> other JS engine developers aren't necessarily enthused about signing on to support a spec that [company] developed mostly in secret without input from other JS engines, relies extensively on engine quirks, and doesn't yet solve many hard problems
And I would argue that any spec tagged with these kinds of complaints rarely ever sees broad adoption, regardless who made it. From my perspective, the only thing asm.js has going for it that offsets the downsides is that the code at least runs in other JS engines even if they don't explicitly support it. And that alone might actually be enough.
Wow, where have I heard that before? ;)