W3C home > Mailing lists > Public > public-web-perf@w3.org > January 2011

Re: About window.performance namespace

From: Zhiheng Wang <zhihengw@google.com>
Date: Fri, 7 Jan 2011 13:50:45 -0800
Message-ID: <AANLkTi=TVFaioGHQv0Eu3aM2cfVbEZKoHOBszYaMs=SF@mail.gmail.com>
To: Jonas Sicking <jonas@sicking.cc>
Cc: Tony Gentilcore <tonyg@google.com>, Patrick Meenan <pmeenan@webpagetest.org>, Anderson Quach <aquach@microsoft.com>, Sigbjørn Vik <sigbjorn@opera.com>, public-web-perf <public-web-perf@w3.org>
Hi, Jonas,

On Fri, Jan 7, 2011 at 1:26 PM, Jonas Sicking <jonas@sicking.cc> wrote:

> On Fri, Jan 7, 2011 at 12:45 PM, Zhiheng Wang <zhihengw@google.com> wrote:
> >    Thanks to all for chiming in with many perspectives! Given the
> imminent
> > timing line let's make the call on this
> > topic and move forward: NavigationTiming will use window.performance
> object,
> > which itself is replaceable.
> >    Here are the points we've covered so far to reach this decision:
> >   - There is interest to keep the interface concise, which has its own
> > advantage in the long run.
> >   - Most of us agree that we should avoid the potential conflicts with
> > existing pages. Having window.performance replaceable
> >     address that.
> >   - Probably a common practice, having window.performance replaceable
> could
> > still confuse some. But so far there
> >     doesn't seem to be any objection to that.
> >   - There is some risk allowing developers to replace window.performance.
> > But considering most objects/functions
> >     are replaceable in ECMA scripts, protecting window.performance alone
> > could be a half-way solution to integrity
> >     of the collected. So best intentions are assumed.
> >   - Less of an argument... but this is the current implementation adopted
> by
> > IE and Chrome.
>
> I'm still not understanding this. As far as I can see only the last
> point, which you are saying is "less of an argument", seems to be an
> argument for using the name "performance".
>

    Yes, it's not a deciding factor but still an argument in practice.


>
> I'll point out, yet again, that my question of why not using
> "pagePerformance" or "performanceMetrics", still remains unanswered.
>

    window.performance is short and terse so it bears more weight than
others.
And, when it's on billions of pages, every byte counts. :-)

cheers,
Zhiheng



>
> / Jonas
>
Received on Friday, 7 January 2011 21:51:15 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:04:29 UTC