Re: Moratorium on the spec-splitting discussion

On Jan 28, 2009, at 6:02 AM, Sam Ruby wrote:

>
> Maciej Stachowiak wrote:
>> On Jan 28, 2009, at 3:58 AM, Sam Ruby wrote:
>>> Maciej Stachowiak wrote:
>>>> I would be happy to see a structured discussion led by the chairs
>>>
>>> Let's test that assertion.
>>>
>>> The structured discussion that I would like to propose is that we  
>>> produce concrete specs, distribute those specs widely, identify  
>>> specific issues with those specs, and block progress of any and  
>>> all such drafts to the Proposed Recommendation stage until those  
>>> issues are disposed of one way or another.
>> I'd like to ask for a clarification.
>
> Thanks!
>
>> Is this meant to imply that we would not block progress to any  
>> stage earlier than Proposed Recommendation, or do we still take a  
>> group decision at those other transitions?
>
> No.  There are transitions that require consensus, and there are  
> transitions that do not.
>
> I believe that previous attempts to apply standards of consensus to  
> transitions which are clearly documented as not requiring such,  
> while well intentioned, were misguided.

I'm asking about "group decision", not necessarily consensus. Per the  
W3C Process, when consensus cannot be achieved, a group decision  
should nontheless be made by other means, such as voting.

My understanding is that the following transitions require group  
decisions:

- First Public Working Draft: "The Chair must record the group's  
decision to request advancement.  ... the requirements for publication  
of a Working Draft are limited to an agreement by a chartered Working  
Group to publish the technical report and satisfaction of the Team's  
Publication Rules ... Consensus is not a prerequisite for approval to  
publish; the Working Group may request publication of a Working Draft  
even if it is unstable and does not meet all Working Group  
requirements."

This says to me that FPWD requires a group decision to start a spec  
down the standards process, though not necessarily consensus on the  
technical content of the spec.

- Last Call: "A Working Group's Last Call announcement is a signal  
that: the Working Group believes that it has satisfied its relevant  
technical requirements (e.g., of the charter or requirements document)  
in the Working Draft... Before announcing a Last Call, the Working  
Group must do all of the following: 1. Record the group's decision to  
request advancement."

Seems pretty clear - LC requires a group decision as well as agreement  
that technical requirements have been satisfied.

- CR/PR/REC: "n preparation for advancement to Candidate  
Recommendation or subsequent maturity levels up to and including  
publication as a Recommendation, the Working Group must: Record the  
group's decision to request advancement.... Formally address all  
issues raised about the document since the previous step... Report any  
Formal Objections."


So my understanding is that issues and objections should be raised  
well before the PR transition, that every transition requires some  
form of group decision, and that even the FPWD decision amounts to a  
group decision to put something on the standards track (though it is a  
decision the group can reverse later, and of course it is possible to  
public a clearly non-REC-track WD). Does that match your understanding?

My apologies for quoting the rules, but since you couched your  
proposal in W3C Process terms, I thought that might be the best way to  
get clarity.

So, given these assumptions, what remains to decide is what is a  
reasonable decision process for such transition decisions, and how to  
ensure that decisions are appropriately informed. I think our existing  
process of web survey following sufficient discussion of relevant  
issues is an acceptable way to do things. I would agree that we  
shouldn't let a very few objectors throw sand in the gears - at some  
point, their issues or objections should be recorded, and we move on.  
But neither should we let a very few proponents get through without  
engaging opposing arguments, and without seriously checking the actual  
levels of support, opposition and apathy.

>
>
>> In other words, are you proposing that advancement to FPWD, LC and  
>> CR should be essentially automatic?
>
> No.
>
>> Or was your "to" meant to be a "towards"?
>
> I meant "to" to mean that specific transition.  I also meant "to" in  
> a non-exclusive manner.

It's not clear to me why you highlighted that particular transition as  
the key one. In my opinion, that point would be way too late to raise  
objections to placing the spec on the standards track at all. In fact,  
if someone held such an objection until that point, despite long being  
aware of the issue, I would consider it not timely raised, and would  
suspect their behavior was in bad faith.

So, I would prefer the Chairs not highlight PR transition as the key  
point for Working Group members to raise objections, or the  
appropriate time to decide how the spec should be factored.

Regards,
Maciej

Received on Wednesday, 28 January 2009 15:08:25 UTC