RE: [minutes] 2012-01-11 Web Performance WG Teleconference #57

Karen and I have pushed up changes to fix the resource linking issues for test_navigation_type_reload.html, test_timing_attributes_order.html and blank_page_unload.htm. Tony, please review these test cases.

Navigation Timing Test Cases
ACTION 21: Fixed resource linking in test_navigation_type_reload.html
ACTION 75: Fixed resource linking in test_timing_attributes_order.html and added blank_page_unload.htm

Thanks,
Jatinder

From: Jatinder Mann [mailto:jmann@microsoft.com]
Sent: Wednesday, January 11, 2012 1:00 PM
To: public-web-perf@w3.org
Subject: [minutes] 2012-01-11 Web Performance WG Teleconference #57


Meeting Summary:



1.      Timing Spec Updates

a.       ACTION-59, 61, 64 and 74

Jatinder made updates to Performance Timeline, Resource Timing, User Timing and Page Visibility to address ACTION-59, 61, 64 and 74. The WG will review these changes prior to closing the action items.



b.       Navigation Timing Test cases

Karen updated test_navigation_type_reload.html to address ACTION-21 and test_timing_attributes_order.html and blank_page_unload.html to address ACTION-75. The test is currently pointing to a wrong resource location that will be updated shortly. Tony will review the tests and make sure they work in all browsers.



Zhiheng had submitted test_document_readiness_exist.html and test_performance_attributes_exist_in_object.html. IE fails the former and passes the latter; Chrome and FF pass both tests. Karen to investigate and see why IE is failing on the first prior to next week.



Karen will be submitting a test case to test the NavigationStart attribute per ACTION-76 for next week.



ACTION-19 was closed as we don't believe a persistent connection test is necessary as this isn't a standardized feature and required by Navigation Timing.



2.      Client Authentication Phase

The issue of whether the Navigation Timing specification should define a phase that measures client authentication was brought up on the mailing list, http://lists.w3.org/Archives/Public/public-web-perf/2012Jan/0008.html, and conference call. Arvind will follow up with data to understand how common http client authentication is. If data indicates this phase is common enough that it merits inclusion in the spec, the WG feels that this can be covered in a future version of Navigation Timing.



3.      Navigation Timing to Proposed Recommendation

The WG has agreed to close all open Navigation Timing action items, which are limited to test suite updates, within this week and move the Navigation Timing specification to Proposed Recommendation next week, January, 18, 2012.


Detailed Notes:



Web Perf Teleconference #57 1/11/2012



IRC log: http://www.w3.org/2012/01/11-webperf-irc


Meeting Minutes: http://www.w3.org/2012/01/11-webperf-minutes.html



Attendees

Present for Navigation Timing, Resource Timing and User Timing (4-5PM EST/1-2PM PST)

Philippe Le Hegaret, Jatinder Mann, Arvind Jain, Karen Anderson, Tony Gentilcore, James Simonsen


Present for Page Visibility, Efficient Script Yielding, Display Paint Notifications (4-5PM EST/2-3PM PST)

Meeting cancelled.



Scribe

Jatinder Mann



Contents

Agenda

1.       Discuss and review submitted spec and test case updates.

2.       Discuss progress on remaining open action items

3.       Discuss client authentication timing attribute



--------------------------------------------------------------------------------



Topics: Spec and Test Updates from last week

Jatinder: Karen and I have updated the Page Visibility, and Timing specs to complete Action items 59, 61, 64 and 74.
... Please review those changes.
... Karen has pushed out two test cases in Action 21 and 75.

http://www.w3.org/2010/webperf/track/actions/18

http://www.w3.org/2010/webperf/track/actions/19

Arvind: Let's go through Navigation Timing open items and see what remains.

<plh> close action-19

<trackbot> ACTION-19 Investigate tests around persistent connections closed

Tony: I'm inclined to say that persistent connections isn't a part of HTML5 spec and probably can't be tested. Let's close

close action-18

<trackbot> ACTION-18 Follow up on test coverage and propose new test cases. closed

http://www.w3.org/2010/webperf/track/actions/57

http://www.w3.org/2010/webperf/track/actions/75

http://w3c-test.org/webperf/tests/approved/navigation-timing/html5/test_timing_attributes_order.html

Tony: Future test cases should always go to submission folder for review prior to moving to approval folder.

Jatinder: Let's keep action-75 open so that we can fix the typo thats causing the test to fail.

<plh> http://w3c-test.org/webperf/tests/approved/navigation-timing/html5/test_navigation_type_reload.html

Jatinder: Karen to update test_navigation_type_reload.html and test_timing_attributes_order.html to fix the resource issue and Tony will review.

http://lists.w3.org/Archives/Public/public-web-perf/2012Jan/0008.html

Jatinder: What are the WG's thoughts on client authentication modal dialogs? Users could impact aggregate timing by sitting on a modal dialogs.

Tony: There are other slow javascript dialogs, alerts that could also impact the overall timing.

Jatinder: I don't think we would want to throw away the client authentication time, as the overall performance timeline shouldn't be impacted. Our options might be to include a new phase.
... we expect the client authentication dialogs to occur between connectStart and responseEnd.

Arvind: HTTP Client authentication might have smaller usage as time goes on.
... We should be able to get data to see how many webpages in the world use this feature.

Tony: We need a spec update here regardless. Either we add the new phases or how our current phases work around this.

Arvind: Well, the problem still exists for proxy servers. There may be many more such small edge cases like this.
... Either we create a new parallel time for authentication. I don't think we should broaden the scope of redirects.

Jatinder: Arvind will follow up with data on how common these issues are and see if we even want to make any change here.

Arvind: I think we should target getting this spec to PR.

Jatinder: Let's plan to update all test cases for next week and target entering PR then.

Tony: I agree with that plan.