[webkit-dev] Adding high resolution platform timestamp to DOM events

Ojan Vafai ojan at chromium.org
Thu Oct 25 10:55:34 PDT 2012

I don't understand. The www-dom discussion ends with a clear consensus to
use a new property name. There were no objections to systemTime. The
public-web-perf discussion didn't have an objections and just wanted to
wait until the V2 spec:

In what way does this not meet the bar for being checked in behind a flag?
Having an editor's draft spec has not historically (even recently) been a
requirement for checking things into WebKit behind a flag.

On Thu, Oct 25, 2012 at 10:12 AM, Robert Flack <flackr at chromium.org> wrote:

> FWIW, I don't plan to have it enabled by default on any platforms until
> we're sure what we want. The discussion with Anne about reusing timestamp
> was before she found out that timestamp calls for a 1970 based time by spec
> (http://lists.w3.org/Archives/Public/www-dom/2012OctDec/0022.html). I
> think we need a new property for this and systemTime has some consensus (as
> per the post from Anne), but I will of course not expose it to the public
> until we've reached a consensus and a proposed standard.
> On Thu, Oct 25, 2012 at 12:59 PM, Ryosuke Niwa <rniwa at webkit.org> wrote:
>> As Elliott pointed out, this property doesn't seem to be on any working
>> draft or editor's draft yet. And it doesn't seem like either thread on
>> www-dom or www-perf reached a consensus.
>> I'd appreciate if you waited until either thread reached a rough
>> consensus about the feature. Namely, www-dom thread discussion leads me to
>> believe that some people think we can simply modify timeStamp IDL attribute
>> instead of adding new one.
>> - Ryosuke
>> On Wed, Oct 24, 2012 at 6:17 PM, Robert Flack <flackr at chromium.org>wrote:
>>> Hi webkit-dev,
>>> I would like to add platform timestamps to DOM events as the systemTime
>>> property. I have a patch implementing the feature:
>>> https://bugs.webkit.org/show_bug.cgi?id=94987. This will let us know
>>> the time at which the system received an event to be able to accurately
>>> handle it, whereas the timestamp property gives the time the DOM event was
>>> created in an inaccurate milliseconds since 1970 form. This has been
>>> discussed on www-dom (
>>> http://lists.w3.org/Archives/Public/www-dom/2012OctDec/0028.html) and
>>> www-perf (
>>> http://lists.w3.org/Archives/Public/public-web-perf/2012Oct/0046.html)
>>> and use cases for this have been discussed (
>>> http://lists.w3.org/Archives/Public/www-dom/2012AprJun/0092.html).
>>> The platform timestamp comes in as a monotonic timestamp. Since the DOM
>>> Event spec requires that timeStamp() be a 1970-epoch based timestamp it is
>>> not sufficient for a high resolution precise time delta on event delivery.
>>> Instead, we add a systemTime property which uses the Performance API spec (
>>> http://dvcs.w3.org/hg/webperf/raw-file/tip/specs/HighResolutionTime/Overview.html)
>>> for high res timestamps (time since document load timestamp to avoid user
>>> fingerprinting) and provide the platform's high resolution timestamp to
>>> ECMAScript.
>>> Let me know if you have any suggestions. I look forward to everyone's
>>> feedback, cheers!
>>> - Rob
>>> _______________________________________________
>>> webkit-dev mailing list
>>> webkit-dev at lists.webkit.org
>>> http://lists.webkit.org/mailman/listinfo/webkit-dev
> _______________________________________________
> webkit-dev mailing list
> webkit-dev at lists.webkit.org
> http://lists.webkit.org/mailman/listinfo/webkit-dev
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-dev/attachments/20121025/1948a9a3/attachment-0001.html>

More information about the webkit-dev mailing list