Re: Call for Consensus: Please review the draft charter

On Friday, December 2, 2011 at 11:47 PM, Carr, Wayne wrote:

> Nice going!
>  
> Just a couple of comments. The last one is the only serious one.
>  
> "The CG treats specifications as living documents: specifications are not versioned and they are not dated"
> I'd delete that line. Maybe it doesn't need to be (though when working with a WG I think it's going to be necessary to know exactly what was the input for a particular stage). You could just remove the prohibition and see what works.

Agreed. Versioning occurs based on check in version on Mercurial. I will find a way of to plug the two together.   

Do you think that would satisfy the need to have date/version?  

I just don't want dates to indicate or imply milestones (there is better things to do that, such as the "Status of this Document").  

> "A W3C Working Group takes ownership of it to progress it along the Recommendation track."  
> But, you never give ownership do you? You give them a version. They make change requests, you give them another version. They decide to publish some version as a Candidate Rec. You don't actually give them "ownership" - you give them copies of versions. The actual editing of the doc for technical content would be in the CG, not the WG.

I was thinking they were going to fork it. But yes, the above would be good.   
> "Every conformance requirement has a tests"
> Typo - "test"
>  
> "A specification is “done” when"
> I'm not sure about that list - I'd drop it. That's when a version of a spec is gone to REC. But if there are no "versions" it isn't meaningful so say that means it is done. It would be done when no one wants to keep working on it. It seems a spec is done when the CG no longer wants to work on it because they think it is done or decides to drop it.

Agreed. Dropped.   

> "If a decision is necessary for timely progress, but consensus is not achieved after careful consideration of the range of views presented, the Editor of a specification can make a decision on a matter. In extreme cases, the Chair will put a question"
> Instead of "In extreme cases" it would be better to have some process to object to an Editor's decision and call for a vote. One possibility is someone objects, the chair then makes a decision. Someone formally objects to that and it goes to a vote. And it should say that should be avoided whenever possible. But there needs to be a way to not have ultimate decisions be one or two people.

Does this work:  

"If a decision is necessary for timely progress, but consensus is not achieved after careful consideration of the range of views presented, the Editor of a specification can make a decision on a matter. If further objections are raised on the matter by members or the public, the Chair will weight the evidence/arguments and propose a path forward. If one or more members still object, then a vote will be called within the group (allowing for remote asynchronous participation -- using, for example, email and/or web-based survey). Based on the outcome of the vote, the matter should then be considered resolved unless and until new information becomes available. Voting on objections is a matter of last resort and is expected to be used only under exceptional circumstances."

Received on Monday, 12 December 2011 18:11:29 UTC