Web Performance Working Group Teleconference

<trackbot> Date: 10 December 2014

Frame Timing

plh: double cpuTime should be HRT instead

Michael: based on conversation, cpuTime will be the whole clock during from the start of the frame to done processing
... sourceFrameNumber will be the frame index

<scribe> ACTION: Plh to update cpuTime to use HR time [recorded in http://www.w3.org/2014/12/10-webperf-minutes.html#action01]

<trackbot> Created ACTION-146 - Update cputime to use hr time [on Philippe Le Hégaret - due 2014-12-17].

https://github.com/w3c/frame-timing/pull/19

igrigorik: we need to update that one
... and Michael will update it

Michael: yes, will do

Resolution: publish the FPWD once Michael is done with the change

plh: moratorium is tuesday or thursday next week

igrigorik: we had good feedback from mozilla and ms

Performane Observer feedback

igrigorik: maybe related to frame timing
... we have outstanding issues to trigger when new resources are added to RT
... the proposal is to add a subscriber
... but frame timing fires often
... proposal is to define a new "mutation" event for performance timeline

plh: we should make sure anne is in the loop on this imho

igrigorik: we probably need to formalize the use cases here

plh: I'll follow on the list

Server Timing feedback

igrigorik: we circulated a draft with use cases and API
... ideas at the moment
... feedback seems positive
... we need an official repo for it

<scribe> ACTION: plh to make sure the repo [recorded in http://www.w3.org/2014/12/10-webperf-minutes.html#action02]

<trackbot> Created ACTION-147 - Make sure the repo [on Philippe Le Hégaret - due 2014-12-17].

<scribe> ACTION: plh to make sure it's ok with the charter [recorded in http://www.w3.org/2014/12/10-webperf-minutes.html#action03]

<trackbot> Created ACTION-148 - Make sure it's ok with the charter [on Philippe Le Hégaret - due 2014-12-17].

Splitting Resource Hints / Preload

igrigorik: confusion about mandatory hints, etc.
... from TPAC
... best way is to separate preload from the rest
... put the optinal semantic in a separate spec

plh: I can do either way

igrigorik: it's easier to implement
... I'll send a follow up asking for objection within a week

NT and getEntriesByType

igrigorik: chrome and ff returns an entry array
... ie returns data

plh: original thinking was to harmonize the specs and only use performancetimeline extension

igrigorik: in that case, we always return an array of size 1
... we 'll need a test as well

plh: I'll raise an issue to follow up

Resource Timing

igrigorik: in the current spec, we referece RT L2
... that should be removed

plh: I'd like to move the spec back to WD
... that should take care of the note

igrigorik: why didn't we address audio/video in L1?

plh: I think we weren't sure about all the use cases at that time

<scribe> ACTION: plh to update RT in /TR [recorded in http://www.w3.org/2014/12/10-webperf-minutes.html#action04]

<trackbot> Created ACTION-149 - Update rt in /tr [on Philippe Le Hégaret - due 2014-12-17].

Page Visibility 2

plh: did you reach a conclusion with the chrome team?

igrigorik: Nat still opposes the change
... we'd be much better off tackling element visibility
... instead of hacking iframe
... I'll follow up on this front

next meeting

plh: I'll ping folks to see if we need a meeting next week