RE: RFC 3929 on Alternative Decision Making Processes for Consens us-Blocked Decisions in the IETF (fwd)

> From: [email protected]
> [mailto:[email protected]]On Behalf Of Matthew Elvey

> Generally applicable comments on 3929:
> 1)This scheme would put excessive power in the hands of the few: the 
> Chairs, who are appointed (indirectly) by the elite few who can pay 
> non-trivial sums of money to attend lots of IETF meetings, such sums 
> generally provided by large corporations with entrenched 
> interests. 

I don't think that you really identify the problem here. I think that
you certainly have a point on the cost issue. Even though IETF does not
have a membership cost like OASIS or W3C the cost of attending meetings
is a significant barrier.

But I don't think that the right to participate in NOMCON is that 
significant. NOMCON is deliberately constructed in a way that makes
the return of incumbents all but inevitable. If you choose ten people
at random from an organization the chances that they will agree on
upsetting the status quo is very low.

Its exactly what you expect from a group of academics, its a tenure
system. Its also what you would expect from a group of engineers, an
environment where accountability and deadlines are always secondary
to idealized technical perfection.

I do not beleive that the outcome of MARID would have been any different
under any other chairs.

I do however think that the outcome would have been likely to have 
been better in a system where the group choose its own chairs rather
than having them imposed by the exstablishment. This is not because I
think the chairs would have been different, I suspect that the same
people would have volunteered and I doubt that the election would have
been contested, but the group members would have felt that they have
a stake in the management of the group. People are more likely to accept
decisions made by people they feel they have a stake in choosing - even
when they voted for somebody else.

> MARID specific comment:
> What happened here doesn't seem to be a failure, but rather was a 
> success in the following sense: There was not sufficient maturity in the 
> proposals for a rough consensus to be reached that the proposals were 
> adequate to meet the bare minimum of goals that they needed to meet.  

I disagree. SenderID/SPF is currently in deployment and is working just 
fine in production. Maturity was not the problem. 

I think that there was a mismatch of expectations between people who
thought that the purpose of the group was to agree on a refinement of
an existing specification that had already achieved a very large degree
of industry buy in and those who thought that it was an open forum to
propose their alternative idea.

I have never been in a standards group of the second type that has
succeeded. In every case that a group has suceeded the starting point
has been a specification that was already largely complete.

> The process brought the fact to light that the proposals were 
> inadequate.  Many of us weren't aware of this early on.  3929 
> seems to 
> assume that a failure to reach a rough consensus must mean there's a 
> problem in the process, when in fact it could be in the proposals.

I think that as a practical matter the issue of acceptable IPR terms
should not be discussed at the working group level where the bargaining 
leverage is least.

I found it practically impossible to follow the alledged discussion
of technical issues due to the constant heckling and deliberate
recycling of false positions. The lack of consensus inside the group
did not reflect the solid industry consensus for SPF/SenderID that
had already been established before MARID started.

It may not seem fair but an engineer who is representing the issues
of an ISP with 20 million customers has a rather larger stake in
the outcome than an individual contributor responsible to nobody but 
themselves.