~/talk

Predicting User Activity to Make the Web Fast

Talkat Velocity US 2012 · 2012-06-25..06-27· [web-performance, web-platform]
links:none on file
links:none on file
slides:none on file

Chrome uses prerendering to fetch all resources into a hidden tab and swap it in when navigation happens, using an omnibox prediction model that stores hit/miss counts for suggestions to compute confidence. Developers can opt in with HTML5 link rel=prerender and detect prerender state via the Page Visibility API. Results: about one third of omnibox navigations are prerendered, 90% of those are used, 15–20% become instant (<10 ms), saving ~1 second per navigation and ~30 years cumulatively per day.

AI-written summary · about

  1. 0:00Introduction
  2. 1:05Current web speed challenges
  3. 3:51User behavior prediction concept
  4. 5:16Omnibox pre-rendering model
  5. 8:04Pre-rendering demo
  6. 9:27Key results and impact
  7. 9:48Developer recommendations
~/search
type to search 6,518 conferences, 143,918 speakers and 192,133 talks
~/help
cmdctrl k
search every conference, speaker and series
/
focus the filter search
h
home
c
conferences
s
speakers
e
series
t
light / dark — the status bar button does the same
← →
on a talk: 5 seconds back / forward
0–9
on a talk: jump to 0%–90% of the recording
?
this help
esc
close this, then clear every filter — or just the search, while you are typing in it

Ignored while you are typing. Pages without a filter rail have no search to focus.