[rfc-i] [IAB] Comments on RSE models
Joel M. Halpern
jmh at joelhalpern.com
Fri Jan 14 18:08:30 PST 2011
There is a larger difference in the two models, I think.
However, before I go there, I want to elaborate on one of Dave's
comments. If whatever structure we propose is going to work, it will
likely work by evolving the actual relationship between the REOC and the
RSE. By itself, that is fine.
However, if the starting state and the ending state are too different,
we will end up hiring the wrong person for the job.
Of course, if we do badly enough, it will not be able to evolve to
sufficiency, and things will gradually degrade in all sorts of subtle
ways until it is enough of a mess for folks to really notice, and then
it will be VERY hard to fix.
So, to return to what I see as an important difference between Paul and
Glenn. Paul, assuming I am reading things right, wants the committee to
be responsible for making all policy decisions. The committee is
responsible for leading community discussion. And for judging that
discussion. The RSE then does what the committee says. In Glenn's
model, the RSE is responsible for those activities. He then brings his
conclusions to the committee for their advice and consent. If they
think he is wrong, they fix it. They still have the responsibility.
The two structures, while they are similar, operate very differently in
practice. The later draws on the strength of committees for providing
additional perspectives and careful review. The former requires the
committee to perform management, which much evidence from both within
the IETF structure and outside says is between a bad idea and ineffective.
Joel M. Halpern
On 1/14/2011 7:59 PM, Dave CROCKER wrote:
> On 1/14/2011 3:56 PM, Paul Hoffman wrote:
>> On 1/14/11 2:53 PM, Olaf Kolkman wrote:
>>> When I mean 'X implements' I do not mean that
>>> X does the hands-on work herself but that she is the responsible
>>> person for making sure the work happens and is properly budgeted and
>>> In your model, who is is responsible for implementation in that
>> The REOC. In my model, the REOC is responsible for overall series
>> management and
>> stability and creates policy for the RFC series, and the RSE..."takes
>> direction from RSOC on initiatives", which would cover new, ongoing,
>> and changed
>> initiatives such as "a big question got answered".
>> However, the
>> difference in the models is that in mine, if the RSE fails at it or
>> takes too
>> long, it is the REOC that is responsible; in Glenn's, I think the REOC
>> does not
>> have the responsibility to make sure the RSE is doing what the
>> community asked.
> One nice thing about this exchange is that it is discussing the work and
> not the labels.
> There does seem to be comfort with having an REOC. And it appears that
> there is comfort with having it hold a significant degree of authority.
> I'm not sure as sure but it also appears that folks are comfortable with
> having the REOC be something like the top of the RFC Editor chain of
> command, except with the IAB being available for appeals (dispute
> That is, I think that both models actual rest a significant degree of
> ultimate authority in the REOC.
> However, I'm not really able to tell what the pragmatic responsibilities
> and authorities actually are for the RSE in either case.
> No matter what is formally specified, the details of the relationship
> between the REOC and the RSE will be subject to their own adjustments.
> The questions are where things start from and what the likely "dynamic
> range" is? Get someone who only knows how to edit documents and they are
> not likely to know much about running a technical document series. Get
> someone who has no hands on experience with producing documents, and
> they aren't likely to have much insight into the pragmatics
> On the assumption that we do not want a mechanical drone in the job,
> there needs to be some meat on the bones of the RSE's job. It has to
> have real challenges and an ability to effect real improvements.
> So, what can the RSE do on their own initiative, without first checking
> with the REOC?
> * Can the RSE resolve a prioritization issue for the Publication Center,
> without checking with the REOC?
> * Can the RSE resolve any other Production or Publication issues,
> without checking with the REOC? If so, what kinds?
> * Can the RSE initiate work on a policy, without first get approval from
> the REOC? (But notice I said "initiate work" rather than "decide".)
> What other sorts of questions about specific activities will help to
> distinguish between the two models and help understand the kind of
> person (background, skills, style, goals) is likely to apply for the RSE
> job and do it well?
More information about the rfc-interest