[rfc-i] The proper role of an RSE
dhc at dcrocker.net
Wed Jan 5 04:46:35 PST 2011
I hear reports that the IAB will be focusing on the question of the RSE today
and thought it worth lodging a simple statement of support for the proposal that
is on the table.
But unlike most folk in our community, I cannot stop with such a brief, direct
So, some elaboration...
For any interesting topic, a variety of reasonable solutions can be developed.
For any reasonable solution to an interesting topic, there will always be much
that can reasonably be debated, tweaked or replaced. The current proposal is no
I think the current proposal has three core characteristics that are essential:
1. Incremental change to a model that has functioned well for 4 decades.
Major change to the organization or management of a critical service always
carries major, unintended -- and usually undesirable -- side-effects. (For
reference, that's so well established in in management and organizational
behavior circles that it is a cliche.) When there is a compelling reason for
major change, the side-effects might be justified. I have not heard any
compelling reasons put forward to justify major changes to RFC Editor
management, and the current proposal correctly stays with the established model.
2. An RSE who serves as a central point of responsibility and authority for
/integrated/ management of the RFC Editor.
The job of the RSE is considerably more than that of a senior technical
editor or of a supervisor. Management is not the same as supervision. The RFC
Editor has a complex set of document sources, a complex set of document
consumers, a continuing need for enhancement and revision to its operation, and
periodic issues that require policy and design skills that go far beyond
competent editing. They require skills at anticipating and responding to
long-term issues, rather than repetitive performance of immediate tasks. In
addition, these activities cannot be combined with direct operations or the
direct operations activities will suffer. So, different skills and different
focus are needed.
3. Revision to the oversight function that adapts to current pragmatics.
The IAB should not be expected to provide close oversight to the RSE; it has
neither the time nor -- from what I have heard and seen -- the interest in the
effort that would take. And while the IAOC is essential to particular
activities of the RFC Editor, it's role in the IETF does not seem to be to be a
good fit to have primary responsibility of oversight for the strategic work of
the RSE. A small, specialized body of folks from our community can provide that
oversight best, IMO. (Some of what I'm describing differs from the specifics of
Glenn's proposal. But as I said, a good proposal can always be tweaked... But
while I believe the variance I'm suggesting would be an improvement to the
proposal, I do not think it as essential as items 1 and 2, above.)
My view of the need for an RSE who has overall responsibility for the entire RFC
Editor is not a desire to return to an earlier, golden age. The times, the
people and the community are all quite different. (Hence the need for the
oversight committee.) Rather, the requirement comes from a basic, conservative
model of management appropriate to any critical operational service. I don't
think that we are clever or diligent enough to make a radically different model
work, nor do I think there are any benefits in such change.
To the extent that there are claims that we've recently been operating under a
radically different model, I'll note that while the daily work has indeed been
performed well, the longer-term work of the RFC Editor has not, since no one had
responsibility to worry about it...
More information about the rfc-interest