Why can't we elect the WG chairs? Why can't we elect the ADs?
When the IETF pays for the 60% (80%, 100%, take your pick) of
an AD's salary, they can elect ADs. Unfortunately, the
current system is heavily biased towards keeping existing ADs
- who, like career politicians, can secure financial support
from their employers for continued participation based on
their current position. Perhaps it's time for term limits ;-)
I have two problems with the current situation, the first is that the
ADs have very much less authority than their role requires. It is simply
not possible to have authority without accountability and the whole
point of NOMCON is to ensure that no AD or IAB member is ever
accountable to anyone. The committee that appoints them will not be the
committee that reappoints them.
The second is that there is no way for the membership to tell the IESG
and IAB that they should stop dealling in the minutiae and instead focus
on the real problems facing the Internet which have nothing to do with
whether a draft that will soon be forgotten is well written or not.
I want the IAB to be giving architectural leadership. The only way that
is going to happen is if people who are writing specifications feel that
they have a stake in the IAB decision making process. Architectural
advice is not very useful when it amounts to 'don't do it your way' and
the alternative is dependent on deployment of an entire new
infrastructure like DNSSEC, IPv6 or even deployment of new DNS RRs.
There is a complete discontinuity in the style of 1980s era IETF
architecure and present day IETF architecure . In the 1980s era the
architectures were designed to empower and encourage experimentation.
Today everyone seems to have lost their nerve, the thing is too big and
too scary to risk breaking. So instead of there being an empowerment
process there is a gating process and the IESG considers its role as
being a gatekeeper.
The Web is successful because it is designed to circumvent authority and
gatekeepers, including its creators. Everything is a URI, there are no
registries of magic numbers whose permission has to be obtained.
At present the IAB is trying to promote an internet architecture in
which each new application of the DNS will require deployment of new DNS
servers. It's an entirely impractical architecture that is being
rejected in favor of various ad-hoc TXT records. The principled
architectural approach here would be to propose one off changes to the
DNS that would eliminate the need for new RR deployments in the future.
But that would result in a loss of control and so instead we have the
ridiculous situation in which everyone uses a TXT record in the sure and
certain knowledge that this will be the mechanism that is used in
perpetuity but also requests a DNS RR which they know will never replace
the TXT hack and whose only real point is to get the drafts past the
The result is the worst of all worlds, there is neither control nor
architecture. Each protocol will implement TXT security policy records
slightly differently. What could have been a principled architecture
becomes an ad hoc one that is used in practice and an ideal one that
like IPv6 everyone will pay lip service to the theoretical benefits of
without taking the action necessary to deploy.
Ietf mailing list