[rfc-i] Requested follow-up from last night's plenary

Brian E Carpenter brian.e.carpenter at gmail.com
Mon Nov 8 17:03:54 PST 2010

Hi Ted,

> "The RFC Series is the Internet technical community's official
> medium, through which it communicates with itself and the rest of
> the world."

I think I can be blamed for Glenn using the phrase "Internet technical community".
And in my mind, it does not equate to the union of the IETF, the IAB, and the IRTF.
It is much broader - it includes all the *NOGs, the RIRs, product developers,
service providers, open source developers, other standards bodies, writers of
academic text books, and probably some others I've overlooked.

I don't think it is grandiose to identify the RFC series as the medium of
communication with that whole audience. It would be helpful to have
some actual readership data, but I don't know quite how to get it.

That is, in any case, why I believe it is entirely reasonable to put
the RFC series in a different category than other mechanisms such as
I-Ds, I*TF email lists, etc.

I would strike the word "official", however.

> As written, this declares the RFC Editor to be the lead of this
> activity, with this general leadership merely "tempered by" the
> processes and requirements of the streams which produce the actual
> output.
> The focus in RFC 4844 is support and implementation, with expertise
> guiding discussion about editing, publishing, and archiving.  This
> document moves this to the general lead of this activity, a change
> I cannot agree is modest.

I agree that it's a significant change, from a wishy-washy job to
a well-defined managerial job. One of my complaints at the previous
model, and one of the deep reasons (I suspect) why the previous search
for an RSE failed, is that the previous model created a highly visible
position with plenty of responsibility (i.e. a target for rotten fruit
when things go wrong) but *without* the authority needed to exercise
that responsibility. That's a very fundamental mistake in organisational
design that, frankly, the IAB got wrong.

Glenn may not have tuned it right yet, but unless the RSE is given
explicit and clear authority to manage that which s/he will be held
responsible for, this will not work.

I remind you that the IAB's authority in this matter is very specific,
as defined in the IAB charter:

"The RFC Editor executes editorial management and publication of the
IETF "Request for Comment" (RFC) document series...
The IAB must approve the appointment of an organization to
act as RFC Editor and the general policy followed by the RFC Editor."

[RFC 2850 = BCP 39, Section 2]

That clearly gives the RFC Editor a good deal of autonomy and direct
responsibility, with the IAB having an approval role only. I don't
believe that RFCs 4844 and 5620 (Informational RFCs) are fully
consistent with the IAB's authority under RFC 2850. I think Glenn's
positioning of the RSE role is much closer to consistency with it.
And to repeat myself, IMHO this the only approach that can make the
job viable.

> The document states that this leadership would be "as it is
> practiced in a typical not-for-profit organization" along with
> specific community driven practices (seek input, foster volunteers,
> supervise according to procedures).  This is not the correct model
> for a document series 90 per cent of whose output is standards
> documents representing hard-won consensus.  Those have to be led by
> the communities producing the documents.

Indeed, but it is 100% clear that the streams each retain their
unfettered right to approve or disapprove publication. This may be
a drafting problem in the text, since Glenn's made it pretty clear
that this refers only to general matters affecting all streams
("general policy" in the IAB charter language quoted above).

> The document also describes a change to the RSAG model, which it
> describes as "marginally expanded".  In fact,the RSAG in this
> document has a major change, buried in Appendix A, section 2.
> Where the body of the document and RFC 5620 state that the RSAG is
> not responsible for hiring the RSE, this gives the constituents of
> the "search and selection committee", which includes 2 members of
> the RSAG.

You have a point there. I'm quite sure this wasn't intended to
create a cabal, but we can certainly discuss this. To be frank the
current membership of the RSAG is a bit of a by-product of history
and there's scope for reviewing it. RFC 5620 says "The RSAG members
are proposed by the Series Editor in consultation with the sitting
RSAG members, and then confirmed and formally appointed by the IAB."
Members proposed by the Series Editor aren't the ideal constituency
to provide members of the hiring committee for the Series Editor.
otoh, it is a group of people who presumably have more knowledge of
what the RSE does than the population at large. But I agree that
this particular point needs more thought.

> The document further considers the RSE to have executive authority
> over matters relating to "internal" issues of the overall RFC
> editor function.  In "Subjects for which discussion does not need
> to occur", the document lists any issues where:
> "the matter is strictly one of internal management within the RFC
>       Editor,"
> Given that the RFC Editor function is split, an internal matter
> might, in fact, be management related to the work of the
> publication group; this is a seriously different model than where
> we started with this.

Yes. As indicated above, I believe that we (for some value of "we")
started with the wrong model, failed to recruit an RSE because
of that faulty model, and now need to fix that aspect of the model.
I think Glenn is describing something close to a model that will
actually work, and is consistent with the IAB's authority under
its charter.

Full disclosure for those who may not know this: I was IAB Chair
at the time that the current IAB charter was adopted as a BCP,
and I was the document editor.


More information about the rfc-interest mailing list