[rfc-i] new draft summarizing updated Transitional RFC Editor recommendations now available
ajs at shinkuro.com
Fri Nov 26 14:19:57 PST 2010
On Tue, Nov 23, 2010 at 12:47:49AM -0800, Glenn Kowack wrote:
> A new draft summarizing updated Transitional RFC Recommendations,
> "draft-kowack-rfc-editor-model-v2-overview-00" is now available.
I have read the document. I have some comments.
I'm going to treat it as a hastily-made draft and therefore not bother
with particular comments about structure, grammar, and usage, but I
sure hope a fairly serious pass will be made over the draft to address
the quality of the English before the final version. Normally, I
wouldn't even comment at this stage; but this _is_ a document about
the editor, and having to re-read sentences multiple times to make
sense of them is pretty jarring.
The reporting structure is not clear to me. The diagram in section
2.1 says that the RSE reports to the IAB. Section 4.1 says the RSE
reports to the REOC but is appointed by the IAB. I suppose the latter
somewhat unorthodox arrangement could work in practice, given that the
members of the REOC apparently to serve at the pleasure of the IAB,
except that neither of these bodies is in a position to write
contracts. So the IAD (right? anyway, whoever is in a legal position
to write those contracts) has to appear in here somewhere, and doesn't
as far as I'm able to make out. This needs serious work. I guess I
think what is proposed is that the RSE works under the general
direction of the IAB, with day to day supervision provided by the
REOC. The IAD undertakes contracts with the RSE at the direction of
the IAOC and according to the express wishes of the IAB. Is that
Section 4.2 includes some responsibilities I think should not be
included. In particular, I don't get these two:
o representation of the Series to the community, and
o representation of the Series to the rest of the world.
I think they should be removed.
On bullet 1, I think the only possible reason the RFC series is
important is that people think it is useful. Those people who think
it is useful are the only people I think could possibly qualify as
"the community" in the above. So the series is tautologically
represented to them. The specific details of this responsibility are
outlined in 4.2.4. It seems to me that bullets one and three in that
list are just a basic part of doing the document quality work outlined
in 4.2.3; that the "as required" in bullet two amounts to saying
either that this should all be done in support of the first three
tasks, or whenever the supervisor (IAB? REOC?) says so; that bullet
four is a simple description that the RSE has to tell his/her boss
what s/he is doing; and that the last bullet is empty of any meaning
On bullet 2, the claim is either preposterous or false. "The rest of
the world" broadly construed is not a class of people needing the RFC
series represented to them. My mother does not care even a little bit
about the RFC series. "The rest of the world" narrowly construed --
i.e. people participating in areas touched on by the various input
streams but not part of that input stream community -- should not have
the RSE involved in their issues, either. That's a responsibility of
the relevant liasons to those other communities, I think. The
expansion of this responsibility in section 4.2.5 is, to me, a
demonstration of what I'm suggesting, because it seems unlikely
that the press is ever going to care about the series _itself_, except
on occasions so rare that it would be better handled as a one-off
event and not a part of any job description. More likely, the press
will be interested in particular controversies, and that sort of thing
should never be handled by the RSE. I have no idea what the second
bullet means, so I think it should be removed.
I think the rest of the responsibilities are ok. In particular, I
think I have no substantial objections to the items outlined in
sections 4.2.1, 4.2.2, or 4.2.3 (although as a nit I'd like to see the
parentheses removed from the second bullet in 4.2.3).
I have no objection to any of the items in 4.3.1.
Section 4.3.2 appears to have a number of special projects bound up in
it. These are one-time events that, I suspect, don't belong in the
job description of the RSE unless we plan to go through this root
canal again in the near future. 18.104.22.168 contains one such project.
22.214.171.124 doesn't actually define Production Quality at all, but talks
about a possible service that might be investigated. These sections
should be cut.
Section 4.4 seems reasonable to me.
I sent comments about the multiple committees under separate cover.
Specifically on the REOC, I agree with Ted Hardie that the strictures
on conflict of interest are too high on the REOC given that similar
conflicts have been explicitly permitted for the RSE itself. Surely a
simple recusal model for the dispute resolution would be enough to
solve any worries about conflicts?
In section 6, we have this:
If one party still disagrees after the reconsideration, that party
should ask the Series Editor to undertake a formal review. The RSE
must inform and engage the REOC in their oversight capacity, and may
call for a review committee including members of the REOC. The RSE
and REOC should seek to reach rough consensus on the resolution of
It strikes me that this is underspecified. The RSE may call for a
review committee, but it doesn't say who appoints the committee, how
they're selected, or what the membership must be. Is that ok? I
think probably not. This is exactly the sort of case in which it's
useful to have sound rules. While most of the time I am very much in
favour of loose rules with general principles, we need to have fairly
automatic and completely specified rules for who gets picked to do
conflict resolution. Otherwise, the process of picking the dispute
resolution body is just another way to fight the same dispute.
Perhaps the next paragraph is supposed to constrain things, but if it
is I was unable to understand how.
I don't have anything that I would like added to the document.
With the changes I outline above, I think the position ends up
slightly on the "broad" side of the two constructions for the position
I outlined some time ago, but I'd entertain arguments to the contrary.
I have no idea why to select between those two constructions, because
there are no reasons in the document for either view. Therefore, I've
just assumed the broad construction is correct for the purposes of
these comments, and suggested (among other things) where I think the
document has erred in the direction of "too broad".
ajs at shinkuro.com
More information about the rfc-interest