Re: draft-housley-two-maturity-levels

[John Leslie <[e-mail]>]

Eric Burger <[e-mail]> wrote:
> 
> The known problem is it takes well over four years to get anything
> published.

   We've actually been making progress on that, at least for the part
where the IESG is involved.

   Russ's draft _would_ reduce one issue: where a reference should be
Normative but hasn't been called out, the LastCall must be repeated.
This is an irritant, and never seems to actually change anything.

   I frankly cannot see how Russ's draft would speed the process any
other way. And I worry that changing to two levels would make folks
even more hesitant to approve Proposed Standard status. (This is more
a worry than an actual issue with what Russ's draft _says_...)

> I am experiencing in one never-ending saga the logical conclusion
> of the logic: since Proposed Standard is the new Draft Standard,
> and since the IESG has to make sure any proposal is beyond bullet-
> proof, the industry has long since implemented draft-mumble-21,
> which has not changed for over a year, and few in industry cares
> if the document publishes as an RFC, because from their point of
> view, if something has been working in the field for three years,
> has 18 independent implementations, and has not yet crashed the
> Internet, it is probably ready, whether the IETF formally says so
> or not. That is the fast track to making the IETF irrelevant.

   This is a function of WGCs worrying too much before asking the
IESG to act on Working-Group submissions. I can't see Russ's proposal
reducing that problem. :^(

   BTW, the Individual Submission route bypasses that problem, and
is being used fairly liberally. Individual Submissions _never_ take
four years (unless the submitter lost interest along the way).

> The very real danger here is while that attitude may be OK for a
> small media application, that attitude could be a disaster in,
> for example, the routing area. Something really has to be done.

   Indeed, we _have_ seen a disaster in the routing area... :^(

> Now, I do agree with Scott here there is absolutely no incentive
> for anyone to bring a protocol to Draft/Internet Standard level.

   Scott exaggerates a bit, but the incentive _is_ usually lacking.
It's the lack of incentive, combined with a belief that so-and-so
"should" shepherd the advancement, that has led to IESG desires
to remove known problems from PS drafts.

   (Thus, IMHO, the fix for this problem is either better incentives
or assigning responsibility for advancement to someone who _has_
an incentive to complete it.)

> What I *am* hoping is that with two, clearly defined maturity
> levels, we can go back to publishing Proposed Standards in about
> a year, and have Internet Standard mean something.

   "Hope springs eternal..."

   Alas, I can imagine no rational basis for that hope. :^(

> Otherwise, we will be perpetually running the Internet on Internet
> Drafts, which is something I do not think anyone really can say
> is a good thing.

   We _cannot_ stop folks from "running" the Internet on I-Ds. And,
to be honest, the Internet is based on RFCs that were _easier_ to
publish than I-Ds are today.

   Although Russ's draft _would_ fix the downref irritant, I can't
see it doing anything to reduce the PS logjams -- indeed, human
nature ensures we'll invent _more_ logjams once PS is "only one step"
from final Standard.

   Lots of us would like to reduce the PS logjams, but we need to
look outside the part of 2026 that deals with number of levels.

   We might, for example, specifically _allow_ Individual Submissions
to bypass a congested WG if no specific objections could be listed.

   We might reduce the level of "considerations" (such as Security
Considerations) expected for Proposed Standard (though, IMHO, that's
an increasingly dangerous place to go), or assign responsibility
for adding such Considerations to something like a Security Area
Directorate.

   We might forbid WGCs from putting things off until the next IETF
week...

   We might formalize issue-tracking for WGs considering something
for PS...

   We might formalize the Document Editor role to prevent "Document
Editors" from pushing their own viewpoint...

   We might formalize a WG Secretary role to clarify specific issues
raised during in-person and virtual meetings...

   We might _require_ virtual meetings whenever a milestone is in
danger of slipping...

   And, frankly, I could go on, and many others could go on listing
things which _would_ reduce some of the logjams that have led us
to such long leadtimes to reach PS status.

   We may be nearing the point, universally found in legislatures,
where we'll "vote for anything so we can adjourn". If so, we could
do worse than Russ's proposal.

   But "vote for anything" _very_ often leads to bad law. I fully
agree with Scott Bradner that changing the number of levels isn't
something we should do lightly.

--
John Leslie <[e-mail]>