[minutes] Web Performance Group 2015-02-11

Available at
 http://www.w3.org/2015/02/11-webperf-minutes.html

Summary:
- We needs additional security review before remove "navigation"
requirement from NEL
- PerformanceObservers proposal looks good but needs refinement on list
- User Timing can have negative duration
- Resource Timing/User Timing will get exposed to Service Workers

- Server Timing to be published as FPWD
- Editors are allowed to update Working Drafts at will, as long as they
  notify public-web-perf@w3.org of the update



Text version:

              Web Performance Working Group Teleconference

11 Feb 2015

Attendees

   Present
          Plh, Yoav, Ilya, Todd

   Regrets
          Eli

   Chair
          plh

   Scribe
          plh

Contents

     * [2]Topics
         1. [3]NEL
         2. [4]PerformanceObservers
         3. [5]Server TIming
         4. [6]User Timing
         5. [7]nav timing 2
         6. [8]RAF next steps
         7. [9]Automatic publishing
         8. [10]Service Workers support for
            window.performance/resourcetiming/usertiming
     * [11]Summary of Action Items
     __________________________________________________________

   <trackbot> Date: 11 February 2015

NEL

   Ilya: I'm making progress on feedback
   ... looking for feedback on removing navigation
   ... embedded case is missing
   ... if we remove navigation restriction, it becomes pretty
   simple

   Yoav: I think the use case is useful
   ... concern is security
   ... should we expose it?
   ... need to get security folks involved

   Ilya: we wouldn't be reporting info that would have been
   reported if it had been successful
   ... but third party reports are more problematic
   ... third.com has a resource, someone tries to load on a
   different site. That should be fine to report

   Yoav: notifying the embedder might be problematic

   Ilya: right. maybe opt-in policy

   Todd: still thinking it through

   Tobin: do we have an idea of the clear scenarios?

   Ilya: no, we have an example section
   ... but we could add it
   ... that's good idea
   ... I'll follow with a new draft

PerformanceObservers

   Plh: changing the type of PerformanceEntryList is problematic
   ... right now, it's an array
   ... forEach, map, filter, etc
   ... we would remove those functions otherwise
interface PerformanceEntryList {
  getter PerformanceEntry item(unsigned long index);
  readonly attribute unsigned long length;
  PerformanceEntryList getEntriesByName(name);
}

   Plh: I'll follow up on the list
   ... otherwise, proposal looks fine

   Tobin: I like the idea
   ... better object model for mass events
   ... will keep thinking about it

   ??: in general, high level is looking good

Server Timing

   Yoav: looks good
   ... was wondering about the possibility of allocating more
   ... but other than that, it's fine
   ... looks useful as well
   ... why 150 limit?

   Ilya: copied from resource timing
   ... we could change it

   Yoav: why not start higher?

   [...]

   Yoav: ok, makes sense

   Resolved: Server Timing is published as FPWD

User Timing

   Plh: can't we remove SYNTAX_ERR limitation?

   Ilya: don't see any reason why not

   Plh: negative duration restriction...

   [back and forth around whether negative values are useful]

   ??: given the scenario in the email, it sounds legitimate

   Todd: I'll test it in IE

nav timing 2

   plh: Why can't we get
   performance.getEntriesByType("navigation") implemented?

   <scribe> ACTION: plh to send an email about what's new in
   navtiming2 [recorded in
   [12]http://www.w3.org/2015/02/11-webperf-minutes.html#action01]

   <trackbot> Created ACTION-150 - Send an email about what's new
   in navtiming2 [on Philippe Le Hégaret - due 2015-02-18].

RAF next steps

   Ilya: we also redirect handling for R and NT

   <scribe> ACTION: plh to send an email on RAF [recorded in
   [13]http://www.w3.org/2015/02/11-webperf-minutes.html#action02]

   <trackbot> Created ACTION-151 - Send an email on raf [on
   Philippe Le Hegaret - due 2015-02-18].

Automatic publishing

   plh: can we allow editors to publish in /TR if the editor is
   willing to send email to the list?

   Resolved: Editors are allowed to update WD in /TR

Service Workers support for
window.performance/resourcetiming/usertiming

   Ilya: service workers wants to have access to
   window.performance, RT and UT
   ... needs to update the spec to not specifically call out
   window

   ??: could that be addressed by performance observers?

   Ilya: that's different
   ... I'll take a look at it

   <scribe> ACTION: plh to make it clear that duration can be
   negative in UT [recorded in
   [14]http://www.w3.org/2015/02/11-webperf-minutes.html#action03]

   <trackbot> Created ACTION-152 - Make it clear that duration can
   be negative in ut [on Philippe Le Hegaret - due 2015-02-18].

Received on Thursday, 12 February 2015 15:55:05 UTC