Web Performance Working Group Teleconference

24 Jun 2015

See also: IRC log


Plh, Ilya, Bernhard, Todd, Michael


Review charter feedback and proposed process updates


Plh: I'll make it official tomorrow

<igrigorik> https://lists.w3.org/Archives/Public/public-web-perf/2015Jun/0080.html

/TR/performance-timeline would be in sync with the github draft.



todd: should we have a branch in the repo that also contains that snapshot?

Ilya: yes, snapshot with a branch seem reasonnable
... so the most recent would be gh-pages?

plh: yes

Resolved: Performance Timeline 2 can be published as a FPWD

Resolved: User Timing 2 can be published as a FPWD

Resolved: Server Tming can be published as a FPWD

Resolved: Preload can be published as a FPWD

/TR/performance-timeline -> /TR/performance-timeline-2


Resolved: adopted the new versioning and branching

Resolved: WDs should be automatically published


time origin & workerStart

<igrigorik> https://github.com/w3c/hr-time/pull/6

Ilya: we should deprecate workerStart in favor of a translated time based approach

Todd: didn't get feedback from Justin yet, makes sense from him.
... we should wait for Boris to give a +1 as well

plh: you don't need to deprecate workerStart. it never made it into a REC. just removing it is fine.

attribute filtering

<igrigorik> https://github.com/w3c/performance-timeline/pull/11

Ilya: [summarizing the history]. https://htmlpreview.github.io/?https://github.com/w3c/performance-timeline/blob/592576c94599a88aa6183d827e54fa46d822cdfb/index.html

Todd: ok to merge

Ilya: I'll merge and ping the list

revise "return copy"


Ilya: Chrome and Firefox do not return a copy of the array
... IE does
... can we remove the "copy"? it's still ok to return a copy

Todd: looks good to me

<mpb> Do we need to change language on the Performance Observer interface too?


Ilya: yes, perf observer shouldn't return a copy either

credentials in URLs


Todd: it's been suggested that exposing credentials in URLs through RT makes it easier to capture the credentials
... the move to Fetch will disallow this and fetch will speak to it

Ilya: we can drop the credentials, but you won't get the same value for the url. and if you use different credentials, you won't be able to differentiate it
... I propose to close it as "wontfix"
... and let the browsers make their own decision (ie do as IE by disallowing that type of request)

Resolved: wontfix


plh: drop it?


Ilya: we haven't been successful at implementing
... we could ressurect it in the future if we have access to the data.

Resolved: drop linkNegotiationStart

Navigation Timing 2

Todd: ALPN id / nextHopProtocol . has value and no plan for implementation
... ditto for redirectCount
... should we do a snapshot of NT2?

plh: pb is that only IE supports it on top of performance timeline

[discussion on the need to do snapshots of HRT2, PT2, and RT first]

Ilya: open question related to nextHopProtocol into exposing too much networking information

Todd: didn't think of that yet

Ilya: redirectCount. was about making resource timing multi-fetch friendly
... if NT2 gets defined on top of RT, we won't need that

Todd: also some concerns about security/privacy of that one

Call schedule for the summer

Ilya: next one is 07/15
... then 08/12


Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.140 (CVS log)
$Date: 2015/06/26 00:00:00 $