[minutes] 20110629 Web Performance WG Teleconference #39

Meeting Summary:



1.      Browser feature support

a.       IE10 Platform Preview 2 supports new Web Performance APIs

The second IE10 Platform Preview was released yesterday with support for the new web platform features we have been working in the Web Performance WG: requestAnimationFrame API, Page Visibility API and setImmediate API. Thank you for the WG on making such fast progress on these specs.



b.       Firefox 7 to support Navigation Timing

With Firefox support, three major browser vendors will support this API. Congratulations!



2.      TPAC 2011 to be in held in Santa Clara first week of November 2011

TPAC 2011 is taking place in Santa Clara, CA the first week of November this year. Based on whether there is interest in the working group for a face to face meeting, we can book November 3rd or 4th. This will be a good opportunity to not only work face to face, but also evangelize with the working groups that will be present. E.g., Web Applications earlier in the week and HTML5 later in the week. Philippe has sent mail to the mailing list to solicit interest.



3.      Discuss Unified Proposal

a.       Updates to include inheritance model into Navigation, Resource and User Timing forthcoming

Nic, Karen and I have been working on making updates to these specs to use the inheritance model. I will send out a draft by the end of this week for the working group to review.



b.       Cancelling first half of next week's meeting

Due to the number of working group members on vacation next week, we will the Navigation, Resource and User Timing portion of next week's conference call (first half of next week's meeting).



4.      Feedback and discussion on the setImmediate specification

The setImmediate spec draft was published on Tuesday as an Editor's Draft. Based on the conference call conversation, the API conceptually made sense, however, the working group wanted time to read the spec and respond with feedback on the mailing list.



5.      Discuss requestAnimationFrame open issues
The following action items remain on Cameron McCormack and James Robinson:


1.       ISSUE-1: Scheduling processing model needs to be more tightly defined 2011-05-18

No major re-write here issue, text just needs to be more clear.



2.       ISSUE-2: Callback time parameter needs definition 2011-05-18

The spec needs to be more clear that callbacks should match the refresh rate of the display.


3.       ISSUE-3: Animation frame times should be monotonically increasing 2011-05-18

The spec needs a section to clarify that the animation frame times should be monotonically increasing, similar to what was stated in the Timing specs: http://www.w3.org/TR/2011/WD-resource-timing-20110524/#monotonic-clock. The spec should also include window.animationStartTime as a standardized way to get the start time for all animations in a given refresh interval, allowing them to remain in sync with one another.


4.       ISSUE-4: We perhaps should support an element parameter to requestAnimationFrame()

The spec will attempt to define an optional element parameter to this API.


5.       ISSUE-5: Expected callback rates should be documented 2011-05-18

The spec needs to state that the expected callback rate when the animation is visible should be matching the display refresh rate, and when not visible, a throttling algorithm should be used (fixed slow rate, exponential backoff, no callbacks). The throttling algorithm used shouldn't be a normative requirement, just that throttling is used.


6.       ISSUE-6: Spec needs to clarify expected behavior for duplicate calls of the same callback 2011-05-25

The spec needs to be updated to state that duplicate calls for the same callback will be respected.


7.       ISSUE-7: FrameRequestCallback interface should be designated as Callback=FunctionOnly

The spec will continue to use the default value for the specIDL. The webIDL spec will make clarify what the correct default behavior should be.



6.      Discuss Page Visibility open issues

a.       Update document.visibilityState definition

The working group agrees that we should keep the document.visibilityState attribute, as both a place to place vendor prefixed visibility states and future visibility states. The WG agreed to keep PAGE_VISIBLE and PAGE_HIDDEN as required states, and PAGE_PREVIEW as an optional state. The WG also agreed to remove the PAGE_PRERENDER state from the spec. Further, a section on the syntax for adding a vendor prefixed state is to be added to the spec by Jatinder. These updates are due for July 6th.



b.       Creating a test suite for Page Visibility

To ensure that browser implementations are interoperable, we are to add test cases to the test suite. Karen Anderson took an action item to add Page Visibility test cases. This update is due for July 6th.



c.        Last Call for Page Visibility

Considering there are two browser implementations for Page Visibility and the spec has been stable, we are going to set July 13th as the date for Page Visibility to enter Last Call. By July 6th, all spec updates and test case updates should be made. This will give a week prior to entering last call. The last call period will be 4 weeks.



