[minutes] 2012-06-06 Web Performance WG Teleconference #75

Meeting Summary:



1.       Timing-Allow-Origin Processing Model

The Timing-Allow-Origin HTTP header was designed based on the CORS design. However, the CORS Section 7.2 Resource sharing check, http://dvcs.w3.org/hg/cors/raw-file/tip/Overview.html#resource-sharing-check-0, processing model has some steps that don't make sense in Resource Timing. E.g., the concept of 'omit credentials flag' in steps 4 and 5 don't make sense. As the processing model here is only a few sentences, we will include a tailored processing model for Resource Timing instead.



2.       Resources initiated within Web Workers

Web Workers are capable of initiating resource downloads and currently such resource are not tracked in any of the document's timeline. This is because the Performance interface currently doesn't implement WorkersUtil, making this interface not available to web workers. It is clear that if Performance interface available to web workers, the browsing context that created a web worker would have any resource downloaded by that web worker included in its timeline, even if other frames within the top level browsing context shared that web worker. However, as shared workers can be shared across top level browsing context and are their own browsing context, this will need additional thinking. The WG has agree to consider this work in V2 of this spec and to continue to not implement WorkersUtil in this version.




Detailed Notes:



Web Perf Teleconference #75 6/06/2012



IRC log: http://www.w3.org/2012/06/06-webperf-irc



Meeting Minutes: http://www.w3.org/2012/06/06-webperf-minutes.html



Attendees

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

Jatinder Mann, James Simonsen, Karen Anderson


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

Meeting cancelled.



Scribe

Jatinder Mann



Contents

Agenda

1.       Resource timing feedback



--------------------------------------------------------------------------------
James: Now that I've looked at the CORS Section 7.2, I think it is probably better to just write our own processing model.
Jatinder: I agree that a processing model that is more tailored to Resource timing will be better. For example, 'omit credentials flags' doesn't make sense. I will share my proposed text update with the WG.
James: For the question about Web Workers, I think your mail Jatinder made sense (creator of web worker should get web worker's resource in their timeline).
... The shared worker seems a bit complicated and I'm a bit inclined to look at that in V2.
Jatinder: I think that's fair.
... What were your thoughts on the performance.now() discussion?
James: From a timing point of view, I think it makes sense. I'm fine with the current design as well. Hopefully, others can chime in.
Jatinder: The concern of shared timebase gets resolved when you add performance.timing.navigationStart to performance.now().
Karen: James, can you iterate through the reasons why another proposal is better than the current design? I feel like the current design is simpler and meets all the goals, including the issue raised.
James: I didn't have additional concerns. JamesR and others were looking at it from animations point of view. They may not have been famaliar with performance.timing.navigationStart as we were.
Jatinder: That is all I had on the agenda for today.