Page Interactive versus Dom Interactive

The W3C resource timing has a metric for window.performance.domInteractive which as I understand it "marks the point when the browser has finished parsing all of the HTML and DOM construction is complete. (per Measuring the Critical Rendering Path)

However, in running various tests in Chrome and looking at the the WPT Page is Interactive , I have seen instances in which the page may not actually be interactive even though the dom.interactive metric says it is. So wanted to know if the dom.interactive is really not the metric to look at to determine if the page is interactive, since browser threading may be occuring thus causing the page to not be interactive. My reason for asking this is that some have told me to use the dom.interactive metric to see when the page is interactive, but it seems that is not really the case.

No, it’s not a great metric because it will depend on the page and app. A lot of them actually run most of their code after that which is why the Chrome team has been working on “Time to Interactive” which uses much more complex heuristics: GitHub - WICG/time-to-interactive: Repository for hosting TTI specification and discussions around it.

Patrick - Thanks for the clarification on this.