To the IESG:
On Wed, Mar 10, 2010 at 05:43:12PM -0500, Russ Housley wrote:
The IESG has received an appeal. It can be found here:
http://www.ietf.org/iesg/appeal/morfin-2010-03-10.pdf
I have read the document, though I cheerfully concede that some of the
text eludes my understanding. I was an active participant in the
IDNABIS WG. I supported the WG document set.
As near as I can tell from the PDF, the actual substance of the appeal
is entirely in section 3. There are scanty arguments in favour of the
substantial appeal items. The rest of the document comprises a great
deal of background material, and an argument that a lot more work is
needed on IDNA in order to meet real users' needs.
The opening sentence of section 3 says, "This is why this appeal does
not concern the IDNA 2008 document set, as approved by the IESG, which
is now of prime stable importance when considering the Internet
architecture from a user perspective, but also concerning the way the
IESG has approved this IDNA2008 document set . . . ." As near as I can
tell, that says that it is _not_ an appeal of the document set itself.
It is instead an appeal that the documents were not published with
disclaimers attached.
The bullets following the first sentence seem to support to support my
interpretation. I think the bullets are preposterous. If
generalized, they effectively argue that the IESG needs to insert or
obtain disclaimers about a protocol for the architectural consequences
of the protocol, or whenever a protocol is somehow incomplete. Two
major premises are offered for the conclusion. First, we have a claim
that the WG failed to reach consensus on a topic on which the
appellant thinks the WG should have done; this is not a generalizable
principle. Second, the appellant claims that the IESG has a
precautionary duty that makes the disclaimers necessary.
Let us start with the easy premise. Given that there are sometimes
disclaimers attached to documents, and given that the IESG lies on the
path to publication of documents on the standards track, it is either
trivially true or false that the IESG has a precautionary duty when it
finds that a document that otherwise is the product of IETF consensus
is dangerous to the operation of the Internet. It is trivially true
in the sense that the judgement of the IESG is exactly what we have an
IESG for, and part of the process of judging is ensuring that a given
document does not entail harm. It is false in the sense that _any_
change on the Internet has the potential to cause trouble. If there
is a strong precautionary duty, then in fact every document that makes
any innovation of any kind would need to come with a warning saying
that the innovation has not been tested in ubiquitous deployment on
the public Internet.
It is that "false" meaning that shows the deep confusion underlying
the rest of the appeal. For the appeal at bottom assumes that RFCs
are near-scriptural in their force, as though the Internet proceeds
from the Word made RFC. The plain fact is that, if the IESG always
needed to obtain or insert disclaimers whenever there are
architectural consequences or whenever the protocol was incomplete,
then such disclaimers would become new decorations of nearly every
RFC. It is only when there is significant danger of harm that such
disclaimers are needed, and not at other times.
The standards maturity levels of the IETF are the real mechanism to
provide the indicators the appellant seems to want: new, untested
protocols do not come out as STANDARD, but at less mature levels.
That very lower level of maturity is, all by itself, an indication
that there are possible as-yet-unseen issues with the protocol. The
only test of whether there will be issues is experience. In new
protocols, that experience is by definition not there.
This is why the argument from the failure of the IDNABIS WG to reach
consensus on a mappings document does not provide the support the
appellant needs for the conclusions. If a WG cannot reach consensus
on exactly what to do about a given area, that is nowise proof that
one thing, another thing, or indeed anything ought to be the right
answer. It is instead evidence that we don't know where to go yet,
and so we have to try things out and learn what the best thing to do
is. The very idea that protocols should spring fully formed in the
RFC repository is completely mistaken.
Ignoring the background material, the rest of the appeal document is
an argument that more work is needed in the IDNA area. If more work
is needed, then getting the documents out is an important first step,
and not something to be lamented. So the "more work needed" argument
does not support a conclusion that publication without a disclaimer is
harmful: anything that is a good foundation for additional work
cannot, by definition, be itself bad for the Internet.
Given all of the above, I believe the IESG should reject the appeal.
Best regards,
Andrew
--
Andrew Sullivan
ajs(_at_)shinkuro(_dot_)com
Shinkuro, Inc.
_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/ietf