Re: User Timing Mark Name for "Critical Content Loaded"?

Awesome.  navigationLoaded looks to be exactly what I was describing.  The
others look pretty similar to the ones defined in the spec but with
different names - any reason in particular (or were they created before the
named ones in the spec existed)?

I'd be happy to track the same named marks in Chrome though it would be
nice if we could get the spec and the Mozilla names to converge.

I opened a crbug for the chrome implementation:
https://code.google.com/p/chromium/issues/detail?id=503661

On Tue, Jun 23, 2015 at 12:37 PM, Eli Perelman <eperelman@mozilla.com>
wrote:

> Hey Pat,
>
> In Firefox OS, we adopted a series of de facto performance markers that we
> use for performance testing. There are five markers that are core to all
> applications and are quite generic to all apps that we create. Feel free to
> give it a look over and ask any questions about it.
>
>
> https://developer.mozilla.org/en-US/Apps/Build/Performance/Firefox_OS_app_responsiveness_guidelines#Implementation
>
> Thanks!
>
> Eli Perelman
> Mozilla
>
>
>
> On Tue, Jun 23, 2015 at 9:33 AM, Patrick Meenan <pmeenan@webpagetest.org>
> wrote:
>
>> There are some recommended mark names
>> <http://www.w3.org/TR/user-timing/#introduction-1> in the user timing
>> spec but none of them seem to be a good fit for what I am trying to track.
>> I'd like to see a well-defined name that sites can use to mark when they
>> believe their critical content has loaded.  i.e.:
>>
>> - "Time to first tweet" for twitter
>> - "Time to first pin" for pinterest
>> - Time when the hero image loads for an e-commerce or news site
>>
>> "mark_above_the_fold" is the closest but is semantically quite
>> different.  Any chance we can update the spec with one more recommended
>> name?  Something like "critical_content_loaded" (not sure why the others
>> are prefixed with mark_ since they are all marks by definition)?
>>
>> Understanding that it's just convention and nothing is guaranteed, a
>> standard convention for "this is the time I care about most" would be
>> helpful for a bunch of cases:
>>
>> - RUM providers could automatically collect it and report on it as a
>> special case
>> - Synthetic testing (like WebPageTest) could extract it and report on it
>> as a top-level metric
>> - Browsers could collect it as part of their field metrics and optimize
>> for it
>>
>> The last case in particular is one that I'd like to implement in Chrome.
>> We track all marks right now as a histogram but the data isn't terribly
>> useful because we don't really know what site developers are tracking and
>> onload is not terribly useful to optimize for. Having developers (even a
>> small set of high-profile sites) instrument a well-known "optimize for
>> this" metric would be significantly more useful.
>>
>> We could advertise a name that we'd be tracking but I think it would be a
>> lot more impactful if we could have the standard name added to the spec and
>> just point to that.
>>
>> Thanks,
>>
>> -Pat
>>
>
>