Re: [mpowr] Why MPOWR?

[John C Klensin <[e-mail]>]

I want to seize on, and expand a bit on, one aspect of Bernard's 
note (I generally agree with all of it)...

--On Wednesday, 04 February, 2004 08:54 -0800 Bernard Aboba 
<[e-mail]> wrote:

>...
> I also believe we could cut significant time off the beginning
> of the process by streamlining the BOF process and moving more
> quickly to form "study groups".  There are quite a few cases
> where it has taken more than a year to form a working group,
> and even at least one case where formation took several years.
> Streamlining the BOF process seems quite likely to shave
> several months without a lot of effort.
>...

This, unfortunately, is another example of the process-rot 
problem that I've been trying to point out.  A decade ago, a BOF 
(much less a couple of BOFs) were not requirements for creating 
a WG.   BOFs were used if there was a need to get more 
information about interest and focus; if that need did not 
appear to exist, we went directly into a charter process.   If a 
BOF is unlikely to produce new information or insights, moving 
directly to chartering at least avoids having the nit-picking 
and grand philosophy arguments twice (or more) --once as to 
whether the BOF(s) should be permitted and with what agenda(s) 
and again over the details of the charter.  More about the grand 
philosophy arguments below.

In addition, there was a somewhat more relaxed attitude toward 
chartering WGs.  Several ADs took the position that, if a group 
seemed coherent but was unfocused, it was rational to charter 
them as a WG and watch them closely for a while to determine 
whether they were going to produce something useful or go off 
into the weeds.  If the latter, they got pruned, if the former, 
they flowered.  The assumption was that the best way to evaluate 
whether a WG was going to do something useful and competent was 
to let it try.  This direct approach was believed to be much 
more efficient, and less likely to waste a lot of everyone's 
time, than interminable discussions about charters, wondering 
whether WG proponents were trying to tune charters to what 
people wanted to hear even while planning something else, 
speculation about results, and so on.

Growth, and the "just too many WGs" problem would make that 
model more difficult if applied today.  More important, the IESG 
became shy of using it, probably for several reasons.  Certainly 
one of them was hearing from too many WGs "we struggled to get a 
charter, you chartered us, we did all this work, so we are 
_entitled_ to complete our work and have our output 
standardized".  Of course, the more time and effort goes into 
the chartering process, the stronger that argument gets.   I 
also assume that most ADs would prefer to spend time managing/ 
coordinating WGs that are trying to do technical work --and 
shutting down those that are hopelessly wedged-- than to spend 
the same time in charter debates, especially philosophical ones. 
I might even suggest that people who don't have that preference 
are unsuitable to serve as ADs.

It is important to note that the key procedures have not changed 
at all in the last decade.  The evolution to "you need a BOF or 
two" and "charters are subject to long community debate" both 
occurred without formal process changes.  If we don't like them 
--and they certainly add a lot to the time between "proposal to 
do work" and "finished product"-- they can presumably be undone 
by the same quiet and efficient mechanisms.

Now these changes occurred mostly because of IESG perceptions (I 
think accurate ones) that the community wanted more input into 
early-stage WG establishment efforts.   Well, that greater 
opportunity for discussion and greater openness now has a track 
record.  The track record is dismal, at least with regard to a 
"how long does it take to get things done" criterion.  We start 
a discussion.  It generates a few comments and, especially in 
"hot button" areas, a lot of noise (see below).  So a BOF is 
held, either because everyone believes by now that they are a 
requirement or to see if a face to face meeting resolves some of 
the controversy.   It doesn't resolve the controversy, partially 
because it is easier for someone to block progress by seizing a 
microphone and grandstanding than it is to do so on a mailing 
list -- if only because most of us have adopted "delete before 
reading" approaches to comments beyond a certain number by some 
people on some threads (that has its own bad effects, but they 
are another matter).  The mailing list is then filled, for the 
next few months, by the same noise, the same passionate 
arguments repeated many times, and by debates about what the 
first BOF actually concluded.  So a second BOF is held in the 
vain hope of sorting that out, and the process repeats.   If the 
effort then starts a charter process, the same arguments are 
replayed again, typically with little evidence that the BOFs 
were held or accomplished anything.  But six months to a year go 
down the rathole -- much time spent, little real work done.

If we want this to stop, we need to make it _very_ clear to the 
IESG, clear enough to overwhelm the noise, that we are tired of 
it.  No more BOFs, and especially no second BOFs, unless it is 
clear that useful information is likely to come out of them. 
An accelerated chartering process with clear community support 
for shutting down WGs that looked marginal at charter time, were 
given a chance anyway, but aren't producing (there may be 
elements of the O'Dell-Klensin and/or Huston-Rose proposals from 
early in this reform process that might be useful here).

And we need to be able to try that way of doing things (again?) 
without getting bogged down for months in BOFs, chartering 
discussions, and philosophical arguments about how to plan the 
process to do so.  We didn't need any of that process to get 
into this mess, and we shouldn't have to exercise the symptoms 
of the mess to get us out of it.   (A couple of us are finishing 
up an I-D that suggests a radical way out of this impasse.  It 
will be queued before the Monday deadline.   But it doesn't 
suggest anything that couldn't be done tomorrow if the IESG sees 
a requirement and mandate for it, sees community rough consensus 
about that mandate, and is actually inclined to make changes 
rather than fostering debate about them.)

Finally, that comment about philosophical debates.  There are a 
number of hard philosophical and/or architectural questions 
facing the IETF and the Internet.  They include a number of 
issues surrounding the NAT and middlebox questions; how much we 
are willing to distort the operations of well-established 
protocols in order to provide small, possibly-useful, patches 
for high-impact problems that arguably really lie elsewhere 
(spam fighting and the impact of email-spread and 
http/html-embedded malware being good examples here); questions 
about configuration alternatives; variations on the "open 
source" and/or "free software" debates; and so on.  They are 
important issues.  We need, IMO, to face and debate them in 
clear and open ways and, ideally, reach some conclusions.  But 
passions about them run high on both (all?) sides.  If every 
attempt to charter a WG or review a proposal for standardization 
that might infringe on one of those "hot button" areas restarts 
the debate in the context of that charter or proposal, things 
will move slowly --perhaps to the point that we will never get 
anything done-- and the passion of those debates will drown out 
any real analysis and discussion of the specifics of the 
proposal.   Unfortunately, we have ample worked examples of that 
pattern, more than enough that we should be learning from them. 
The people who are doing this are presumably sufficiently 
self-aware to know who they are.  In the interest of progress in 
the IETF --of having an IETF that, a few years from now, will 
[still be?] perceived as able to do useful work -- they need to 
stop attacking individual WG proposals and work on a more 
general debate and discussion.  And, if they can't or won't 
--especially after the general decisions of what the IETF will 
and will not do in a particular area have been made-- then the 
community needs to understand that this behavior, no matter how 
well motivated, is as destructive to progress in the IETF as 
personal attacks, repeated off-topic postings, and so on, and 
deal with them accordingly.

     john

p.s. While triggered by a note on MPOWR, this note probably is 
applicable across much of the "solutions" spectrum.  If someone 
feels that it has enough value to call attention to it on some 
other list, feel free.


_______________________________________________
mpowr mailing list
[e-mail]
https://www1.ietf.org/mailman/listinfo/mpowr