RE: [NavigationTiming] another update

Hi all,

Here are some points of feedback on the Navigation Timing [1].

Section 4.2
navigationStart
-- comments
Account for when there is no previous document that was unloaded
-- recommended text
This attribute must return the time immediately after the user agent finishes prompting to unload the previous document. If there was no previous document, return the time when the navigation was committed by the user.

redirectStart
-- comments
Typo, referring to the attribute navigateStart
-- recommended text
Note: redirectStart is the equivalent of the navigationStart attribute of the navigation that initiates the redirect.

redirectEnd
-- comments
Consistency with the term last byte/chunk/trunk
Typo s/direct/redirect
Not sure what the second paragraph is referring to, perhaps just add the non-redirect clause to the previous paragraph in the definition.
-- recommended text
If there are HTTP redirects or equivalent when navigating and if the document before the redirect and the document at the end of the redirect are from the same origin, this attribute must return the time immediately after receiving the last byte of the response of the last redirect. Otherwise, this attribute must return zero.

domContentLoaded
-- comments
We should be consistent with how the load event is treated, that is break-out this marker to include domContentLoadedStart and domContentLoadedEnd

Section 4.4
-- The example includes a typo s/loadEnd/loadEventEnd
-- The navigation type is invalid, e.g. TYPE_NAVIGATE

Diagram
-- comments
Typo, s/navigateStart/navigationStart
[ prompt for unload ] comes before [ redirect ]
[ app cache ] should also include http cache and loading from a local file
-- feedback
Statements where the time associated to with fetching from the user agent cache needs to be a non-normative requirement as each user-agent will implement caches differently.

Undefined times, such as gaps between the phases need to be considered as non-normative requirements, as this can also be user-agent specific.

Processing Model
-- comments
1. Sub-bullets underneath 1 and 15, should use roman numerals or ordered list with letters.
3. s/then go to step 6/then go to step 5
6. Roll the constraint of loading a local file or access from cache into step 5.
15. sub-bullet 5. Do we want to zero out redirectCount?

-- recommended texts
1. If the navigation is aborted or there is no navigation, do not collect timing information.
3. If the current document and the previous document are not from the same origin set both unloadEventStart and unloadEventEnd to 0 then go to step 5. Otherwise record unloadEventStart as the time immediately before the unload event fires.
13. s/trunk/byte
14. s/entire response/last byte of the response
15. Sub-bullet 4
21. consistent wording with the definition
Set all the attributes in window.performance timing to except
window.performance.timing.navigationStart
window.performance.timing.unloadEventStart
window.performance.timing.unloadEventEnd
window.performance.timing.redirectStart
window.performance.timing.redirectEnd
21. Record the time as loadEventEnd immediately after the user agent completes the load event.

[1] http://dvcs.w3.org/hg/webperf/raw-file/tip/specs/NavigationTiming/Overview.html


From: public-web-perf-request@w3.org [mailto:public-web-perf-request@w3.org] On Behalf Of Zhiheng Wang
Sent: Tuesday, October 12, 2010 11:49 AM
To: public-web-perf
Subject: [NavigationTiming] another update

Hey, folks,

   Just want to pitch another update on the NavigationTiming draft, which fixes the links and adds a graphical overview of the timing attributes.

   https://dvcs.w3.org/hg/webperf/raw-file/tip/specs/NavigationTiming/Overview.html

cheers,
Zhiheng