Olaf M. Kolkman writes:
+ RFC3??? AXFR clarify to Draft Standard.
Replacement charter item: ``See whether consensus can be achieved on
AXFR clarifications.'' I fully understand the desire to clarify how the
DNS protocol works---but that doesn't justify pushing a horribly flawed
document to ``Proposed Standard,'' never mind ``Draft Standard.''
At least seven people have gone on record as objecting to axfr-clarify;
see http://cr.yp.to/djbdns/axfr-clarify.html for details. The BIND
company admits that axfr-clarify does not correctly describe the
behavior of most DNS servers on the Internet.
My web page http://cr.yp.to/djbdns/axfr-notes.html contains an accurate
description of the AXFR protocol, quite different from axfr-clarify.
Along the way, it identifies nine clear errors in axfr-clarify: six
points where axfr-clarify (in violation of RFC 2119 section 6) prohibits
existing working behavior, and three points where axfr-clarify creates
new interoperability problems.
What the BIND company has given us is not, in fact, ``clarification.''
It is BIND 9 documentation fraudulently posing as clarification. It is
blatantly biased in favor of one vendor---the same vendor that produced
the document, that has generated most of the messages supporting the
document, and that previously paid one of the WG chairs (Gudmundsson) to
work on its software.
The problems with axfr-clarify, both in content and in procedure, have
been thoroughly documented. Anyone who wants to ignore IETF documents---
even the occasional high-quality ones, the ones that are important for
interoperability---simply has to point to the progress of axfr-clarify
as an IETF ``standard.'' A legitimate standards organization would never
have allowed axfr-clarify to get anywhere.
---D. J. Bernstein, Associate Professor, Department of Mathematics,
Statistics, and Computer Science, University of Illinois at Chicago