[minutes] 20150812 Web Performance

Available at

   http://www.w3.org/2015/08/12-webperf-minutes.html

Text version:

                      Web Performance Working Group

12 August 2015

    [2]Agenda

       [2] 
https://lists.w3.org/Archives/Public/public-web-perf/2015Aug/0007.html

Attendees

    Present
           Michael, Todd, Ilya, Eli, Mark

    Regrets
           plh

    Chair
           Todd

    Scribe
           igrigorik

Contents

      * [3]Topics
          1. [4]Next steps with Animation Timing spec...
          2. [5]requestIdleCallback to FPWD
          3. [6]Navigation Timing
          4. [7]Frame Timing
          5. [8]Beacon issues
      * [9]Summary of Action Items
      __________________________________________________________

Next steps with Animation Timing spec...

    ToddReifsteck: we should wait for plh@ for right steps to do a
    crisp handoff
    ... we should ask plh@ how to handoff outstanding issues

    AI: we'll wait for next call with plh@ to resolve this one

requestIdleCallback to FPWD

    mpb: any concerns over visited, does that need to block FPWD?

    ToddReifsteck: doesn't need to block, we can publish FPWD

    igrigorik: I'll check if we can publish, might need plh to flip
    right bits

Navigation Timing

    <ToddReifsteck>
    [10]https://github.com/w3c/navigation-timing/pull/32

      [10] https://github.com/w3c/navigation-timing/pull/32

    ToddReifsteck: no objections, will merge

Frame Timing

    Next item: [11]https://github.com/w3c/frame-timing/issues/46

      [11] https://github.com/w3c/frame-timing/issues/46

    <ToddReifsteck> Upcoming:
    [12]https://github.com/w3c/performance-timeline/issues/41

      [12] https://github.com/w3c/performance-timeline/issues/41

    <ToddReifsteck> Upcoming:
    [13]https://github.com/w3c/performance-timeline/pull/43

      [13] https://github.com/w3c/performance-timeline/pull/43

    mpb: keeping them separate allows us to add different
    attributes in the future (something we discussed but don't have
    today.. yet)

    eliperelman: the paths to both of them are very different,
    there might be some associated overhead..

    ToddReifsteck: we can add the counter event interface in the
    future if we get a 3rd or 4th
    ... wontfix, will followup on the bug

    <ToddReifsteck> Next item:
    [14]https://github.com/w3c/performance-timeline/issues/41

      [14] https://github.com/w3c/performance-timeline/issues/41

    ToddReifsteck: thinking is.. if PO has ability to report
    existing timeline then JS interface is much simpler (no races)
    ... same thing is possible via other means but we need a crisp
    processing model

    some issues: lookback doesn't play well with our current buffer
    model; we should nail the processing model first

    mpb, eliperelman: there is potential for missed events (e.g.
    exceeded buffer)

    looking at Chrome, changing buffer size is very rare

    mpb: a plausible v2 enhancement?

    [15]https://github.com/w3c/navigation-timing/issues/1

      [15] https://github.com/w3c/navigation-timing/issues/1

    <ToddReifsteck>
    [16]https://github.com/w3c/performance-timeline/pull/43

      [16] https://github.com/w3c/performance-timeline/pull/43

    ToddReifsteck: we should nail down the processing model, we'll
    resolve current issue as possible future enhacement

    any objections to going back to "new PerfObserver" approach?

    ToddReifsteck: no, we should iron out the language questions
    and get Anne/Boris to sign off

    Eli: no objection

    mpb: we'll follow up on the thread

Beacon issues

    <ToddReifsteck> [17]https://github.com/w3c/beacon/issues/10

      [17] https://github.com/w3c/beacon/issues/10

    [18]http://w3c.github.io/beacon/#sec-processing-model

      [18] http://w3c.github.io/beacon/#sec-processing-model

    <ToddReifsteck> [19]https://github.com/w3c/beacon/issues/5

      [19] https://github.com/w3c/beacon/issues/5

    Re, #10: igrigorik to followup, current spec is already defined
    via Fetch

    Re, #5: we'll define Beacon-Age, the similar use case for NEL
    is no longer relevant because NEL can deliver multiple reports
    in same payload

    <ToddReifsteck> Preload:
    [20]https://github.com/w3c/preload/pull/26

      [20] https://github.com/w3c/preload/pull/26

    <ToddReifsteck> 8/26 for next call

Received on Monday, 14 September 2015 14:29:14 UTC