Mozilla is happy to support Facebook in forming a Core Mobile Web Platform W3C Community Group in which to curate prioritized, tiered lists of emerging and de facto standards that browsers should support in order for the Web to compete with native application stacks on mobile devices.
The W3C Community Groups do not create normative specifications; their work is informative at most [UPDATED per Ian Jacob's comment]. However I believe they can add significant value, especially by helping developers make their priorities clear to the implementors who tend to control the normative specs (W3C Recommendations).
Standards-making like law-making is definitely sausage-making. How could it be otherwise, with intensely competitive companies trying to work together?
On top of this, consider how conflicted many standards bodies are by pay-to-play, however muted and tamed by “process”. Anyone can join with enough money, and inject a divergent agenda or random noise into the process.
One inevitable outcome of these conflicts is too many proposed and even finalized standards for all browsers possibly to implement correctly and completely. The nice thing about standards is….
Who is best situated to advise implementors (mainly browser vendors) on which standards to prototype and finalize first? In my view, developers. But of course you can’t ask developers questions to answer with one voice. Developer communities must acclaim their own leaders, who then speak to standards bodies.
Last year, Facebook joined the W3C. I thought at the time “there is a company with skin in the Web content game, not only for pages but especially for apps.” Facebook relies heavily on HTML5, CSS, and JS. Facebook has no browser in the market to pull focus or inject asymmetric browser/service integration agendas.
So I encourage everyone interested in helping to join with James, Tobie and others in the new Core Mobile Web Platform community group. Together we can get the specs that Web developers deserve, completed in the right order with multiple interoperating implementations.