Re: [hybi] Issue tracker

[Ian Hickson <[e-mail]>]

On Thu, 13 May 2010, Salvatore Loreto wrote:
> On 5/13/10 4:36 AM, Ian Hickson 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)
> >      
> > That sounds like a huge amount of overhead. Here are the issues that I'm
> > currently aware of for Web Sockets:
> > 
> >     http://www.whatwg.org/issues/#websocket
> > 
> > Do I need to send an e-mail for each one of these to raise it as an 
> > issue?
>
> If you really think that an issue is important and it is worth to 
> discuss it within the community, then you should send by mail to the 
> mailing list and ask to add it to the Issue tracker

Could you give me some guidelines for how I should decide whether an issue 
should result in my following the process above or whether it should 
result in me just fixing the spec? The process you outline above doesn't 
seem to leave any room for me not raising issues. Should I raise all 
issues? If not, how should I determine what issues to raise?


> Discussing and building consensus is always a time consuming task, but 
> it helps a lot in writing a good and robust protocol leveraging on the 
> help and the experience of a lot of people

Consensus is not required to write a good and robust protocol leveraging 
on the help and the experience of a lot of people. Determining consensus 
also doesn't need to be time-consuming. However, while my preference is 
for a trivial (and fast!) proven process such as the WHATWG's, as used so 
far in the development of Web Sockets, I'm willing to use whatever process 
you come up with, so long as it is clearly defined.

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