ietf
[Top] [All Lists]

Re: Last Call: <draft-resnick-on-consensus-05.txt> (On Consensus and Humming in the IETF) to Informational RFC

2013-10-27 11:01:55
On 10/9/13 5:20 AM, Eliot Lear wrote:
1.  No matter how you slice the definition of rough consensus, if the
chair does not act in a fair and balanced way, the outcome will be
incorrect.  This is what the appeals process is for, and I would suggest
mentioning it, perhaps in some detail.

I do allude to this at the end of section 3, and I've highlighted it a bit better with one of the edits Dave suggested. But every time I tried to elaborate the "detail", I found myself talking more about the theory of appeals rather than the theory of consensus. I'm open to some suggested text if you have some.

2.  The case of Section 7 is, as you say, a mind bender.  I would
suggest adding another use case: what if those 100 people write their
own draft.  Can they use the exact same process to get the draft adopted
and approved, so long as they answer the technical issues that arise?
In other words, if there are multiple valid alternatives, and one suits
one vendor group and another suits another, can there be just one
standard?  At the neck of the hourglass, perhaps so?  What happens in
this case, from your point of view?  What makes group (a) more special
than group (b)?

Though I think your example adds delightfully to the mind bender, I'm inclined to avoid it. At one level, 100 proposed drafts is no different than 100 proposed objections, though I would contend that if the drafts are actually well written, they would have to have good reasoning as to why the WG's original choice was not the best one. One way to sort your proposed mess is for the chair to go back to the charter of the WG and see which proposal actually meets the charter requirement. Or the chair can make the question about who the assigned document editor is. But really what your proposing is just a step worse into the "pathological WG" problem. And really we could do this ad infinitum. (I'm reminded of the old George Carlin routine about his days in Catholic school and "Ask the priest" day: "So, Father, say you haven't received your Easter communion. And it's Pentecost Sunday, the last day. And you're on a ship at sea. And the chaplain goes into a coma. But you wanted to receive. And then it's Monday, too late. And then you cross the international dateline...".) The fact is that though rough consensus is a resilient process, it's not bulletproof to every attack. This document can only go so far into how it deals with problems. I've added a good bit to that section to address Ted's comments. Hopefully you find that satisfying.

pr

--
Pete Resnick<http://www.qualcomm.com/~presnick/>
Qualcomm Technologies, Inc. - +1 (858)651-4478