Re: Another IESG Charter revision

[Thomas Narten <narten@us.ibm.com>]

> > so this has been the almost universially used path for as long
> > as I was on the IESG - I expect that there were a few docs
> > (though only one comes to mind - the power over mpls one)
> > that took the individual-ad-iesg-path but not many

I agree that today, most docs go through the RFC editor. But IMO, this
has been a gradual shift over time. 5-6 years ago, a lot more
documents came in via individual ADs. Maybe other folks' perceptions
are different.

> > also Klensin specifically complained that the individual-ad-iesg
> > path was being improperly used to give priority to some individual
> > IDs (he did this at a plenary session in the last year)

I object to the blanket assumption that going directly through an AD
gives documents priority. I suppose in some cases it does in
practice. But I've always just felt it was more efficient to take on a
document directly if it was something that I would end up being a
stuckee on anyway. And when I take on such documents, they just go in
the queue along with other documents.

If I give a document priority, its because I think it's important, or
it's a no-brainer to get through the IESG (little work expected,
document is simple, etc.), that sort of thing. There are lots of
factors.

But yes, this can create perception problems.  

> > his complaint was that he had heard a number of people say that
> > specific ADs said that they would push thorugh individual IDs
> > quickly if the ID was sent to them (and also the ADs were saying
> > that this was better than going the RFC Ed route because that was too
> > slow)

I also object to the assertion that RFC editor queue documents are
given lower priority and that we somehow encourage stuff to go there
when we want to somehow delay it. This is one of the things Klensin's
has asserted in the past. I.e., He has asserted at times that the IESG
has consciously put in place some sort of policy that rfc editor
documents get low service (including the IESG processing of those
documents).

> > fwiw - I think that there should be a public discussion if
> > the IESG wants to have a way to give some individual non-wg IDs
> > priority over others  (which is what I think is being asked here)

I suppose it's no a big deal to ask any individual who wants to
publish a document to send it to the rfc editor in order to make it
look like there is a uniform policy. But this also ignores things like
having ADs look at documents first to see whether they should be
standards track or not. I.e., don't we already have a "problem" if
someone has to find an AD to take on a document for standards track?

> > my main message is a request to define clearly what IDs can get this
> > treatment in a way that folks can understand and so that people
> > can not complain of unfair treatment

The problem is I don't see a hard and fast rule here. What is the
criteria for deciding whether a document fits exactly in my area or
that I have relevant experience on the topic to be a competent
reviewer/shepherder? That is my basic criteria.

> If we want to claim to be bound by 2026, we shouldn't break a "should" from 
> that document without any particular reason to.

AFAIK, we don't do this all that often, so I'm not sure how existing
practice conflicts with the should.

> I suggested "AD thinks it's important to the IETF" as a criterion, which 
> should give us the latitude for doing this when important - and if it's not 
> important, why bother?

I don't like this wording because I take stuff on not because I think
its important (how many of our documents really are?) but because as
an AD with expertise in some areas, it's my responsibility to look at
documents that fall in my area.

> Remember - these procedures are for Informational/Experimental individual 
> submissions ONLY.

Yep.

> Note to Thomas: If I have the SLP "set" right, at least one is currently 
> headed for Proposed (slp-remote-da-discovery), and another in the queue for 
> Experimental was originally proposed for Proposed (slp-da-interaction).
> slp-customization seems to be the only one that has in fact followed this 
> path. (Standard disclaimer about finding all the right docs apply....)

Yeah. It's more complicated. Some were requested to go experimental,
I think I argued that some should be PS and there was discussion about
what category they should be in and so forth. I'd hate to formalize
things to where we have the discussion, and then we formally have
submit things to the rfc editor. Just seems like a rule for the sake
of a rule.

But the point is, all 4 documents were sent to me, I put them in
tracker and then slowly started processing them, along with other
documents on my plate. I don't see how forcing them through the  rfc
editor submission would have been a better thing to do.

Thomas