personally
believe that they should be published, but only if clearly flagged to
indicate to the reader that the views contained do not represent how
the rest of the IETF understands these issues. In other words, don't
expect this to work on the big I Internet; it works only in the context
of the handle system as defined by CNRI and will not be integrated into
any of the other systems which use or depend on URIs, URNs, LDAP
identifiers, or DNS names. This is not something that can be clarified
in the IETF documents, since those documents are complete and the
working groups closed, so it must be included here. Without that, I
would have recommended that these documents not be published at all.
Fourth, my use of the term "war of attrition" was meant to
be evocative rather than exact, but I stand by it. In my discussions
with the authors, it has been clear that they are not working in the context
of the larger community to develop whatever system would be best for the
Internet, but are focused on the advancement of their system. One of
the tactics here has clearly been simply to outlast the objections and to demand
that the same points be explained to them over and over, in the hope that
the individual involved will give up and accede. Having held on until the
working groups were closed and then exhausted this tactic in the IESG, it
appears that they are now taking it up again with the RFC Editor. I have
personally now generated more than four thousand lines of prose on this
topic. That's far more than it is worth, to be honest, but the result of this
should not be that they succeed in putting forth their views of the relationship
of their work to other systems of identifiers without appropriate marking
that the community has considered their points and disagreed.
regards,
Ted Hardie



At 2:32 PM -0700 8/5/03, Bob Braden wrote:
*>
*> Hi Bob, RFC-Editor folks,
*> To be frank, I am beginning to view this as a war of attrition
*> rather than a technical discussion. I've had this exact conversation

Ted,

I am still trying to figure out what is going on here. I understand
that there are strongly-held views on both sides, but I have trouble
understanding why there should be two sides at all. "A war of
attrition" seems a little strong.

*> with the draft authors no fewer than four times since I came onto the
*> IESG in March, and the substance of this discussion has been going on
*> in various working groups within the IETF for nearly ten years.
*> The URN and URI working groups made substantive changes to their documents
*> in order to aim for convergence, but the CNRI folk have continued to
*> see their work as something far broader and indeed, something that could
*> subsume nearly every other form of resolution on the Internet. That's clearly
*> not a consensus view, and their descriptions of the current system of
*> identifiers and resolution is very different as a result of that take on the

I understand that that is Bob Kahn's view, and after 30 years of
experience with his technical views, frankly I have to take them very
seriously. Which is not to suggest that he is omniscient, but if you
have been right as often has he has, he deserves a hearing. Among
other things, he literally founded the Internet technology which is the
basis for my employment and yours.

But I don't get the relevance of its not being the consensus view. The
issue is publishing Informational RFCs as an individual submissions.
That raises no issues about consensus.

However, I do look forward to better understanding the precise
technical issues that separate the handle system from the current IETF
consensus view.


*> world. The aim of the note is to highlight that difference and I remain
*> convinced it is an important part of the context of this document.
*> Further, a message from Sam on July 3rd confirmed that he had
*> discussed this wording with the other authors and they had agreed it
*> was acceptable, after I toned this down from the wording Patrik
*> had proposed. I think Bob's message asking for a right to edit it
*> is contrary to that agreement, and I don't think it is appropriate.

Well, you have to understand Bob's role here...

*> The document
*> authors and the IESG agreed on this compromise wording to set the context,
*> and, frankly, I would have drafted a DNP note without it. I urge you
*> to publish
*> this document with the note as expeditiously as possible, so that we can move
*> past this long-fought disagreement.

I talked at length with Bob earlier this week, and we agreed that he
will come by ISI here in Marina del Rey on Tuesday Sept 23 to discuss
these issues with us. We want to invite you to attend, if you can.

There is an alternative ... it turns out that Bob will be at Qualcomm
on Sept 22 (in fact, that is his reason to be on our coast). You could
arrange to chat with him for a few minutes on the afternoon of Sept 22.

I understand that you have devoted significant effort to this issue
already, and I can only ask for your continued help in resolving it.
The RFC Editor thinks it is important to get this one right.

Bob Braden

*> Thanks for your attention,
*> Ted Hardie
*>
*>
*>
*> >X-Sender: rkahn@newcnri.cnri.reston.va.us
*> >Date: Fri, 18 Jul 2003 01:56:10 -0400
*> >To: ssun@cnri.reston.va.us
*> >From: Robert Kahn <rkahn@cnri.reston.va.us>
*> >Subject: Handle System RFC Note
*> >Cc: kahn@cnri.reston.va.us, thardie@qualcomm.com, braden@isi.edu
*> >
*> >Sam,
*> >
*> >Having seen the traffic on this subject, I now understand that
*> >inclusion of a message in the RFC is important to the IESG and even
*> >though they don't usually append anything to informational RFCs,
*> >they feel the need to do so here. I'm prepared to accept that view
*> >for the present.
*> >
*> >However, I would like to suggest a minor compromise that may also be
*> >acceptable to move this forward. Namely, that the first two
*> >sentences of the note be added but the last sentence of the note be
*> >deleted. CNRI does not believe that the Handle System describes an
*> >alternate view of how namespaces and identifiers might work on the
*> >Internet. Indeed, the Handle System can support resolution for a
*> >broad class of identifier systems including those characterized
*> >within the IETF as URIs, URNs or even URLs. I would argue that is a
*> >broader view, not an alternate one.
*> >
*> >I have even more trouble in a sense with the notion that there is an
*> >IETF architecture for identifiers, but this is a semantic concern
*> >for me and more specifically a reflection about what is an
*> >architecture and what is not. If the IESG thinks there is one,
*> >that's probably all that counts. And since this is their statement
*> >about what they think, it has their credibility behind it. And so
*> >I'm prepared to accept it.
*> >
*> >Regards,
*> >
*> >Bob
*>