[cm2187]

It's a bit of a subjective debate. But having to run all that interpreted code in the background is probably not making the engine leaner and faster. I am sure it would be much more efficient to declare your intent in html and having the native renderer handling it. And certainly a lot less verbose and subject to bugs.

Say you add to the input tag a validation attribute with an URL and the browser posts the content of the input box in the background for validation, and give feedback to the user in a standard but customizable way.

For auto-completion, same thing, just adding one attribute with a url and a standard feedback.

Also an attribute to allocate a change to a css property when clicking on an element, to enable css pop ups.

There are probably a dozen more attributes like that and that would pretty much make javascript redundant on 99% of the website without making the HTML syntax much heavier (and overall your page would be lighter, because you don't have to write the plumbing every time). And you save having to interpret javascript so it has to be more efficient in term of performance too.

One counter example is something like bootstrap, where javascript is used to adapt the formatting to the form factor. But this is a good example of a horrible use of javascript, having to use javascript for presentation. It is just a case of html not being fit for purpose (since its purpose has evolved).