We had the following action items from this meeting:

1.       Jatinder Mann: Update NT, RT, UT specs to include inheritance model.

2.       Cameron McCormack and James Robinson: Make requestAnimationFrame spec updates for next week.

3.       Jatinder Mann: Update Page Visibility document.visibilityState definition and add vendor prefixed section for next week.

4.       Karen Anderson: Add Page Visibility test cases.



Detailed Notes:



Web Perf Teleconference #39 6/29/2011



IRC log: http://www.w3.org/2011/06/29-webperf-irc


Meeting Minutes: http://www.w3.org/2011/06/29-webperf-minutes.html



Attendees

Present for Navigation Timing, Resource Timing and User Timing (4-5PM EST/1-2PM PST)
Jatinder Mann, Zhiheng Wang, Philippe Le Hegaret, James Simonsen, Nic Jansma, Karen Anderson, Tony Gentilcore


Present for Page Visibility, Efficient Script Yielding, Display Paint Notifications (4-5PM EST/2-3PM PST)
Arvind Jain, Jason Weber, Jatinder Mann, Philippe Le Hegaret, Nic Jansma



Regrets

James Robinson, Christian Biesinger, Kyle Simpson, Cameron McCormack



Scribe

Jatinder Mann



Contents

Agenda
1.       Discuss Unified Proposal
2.       Discuss feedback on the setImmediate specification
3.       Discuss Page Visibility open issues

4.       Discuss RequestAnimationFrame open issues
5.       Any other business




--------------------------------------------------------------------------------
Discuss Unified Proposal.

Jatinder: I'm excited to announce the second IE10 Platform Preview was released today with support for requestAnimationFrame, Page Visibility and setImmediate. I'd like to thank the working group for designing these specs.
Jatinder: I will have an updated draft for the specs incorporating the inheritance model before the end of this week for your review.
Zhiheng: I had some feedback on the getMarks and getMeasures APIs on the User Timing spec. Some clarification questions that may be resolved with your updates.
... Just wanted to let everyone know that FF7 will support Navigation Timing.
Tony: Sorry, I wasn't able to make last week's call. Is the current proposal the one discussed in Nic's last mail?
Nic: Yes, that's the one. We're working on making updates to the spec and then send out broadly.
Technical meeting at Santa Clara end of October, beginning of November. The Web Application WG Mon Tues Wed, HTML5 WG on Thurs and Fri. Unfortunately, we missed the deadline for getting a room. But if we're interested I could schedule something for Thursday.
Due to the vacations next week, we will be cancelling the first half of next week's web performance teleconference.

Discuss feedback on the setImmediate specification.

Jatinder: We uploaded the setImmediate spec to the mailing list yesterday. Has everyone had a chance to take a look at the spec?
Arvind: We are still taking a look at the spec internally at Google. We can send follow up mail with feedback, as we come with them.
Jason: Considering there are three implementations of requestAnimationFrame, two implementations of Page Visibility, and one for setImmediate we should start looking into locking down the spec. Why don't we go through the action items?

Discuss RequestAnimationFrame open issues.

Jatinder: Based on last week's call, we went through all 7 of the requestAnimationFrame open issues. Most of these are well understood and are just waiting on the editors to make changes.
Arvind: I will follow up with James to see if he's planning on making changes soon.

Discuss Page Visibility open issues.

Jatinder: For Page Visibility, there was one major open item: whether or not to include the visibilityState attribute. Based on the mailing list, it appears that there is value in keeping visiblityState as a host for future visibility states or vendor prefixed states. We may want to leave the spec as is.
... There is one action item remaining to include a section on the expected vendor prefixed syntax, I will add that section.
Arvind: Considering prerender is only in a single vendor and is optional anyway, we should just remove the prerender state.
Jatinder: Agree.
... I will keep the PAGE_PREVIEW state as an optional visibility state.
Arvind: What is the plan for testing our implementations to make sure they are interoperable?
Jatinder: The test suite will help there. I will take an action to release a page visibility test case.
... Considering the spec is very stable, we will have a test suite, we should consider taking the spec to last call.
Arvind: I agree we should take it to last call.
Jatinder: How about we enter Page Visibility Last Call on July 13th? That will give time to finish making the changes to the spec and test suite next week and another week prior to moving to last call.
Arvind: Agreed.