Re: PROTO Process

[Jari Arkko <jari.arkko at piuha.net>]

I wanted to provide some additional background on why we are asking 
about the PROTO process.

The IESG met this week and talked about a number of things related to 
our process. One of the things that we would like to improve upon is the 
progress of drafts through the IESG process. There are several problems 
in this, including:

- the fact that 82% of drafts have to be modified during AD review - 
IESG review (and the rest probably have RFC Editor notes)
- the process takes a relatively long time (avg 9 weeks for drafts that 
stay unchanged and 31 weeks for drafts that change)
- the timing is rather unpredictable, there is great variation among drafts
- the visibility to what is happening and who has the token could be 
improved
- there is at least anecdotal evidence that for many drafts, there is a 
long wait for new version or a response to review comments

Various statistics on these are available at 
http://www.arkko.com/tools/admeasurements/stat/base.html (but do not 
stare at the numbers too directly). The situation seems to have improved 
with regards to various metrics in recent times, but the IESG discussed 
a number of actions to make additional improvements. These include:

- Reminding ADs and shepherds to keep their working groups better 
informed about the ongoing reviews, raised issues, and suggested changes 
to their drafts.

- Starting a discussion in the wgchairs list about the roles of 
shepherds and whether there is something to improve.

- Create a "stall" state in the tracker. This would be used for 
documents where the wait for a response to an IESG or other review has 
taken a long a time. The fact that the draft has entered this state 
should also be communicated to the WG.

- Reminding everyone that if the holdup on a particular issue is due to 
an unresponsive AD, you can contact Russ to send a reminder.

Then I wanted to say a few things about the shepherd process. Please 
remember this works well in many cases, and we get great help from the 
shepherds in resolving issues. Or the authors and ADs converge quickly 
and things just move forward. However, there are cases where I would 
expect more action from the shepherds. For instance, Pasi often reports 
in his AD notes 
(http://www.ietf.org/mail-archive/web/saag/current/msg02643.html) about 
cases where he has raised a Discuss but has not heard from the authors. 
The shepherds and WG chairs should be active in such cases, reminding 
authors to move forward, finding new editors if needed, etc.

I'd also observe that generally speaking the chairs and shepherds and 
more experienced in the IETF process and have worked with more 
specifications than individual authors. This experience is very valuable 
when talking about issues raised during the Last Call and IESG reviews. 
It takes judgment and knowledge of the WG's area to determine whether a 
particular issue is truly a problem. The ADs make these calls, but it 
often comes down to a discussion between an individual AD having 
specific expertise (such as transport or security) and the author; I 
would like to see more shepherds participating in the discussions and 
weighing in. Pushback on Discusses is appreciated!

About the writeups: the ADs do read these, and I at least find them 
useful. I also often read writeups from other AD's documents in the IESG 
review stage, if something about the status or history of the draft is 
unclear. Of course, the writeups are truly useful only if the shepherd 
really did do a final review of the document, checked the nits, and 
provided more useful background information than simply copying the 
abstract from the document :-) But in most cases that I'm aware of they 
do contain useful information.

Jari