[legohead]

This shouldn't be up to the browser. He even notes edge cases where this performs unexpectedly -- unacceptable.

If this is such an issue and can actually be a performance increase, then someone should release a script with the same functionality. Or optionally make it an opt-in option in settings.

[marten-de-vries]

They do not violate the relevant specification. They just implement it in a way that has not been done before with the user's convenience in mind.

That aside, your position is unrealistic: browsers regularly break non-spec conforming websites. They actually monitor such cases (telemetry) and try to work with popular websites to fix the issue before they ship the breaking update, but it's a tradeoff that is regularly made nonetheless.

[xg15]

While the actions were beneficial in this case, that argument would be more convincing if the spec itself weren't a constantly moving "Living standard" maintained by the exact same organisations that also develop the browsers. And even that spec is sometimes consciously broken in an "intervention".

[testvox]

It shouldn't be up to the browser to implement web standards? What should browsers be doing then?

[legohead]

What is the web standard about delaying scripts deliberately encoded into the page?

Why downvote and no explanation? I want to know if I missed something.

[bzbarsky]

Per the standards, scheduling and prioritization of downloads is up to browsers.

The standard also defines that scripts that are not async have to be executed before later content can be parsed (because they can document.write()). They can still be loaded with any priority the browser wants; they just need to block observable DOM construction.

[legohead]

Thank you.