On Wed, 19 May 2010, Salvatore Loreto wrote:
> On 5/19/10 12:05 AM, Ian Hickson wrote:
> > On Tue, 18 May 2010, Salvatore Loreto wrote:
> > > > > > On Wed, 12 May 2010, Salvatore Loreto wrote:
> > > > > > > let me clarify about how we intend to work
> > > > > > > 
> > > > > > > 1) Even if anyone is welcome to suggest issues, only the 
> > > > > > > chairs are allowed to create a new issues!
> > > > > > > 
> > > > > > > 2) We discuss the issues on the list for a while
> > > > > > > 
> > > > > > > 3) Issues related to WG documents will be initially assigned 
> > > > > > > to one of the authors ( mainly because the tool allows to 
> > > > > > > enter only one assignee ).
> > > > > > > 
> > > > > > > 4) the assignee (or the WG document authors) come up with a 
> > > > > > > form of words that they think matches the consensus and 
> > > > > > > attach a diff to the trac
> > > > > > > 
> > > > > > > 5) if there is a consensus on the text, the author will add 
> > > > > > > it to the WG document and the issues will be closed. if 
> > > > > > > there is no consensus on the text go back to 4)
> > >
> > > As editor of the document you do not need to necessarily check all 
> > > the details with the WG before submit a new draft and its up to you 
> > > judge what to do with each specific issue: if you are confident that 
> > > the way you are fixing the spec has the consensus (perhaps because 
> > > it has already been discussed) or it will receive the rough 
> > > consensus from the wg, then go ahead and fix it; if not it is better 
> > > follow the process listed above.
> 
> "/The general rule on disputed topics is that the Working Group has to 
> come to "rough consensus", meaning that a very large majority of those 
> who care must agree./"
>  (The Tao of IETF)
> 
> as chair, I consider "those who care must" all the people that are 
> involved in the HyBi mailing list discussion!

Do you mean all the subscribers to the mailing list, or all those posting? 
If the latter, over what time period?

How large a majority is "very large"?


> So the Web community as a whole is very welcome to join the discussion 
> in the HyBi mailing list, if they do not care to join the discussion 
> then is quite difficult to judge their opinion!

It's actually very easy to determine their opinion; we can just ask them, 
or follow their discussions elsewhere. For example, I regularly partake in 
discussions about HTML5 on Reddit, on IRC, and in other forums, even 
though many people there aren't in the HTML working group. Indeed, using 
Google alerts and Backtype alerts one is able to get a very good sense of 
where people outside the working group stand.

However, while I obviously will continue to look at those sources of 
input, I understand that we'd want to ignore this feedback when 
determining "consensus", if we're going to use a consensus-based approach.


> > Should consensus be considered more important than technical 
> > correctness and interoperability with implementations, or should 
> > making the spec technically sound and making the spec match reality be 
> > considered more important than consensus?
>
> it is quite unlikely that the consensus does not match a solution that 
> is technically correct and interoperable with implementations.

I wish I could share your optimism.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'