RE: [mpowr] Re: [Solutions] Further work on WG (chair) roles - MPOWRWG proposal

[Pete Resnick <[e-mail]>]

On 12/12/03 at 5:48 PM -0500, [e-mail] wrote:

>>A chair does not need the authority to say no in order to get 
>>technical reviews assigned, do a technical writeup, or track 
>>outstanding issues. They already have full authority to accomplish 
>>those tasks.
>
>Some of this may be a terminology issue...
>
>Do you think that a WG chair should be allowed to hold a document 
>(i.e. not submit it to the IESG) while the chair seeks sufficient 
>cross-area review?  Or do you think that a chair is obligated to 
>send a document to the IESG as soon as the WG has consensus to  do 
>so, even if, it is the opinion of the chair that the document 
>requires more cross-area review before it is ready to advance?

I don't think the answers to those questions are mutually exclusive, 
and I think they might both be "yes", but it may take a bit to 
explain. Someone earlier in the week thought I might be able to 
address a comment by Dave Crocker:

>Specifically, there is no consensus that WG chairs should _ever_ be 
>authorized to overrule WG consensus.

by replaying something I talked about at the Vienna chair training 
session about consensus. I think it may also be helpful to answer 
Margaret's query:

Rough consensus is about a meeting of minds in a WG. If a WG is 
working the way it should, with everyone talking to each other "in 
good faith", it's pretty easy to figure out when the group has come 
to rough consensus. However, it is perfectly plausible that a WG 
chair could try to judge consensus in a WG (one of their 
responsibilities in 2418, sec. 3.3) and finds no consensus even if 
almost everyone is saying the same thing. (I used an example of a 
real WG in Vienna; this example is completely hypothetical.)

Suppose 99% of the members of a WG list pipe up on the list and say, 
"In our document on the Foo protocol, we think the 23rd byte of the 
packet should always have the value Bar." One member of the list, 
Pete, pipes up and says, "Wait a minute, if you put Bar in the 23rd 
byte of the packet, every IP router on the Internet will crash in 
this very obvious way", laying out a very nice technical argument 
that seems exactly right, and 5 people give support to the argument. 
The rest of the 200 members of the WG continue to say, "No, we all 
agree that Bar should go in the 23rd byte." When asked about Pete's 
comment, every one of them gives answers of the form, "Well, Pete's 
an idiot, so that can't be right" or "Who cares? I'm just deploying 
this on a closed network" or "I don't like Pete and his supporters, 
so I think we should leave it as-is for spite".

If I'm the chair of this WG, I might very well decide that there is 
*no* consensus to but Bar in the 23rd byte because people in the WG 
are not addressing each other's arguments. In fact, I might say, 
"Since there has been a strong argument by Pete that putting Bar in 
the 23rd byte is bad, some well-reasoned support for that position, 
and no one else in the WG has seen fit to put in a real argument 
against, there *is* rough consensus to *not* have Bar in the 23rd 
byte", instruct the document editor to make the change, and not hand 
the Foo document to the IESG until said change was made. Such a WG 
could appeal my decision. They should lose such an appeal.

On the other hand, if those other 200 people gave reasoned technical 
arguments for why it is perfectly safe to put Bar in the 23rd byte, 
even if they can't convince those other 5, and even if I personally 
disagree with their reasoning, I shouldn't hold that document. If I 
did and they appealed, they should win the appeal (and likely I 
should be fired as WG chair).

It is the chair's job to judge consensus and communicate it. However, 
sometimes that doesn't amount to blindly counting votes or listening 
to hums.

So in answer to your original questions: Yes, a chair can hold a 
document while seeking cross-area review. If the 99% of the WG 
shouts, "No, send it along, we don't care if it's had review", that's 
*not* consensus in the WG that the document has had enough review. 
However, if the WG *has* come to consensus that there's been plenty 
of cross-area review already during WG discussions, yes, I think the 
chair obligated to send it along, even if the chair thinks it needs 
more.

However, both scenarios paint a rather gloomy picture: In the first, 
the WG has lost its mind; they're not coming to consensus about a 
document, they're just making unfounded claims. In the latter, the 
chair has failed pretty significantly. If WG discussion has gone 
along this far and the chair is the only one left at the end of the 
discussion who thinks more review is required (meaning also that they 
didn't get that review done before the rest of the WG started kicking 
and screaming that the document was ready to go), the chair blew it 
months ago. The chair should probably get the boot. What we've got is 
a complete process breakdown at that point, and I'm willing to say 
that it should be the IESG that sorts it out at that point. 
Introducing some sort of new "authority" the chair has to overrule 
consensus isn't addressing the problem in the least.

Chairs have plenty of authority to judge consensus now, and in the 
normal course of events, that will allow them to hold a document when 
there hasn't been sufficient cross-area review. However, it won't 
allow them to usurp the consensus of the WG.

Did that answer the question?
-- 
Pete Resnick <http://www.qualcomm.com/~presnick/>
QUALCOMM Incorporated - Direct phone: [phone], Fax: [phone]

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