[whatwg] "due consideration"

[Maciej Stachowiak]

On Jul 24, 2009, at 5:03 AM, Eduard Pascual wrote:

> On Fri, Jul 24, 2009 at 10:04 AM, Maciej Stachowiak<mjs at apple.com>  
> wrote:
>> Ian gives more careful consideration and more thorough responses to  
>> comments
>> than any other specification editor I have seen in action. I've  
>> commented on
>> many W3C standards and many times I've seen comments raising serious
>> technical issues dismissed without explanation, or just ignored. I  
>> have
>> never seen that with HTML5.
>
> Is that really enough?

I think any process has room for improvement. My main point was a  
comparative one: I've gotten more concrete sense of input being  
considered in the context of HTML5 than almost any other standards  
process in which I have been involved.

> <The point>
> I do not doubt of Ian's good faith, nor of his huge effort in making
> HTML5 the best possible thing it might be. However, I doubt of the
> sanity of having an individual to have the final say about any topic,
> even above expert groups that have been researched and discussed the
> topic for years.

[...]

> Honestly, I can't say for sure which method would be best for HTML;
> but I'm still convinced that having a single gatekeeper with absolute
> power over the next web standard is, at least, insane.
> </The point>

My personal view, as I've said, is that I think the HTML WG has the  
authority to override the Editor by official Working Group Decision,  
whether on a single technical issue or by adopting a separate draft  
wholesale. Of course, Ian may choose to resign as Editor if this  
happens. Personally, I think we are best off if we almost never need  
to bring an issue to a vote. And I would hope Ian would do the right  
things to forestall such a constitutional crisis.

Indeed, in the past, I have seen many formerly contentious issues  
resolved in a way that is satisfactory to everyone. For example,  
degree of requirement for alt="", conformingness of headers="", an  
author-only version of the spec, presence of the SQL database API in  
the spec, and the old version of the WebSocket API, are issues that  
used to be debated constantly, but where I think the current state of  
the spec largely satisfies everyone. Besides discussion, there are  
other things that can change the spec. If there is a feature that  
implementors uniformly fail to implement, or that authors widely  
refuse to use, then I am confident it will fail. Indeed, before our  
final Last Call (perhaps not the first) we will need to prune all  
features that have failed to get traction.

So I guess my position is, a trusted person or small group making  
initial decisions, followed by discussion, followed by the possibility  
of group override in the most extreme cases, is in principle a  
reasonable way to work. And I think we are pretty close to actually  
following that process, even though many would describe it as more  
dictatorial than that.

I snipped discussion of the specifics of RDF and Microdata because  
this is not an area where I have either strong opinions or relevant  
expertise.

Regards,
Maciej