Re: [whatwg/dom] High resolution timing for events (#23)

> Ok let's take the rhetoric down a notch, we're all trying to do the best thing for the web platform.

I apologize for the rhetoric, and I agree fully that everyone here is trying to do what they believe is best for the platform and the browsers they work on.

> So if collecting more data on the impact in Chrome over a few months would be valuable to this debate, then that's fine with me.

I think more data would indeed be useful, if we measure the right things.

Here are some things that (dynamic) telemetry might be helpful if unearthing, if you're willing to run it:

* for each event:
  * comparisons of the value produced by `e.timeStamp` with the value produced by `Date.now()`.
  * whether the timestamp used in such those comparisons came from `e.timeStamp || Date.now()`. For events where Firefox produced ms-since-unix-epoch or 0, this would have previously been guaranteed to produce a date, but now it does not.
  * runtime behavior equivalent to `e.timeStamp || Date.now()` (there are many ways to do this coercion, so a textual analysis is not enough).
  * usage of `date.setTime(timestamp)` or `date.setTime(coercedTimestamp)`

I'd be happy to discuss the exact telemetry that would be most helpful here, if you find this line of reasoning useful.

I'd also add that finding "hits" with this kind of telemetry does not prove that a site is broken, but it will give us a corpus of sites to evaluate manually.

Personally, I'm especially concerned about the vast number of "iOS lookalike" sites that were hand-crafted early in the days of iOS, because those sites were forced to reimplement scrolling, implement high-level touch events, implement animations and more, without the benefit of `window.performance.now()` even existing at all to give them a clue that they should avoid comparing `e.timeStamp` with `Date.now()`.

Also, because those sites are largely "webkit only" in practice, they may well not have noticed the Firefox breakage **at all**, but these sites are on the Internet and most browsers have been attempting to support them.

---
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/whatwg/dom/issues/23#issuecomment-215506260