Re: High performance timers and Web Perf

I agree, it's unclear how this benefits the game use case (which requires
low latency timing info).

-- Ben Garney


On Thu, Nov 3, 2011 at 12:58 AM, Marcos Caceres <w3c@marcosc.com> wrote:

> (cc'ed editor's of DOM4 for comment, given use case [1])
>
> On Wednesday, 2 November 2011 at 17:40, Philippe Le Hegaret wrote:
>
> > Folks,
> >
> > I just wanted to report that the Web Performance group talked about the
> > high performance times requirement yesterday [1].
> >
> > We've been changing our specs recently to allow sub-milliseconds:
> > [[
> > Throughout this work, time is measured in sub-millisecond resolution
> > since the start of the navigation of the root document, as recorded in
> > the navigationStart attribute. For example, a timestamp value of 1.5
> > would be 1.5 milliseconds since the start of the navigation of the root
> > document.
> > ]]
> >
> http://dvcs.w3.org/hg/webperf/raw-file/tip/specs/PerformanceTimeline/Overview.html
> >
> > However, we do realize that the current is still vague and that
> > developers would probably want more guarantees on what
> > "sub-milliseconds" means.
> >
> > We expect to put a short document together within two weeks on that.
> > I'll send a pointer to it.
> >
> > Hopefully, that will address the use case. If it does not, let us know.
> This API seems unrelated. I think we need to have DOM4 modified to support
> higher accuracy in the timeStamp of Events in a backwards compatible way.
>
>
>
>
>