[rfc-i] Discussion about Committees and RSE models
dhc at dcrocker.net
Tue Jan 18 09:53:35 PST 2011
>> On 2011-01-18 14:06, Bob Hinden wrote:
>>> I have read several comments on RSE model stating that committees can do
>>> anything right, make bad decisions, can't be trusted, and are guaranteed
>>> to not have the required expertise.
Except for the last item in the list, none of those things have been asserted
Statements have indeed been about likely expertise, but they have predominantly
been about the nature of committee dynamics and long-standing observations about
better-vs-worse roles for committees. Most of the statements made about
committees are extremely well established. Although the IETF often chooses
unusual structures and styles, it has not invented a new range of group
dynamics, skills, roles or outcomes.
>>> I am confused by this in the IETF context as everything is done by
>>> committees. Working groups are committees, the IESG is a committee, the
>>> IAB is a committee, the IAOC is a committee, the Nomcom is a committee,
>>> the ISOC board of trustees is a committee, directorates are committees,
>>> etc. I don't claim that any of this is perfect, but I am confused by
>>> the statements that committees relating to the RSE model won't work, when
>>> they appear to work in other areas of the IETF.
The core confusions, here, are about how work and decision making actually
happen, when a committee is part of the mix.
Sometimes within the IETF, there really are groups of people who operate as an
integrated whole, throughout the entire process of getting something done. They
formulate the problem statement, explore possible solutions and specify the
choice. But these are actually pretty rare.
The more typical model is that a committee provides a base of authorization for
actions by delegates, review of proposals and specifications, and approval of
the final form. (Please note: that the actual work might be done by one or a
few members of the committee does not mean that the committee did the work.)
Within the IETF and most other places, committees tend NOT to take initiative,
develop the details, nor formulate the specification. They review and approve.
Sometimes they assist with what problem-solving research calls a
"brainstorming" phase of the process, to explore innovative alternatives, but
that's actually pretty rare within the IETF context.
>>> I understand why a committee needs to have the right expertise to work,
>>> but I don't understand the unsubstantiated claim that they are bound to
>>> fail. If we don't put the right set of people on the committee it will
>>> fail, but that also true for hiring an RSE. If we don't hire the right
>>> person with the right capabilities, then that won't work either. But
>>> that's not a reason to not hire an RSE.
Personnel selection never comes with guarantees. However your statement
contains an interesting implication: Whatever the risk of hiring the wrong RSE,
it is multiplied by the need to "hire" more people to properly form the
committee. (Cumulative risk...)
In addition, at best the committee members have the task as a secondary
priority, mixed in with various other, unrelated tasks, usually deemed far more
important. In contrast, the RSE has the task as primary and is being paid for
this /specific work. This means that it is far easier to demand focus and
accountability in a timely fashion from the RSE than from a committee.
The standard management rule is that anything that needs to get done needs to
have someone accountable for making it happen. It is possible to run a
committee that way, where members of the committee have specific items assigned
to them. However at that point the work is no longer being "done" by the
committee. It is being done by the accountable person (possibly with the help
of the committee.)
An added benefit to a two-stage model that has an RSE as a legitimate leader,
and the oversight committee providing... oversight, is that there is better
independent review. If the oversight committee is the primary worker, that
leaves a far more uninvolved and distracted IAB as the only review path.
If folks choose a model in which the RSE can take no initiative and has no
authority to resolve anything directly, we will get a largely useless drone.
If we want someone with serious capabilities, we need to define a job for them
that legitimately calls on those capabilities.
It is one thing to require that major policy changes receive advice and consent
from an oversight committee and quite another to require that the committee be
responsible for initiating everything and that they approve everything.
If folks really mean the former and not the latter, there needs to be clarity
about this, in how the job is defined. The clarity comes from details, not
broad conceptual assertions.
More information about the rfc-interest