W3C home > Mailing lists > Public > public-web-perf@w3.org > July 2013

[minutes] Web Performance WG Teleconference #114 2013-07-17

From: Jatinder Mann <jmann@microsoft.com>
Date: Wed, 17 Jul 2013 18:13:20 +0000
To: "'public-web-perf@w3.org'" <public-web-perf@w3.org>
Message-ID: <d38ae444cfbb4417a966f1f89c9fe709@BLUPR03MB065.namprd03.prod.outlook.com>
Meeting Summary:



1.     New Test Cases

Microsoft has submitted new Performance Timeline and Navigation Timing L2 test cases which need to be reviewed by the working group:

http://w3c-test.org/webperf/tests/submission/Microsoft/NavigationTimingL2/test_navigation_timing2_objects.html

http://w3c-test.org/webperf/tests/submission/Microsoft/PerformanceTimeline/test_performance_timeline_functions.html


The working group will also be looking into moving our test repository to https://github.com/w3c/web-platform-tests. Philippe to follow up with details.



2.     Web Workers in Resource Timing L2

The working group wants to add web worker support in Resource Timing L2. However, there are a few design decisions that need to be closed on: should a dedicated worker share its data with the owning document or not? The working group has raised a few interesting scenarios that need to be considered (see minutes), and will discuss in further meetings or mailing list what design we want to go with.





3.     Publish New Specs as FPWD

The working group has now gone through each of the new specifications and discussed designs. We also have editor's drafts for most of the specs. For the next couple of weeks, we want editors to clean up these drafts, and we can discuss when to publish these specs as FPWD once the drafts are cleaned up.



            Philippe, it may be useful to talk about the exclusion process for IP purposes, as some members were interested in hearing more.




W3C Web Performance WG Teleconference #114 2013-07-17





IRC log: http://www.w3.org/2013/07/17-webperf-irc



Meeting Minutes: http://www.w3.org/2013/07/17-webperf-minutes.html



Attendees

James Simons, Jatinder Mann, Ganesh Rao, Alois Reitbauer, Dan Austin


Scribe

Jatinder Mann



Agenda

1.     Test Cases

2.     Resource Timing L2

3.     FPWDs of New Specs

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



Minutes:


Test Cases

Jatinder: http://w3c-test.org/webperf/tests/submission/Microsoft/NavigationTimingL2/test_navigation_timing2_objects.html
... This test case tests basic Navigation Timing L2 support by testing if the performance.getEntries()[0], performance.getEntriesByName('document'), and performance.getEntriesByType('navigation') return a PerformanceNavigationTiming object
... http://w3c-test.org/webperf/tests/submission/Microsoft/PerformanceTimeline/test_performance_timeline_functions.html
... This test case tests if the performance.getEntries(), performance.getEntriesByName(), and performance.getEntriesByType() functions are supported.
... There was also the question on when we are going to move to using https://github.com/w3c/web-platform-tests? I know Philippe took an action to find out change in process.

James: Should we continue to using the old respository?

Jatinder: I think we should continue using the old process until we change the process.
Web Workers in Resource Timing L2

Jatinder: For web workers, we need to decide whether the resources downloaded by the web worker should show up in the main document associated with the worker, or only when the buffer is requested from the web worker context though the performance.getEntries() methods. The opposite is also interesting, should the resources used in the main document by shown when calling performance.getEntries() methods from the worker context? Sounds like an in[CUT]

Dan: The main idea we were thinking about when this was raised in HTML WG, was to be able to download the timing data in the main page through a background thread via the worker.

Jatinder: The real world use case is to JSON.stringify(performance.getEntries()) to send the data to the server and do analysis there.
... So the option is either JSON from the main document to get all data, or JSON from the main document and all the workers
... We should keep in mind that analystics will just be a single script added to a web page. The analytics would want to have a view of all the resources used, without asking each worker (which isn't possible for the analytics).

James: For Shared workers, we would definitely need to keep the context seperated, as it doesn't have a view into the main document.

Jatinder: Our options might be to either completely seperate workers from the documents, or to keep dedicated workers connected with the main document, and just have the shared worker seperate. I think each has an interesting case, and we should think about it some more before deciding.
... There is also the point that High Resolution Time defines the dedicated worker's time origin as the same as the main document, but the shared worker is the time of creation of the shared worker.

Dan: Yes, let's think about the different customer scenarios more carefully before closing on the direction that we want to go in.
FPWDs of New Specs

Jatinder: At this point, we have discussed most of the new charter specs in the working group and have put together drafts of most of the specs. I want to start cleaning up all of the specs and start publishing them as FPWD. Let's spend the next couple of weeks reviewing open issues and closing up on the drafts.

Dan: I'd be interesting in hearing from Philippe what the process is for exclusion from discussion - I believe its at the FPWD point that we can choose to exclude ourselves or not.
... I think once all of the drafts have been cleaned up, I would like to see them go to FPWD.
... What about the task of putting together a draft of best practices?

Jatinder: I think we are really just looking for a volunteer to put together a draft. Once we have a draft the working group can review.

Dan: I will volunteer to put together a draft.

Jatinder: Let me know if you need any help in putting together a Note with this information. We should loop in Philippe as well to close on what kind of document this will be.
Received on Wednesday, 17 July 2013 18:16:14 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:04:36 UTC