Standardisation Process Re: IRC Log, twitter, responsiveimages.org

Dear CG members,
  
On Friday, 12 October 2012 at 23:21, Shane Hudson wrote:

> Last I heard it seemed Hixie had ruled out the picture element and was staying with srcset?
> 
> Is this wrong or has that change of how html is being standardised changed the situation here?
Although Mat will be providing an update later today, I would like to take this opportunity to clarify some things about the W3C/WHATWG standardisation process and politics. For those that don't know me, I've worked on standards with the W3C for ~6 years, and previously worked for a browser vendor (Opera) for ~2.5 years. 

Lets get something clear as a group: Ian Hickson does not control what goes into HTML - what is in browsers, and what browser vendors are willing to implement is what ultimately goes into the HTML specification - Historically, I think Ian has made this pretty clear: HTML intends to reflect the state of the Web as an interoperable platform (warts and all). 

Thus, Ian's role is to transform features into prose so that an interoperable standard can emerge in a way that works with the existing platform - these features become part of the platform when they appear in browsers. Ian has as much right as anyone else to voice an opinion about a proposed technology - his opinions are highly informed and influential amongst browser vendors (so if he says a proposal has problems, it is important to work constructively to address those). 

But Ian's opinions are just opinions - and things that are in WHATWG's HTML don't often reflect what browser vendors actually want (see [1]!). Ultimately, it is having three or more browser vendors willing to implement a technology that will see it get adopted and become part of the Web Platform. So, it's not Ian who we need to convince (though it would help a lot!), but it's namely at least three of Mozilla, Microsoft, Opera, Apple, and/or Google that we actually need to collaborate with. 

As a community predominantly composed of web developers, we have a good understanding of what we want and need (now formally documented in our "Use Cases and Requirements for Standardizing Responsive Images" document). What technical form that document takes - be it <picture> element, img@srcset, or something else - does not actually matter (think "form follows function"). 

The only thing that really matters in the end is that we end up with a solution that addresses our requirements and that those requirements are clearly articulated to browser makers in a constructive and collaborative way. This is why it's so critical as a first step that we make sure we capture what we need fully and that we continue to push to make sure our requirements are heard AND met through whatever solution emerges. 

Right now, the path of least resistance is to buddy up with Apple through the img@srcset proposal. However, if we are dead-set that the picture element is a better solution then we need to convince some browser vendors to implement it (we don't need a full spec, we just need to describe more or less how it works and what it needs to do). Writing a specification without explicit backing from a browser vendor will ultimately be an exercise in frustration, so we need to avoid that. 

I'll leave it for Mat to articulate a bit more where we go from here:) 

Hope that helps! 
Marcos 

[1] http://lists.w3.org/Archives/Public/public-html/2012Oct/0055.html 
-- 
Marcos Caceres

Received on Saturday, 13 October 2012 11:29:38 UTC