[rfc-i] The RFC Series Manager

John C Klensin john+rfc at jck.com
Thu Jan 22 22:02:52 PST 2009

--On Thursday, January 22, 2009 21:22 -0800 Jim Schaad
<ietf at augustcellars.com> wrote:

>> Ignoring, for the purposes of this note,  the independent
>> submission side of the story --while the Model and RFI spend
>> considerable text on it, it is really just a matter of how one

> Actually I think the case of the independent submission stream
> is even worse as compared to the other streams.  Previous
> discussions on this list appear to state that the Independent
> Stream Editor has the ability to rely on the production house
> to do serious editing work on documents.  (It is assumed that
> documents coming in from the other streams meet some level of
> competence.)   This could easily lead to the production house
> complaining that they are doing more work that was bargained
> for as the documents are passed through the editorial process
> multiple times.

Absolutely.  I didn't explain why I was excluding the
independent submission stream very well.  

I was looking at the situation from the perspective of the RSE.
The RSE oversees (or some other word or phrase) the Production
House and Publisher, but has no other relationship with the ISE
other than receiving documents and setting norms for what is
acceptable to be received.  From that perspective, the ISE and
independent submission stream are no different from any other

On the other hand, the further we move away from the RFC Editor
function being able to say to a stream authorizer "this is just
not good enough (or coherent enough, or clear enough) that it
can be edited with a reasonable level of resources, the more
incentive (at least in principle) a stream process has to dump
documents of very poor editorial quality on the RFC Editor on a
"here, you straighten it out" basis.    Interestingly, under the
present, unified, model, the editing function is least likely to
get an unacceptable document out of the independent stream
relative to anything -- such a handoff would involve the left
hand of the RFC Editor imposing an unreasonable burden on the
right hand (or vice versa).  Under the new model, the ISE has no
such disincentives, as Jim points out.

In practice, it is hard to predict how this would actually work
itself out.  For example, to the degree to which the IESG, IAB,
or IRTF/IRSG feel a need to iterate on documents until every
detail is correct -- well beyond "some [minimal] level of
competence" -- I, like Jim, would not predict many documents
from those directions that were editorial disasters.  And I'd
expect the ISE to exercise an acceptable level of care if only
to be sure that what was approved for publication bore a
reasonable resemblance to what is actually published.

However, Jim makes a point I tried to suggest some weeks ago.
One of the points of fragility of the whole model is that any of
the streams, or the RSE, can impose costs on the Production
House that the latter could not accurately estimate at proposal
submission or contract negotiation time, and can do so by
changes of internal policy that are not subject to review by the
Production House or anyone else.   Coming back to the note that
started this thread, this is one of those situations in which
the choices are to designate an RSM or to just let one happen...
again, unless the IAB wants to make itself available to be part
of the discussion in real time.   I don't know if he will agree
or not, but I believe that Paul Hoffman's recent observations
about design and implementation of tools is a special case of
this more general problem.


More information about the rfc-interest mailing list